misc/libphysfs/lzma/CPP/7zip/Compress/LZMA_Alone/LzmaBench.cpp
author Wuzzy <Wuzzy2@mail.ru>
Tue, 24 Oct 2017 19:30:55 +0200
changeset 12751 838515c4e6c5
parent 12213 bb5522e88ab2
permissions -rw-r--r--
Rewrite INSTALL file: Markdown format, add troubleshooting section, more detailed build instructions, etc.
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
12213
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     1
// LzmaBench.cpp
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     2
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     3
#include "StdAfx.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     4
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     5
#include "LzmaBench.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     6
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     7
#ifndef _WIN32
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     8
#define USE_POSIX_TIME
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
     9
#define USE_POSIX_TIME2
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    10
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    11
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    12
#ifdef USE_POSIX_TIME
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    13
#include <time.h>
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    14
#ifdef USE_POSIX_TIME2
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    15
#include <sys/time.h>
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    16
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    17
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    18
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    19
#ifdef _WIN32
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    20
#define USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    21
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    22
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    23
#ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    24
#ifdef _WIN32
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    25
#include <malloc.h>
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    26
#else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    27
#include <stdlib.h>
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    28
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    29
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    30
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    31
extern "C" 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    32
{ 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    33
#include "../../../../C/Alloc.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    34
#include "../../../../C/7zCrc.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    35
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    36
#include "../../../Common/MyCom.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    37
#include "../../ICoder.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    38
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    39
#ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    40
#include "../../../Windows/Thread.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    41
#include "../../../Windows/Synchronization.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    42
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    43
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    44
#ifdef EXTERNAL_LZMA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    45
#include "../../../Windows/PropVariant.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    46
#else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    47
#include "../LZMA/LZMADecoder.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    48
#include "../LZMA/LZMAEncoder.h"
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    49
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    50
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    51
static const UInt32 kUncompressMinBlockSize = 1 << 26;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    52
static const UInt32 kAdditionalSize = (1 << 16);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    53
static const UInt32 kCompressedAdditionalSize = (1 << 10);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    54
static const UInt32 kMaxLzmaPropSize = 5;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    55
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    56
class CBaseRandomGenerator
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    57
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    58
  UInt32 A1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    59
  UInt32 A2;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    60
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    61
  CBaseRandomGenerator() { Init(); }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    62
  void Init() { A1 = 362436069; A2 = 521288629;}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    63
  UInt32 GetRnd() 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    64
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    65
    return 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    66
      ((A1 = 36969 * (A1 & 0xffff) + (A1 >> 16)) << 16) +
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    67
      ((A2 = 18000 * (A2 & 0xffff) + (A2 >> 16)) );
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    68
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    69
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    70
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    71
class CBenchBuffer
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    72
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    73
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    74
  size_t BufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    75
  Byte *Buffer;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    76
  CBenchBuffer(): Buffer(0) {} 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    77
  virtual ~CBenchBuffer() { Free(); }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    78
  void Free() 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    79
  { 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    80
    ::MidFree(Buffer);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    81
    Buffer = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    82
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    83
  bool Alloc(size_t bufferSize) 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    84
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    85
    if (Buffer != 0 && BufferSize == bufferSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    86
      return true;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    87
    Free();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    88
    Buffer = (Byte *)::MidAlloc(bufferSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    89
    BufferSize = bufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    90
    return (Buffer != 0);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    91
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    92
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    93
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    94
class CBenchRandomGenerator: public CBenchBuffer
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    95
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    96
  CBaseRandomGenerator *RG;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    97
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    98
  void Set(CBaseRandomGenerator *rg) { RG = rg; }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
    99
  UInt32 GetVal(UInt32 &res, int numBits) 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   100
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   101
    UInt32 val = res & (((UInt32)1 << numBits) - 1);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   102
    res >>= numBits;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   103
    return val;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   104
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   105
  UInt32 GetLen(UInt32 &res) 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   106
  { 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   107
    UInt32 len = GetVal(res, 2);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   108
    return GetVal(res, 1 + len);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   109
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   110
  void Generate()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   111
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   112
    UInt32 pos = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   113
    UInt32 rep0 = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   114
    while (pos < BufferSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   115
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   116
      UInt32 res = RG->GetRnd();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   117
      res >>= 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   118
      if (GetVal(res, 1) == 0 || pos < 1024)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   119
        Buffer[pos++] = (Byte)(res & 0xFF);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   120
      else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   121
      {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   122
        UInt32 len;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   123
        len = 1 + GetLen(res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   124
        if (GetVal(res, 3) != 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   125
        {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   126
          len += GetLen(res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   127
          do
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   128
          {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   129
            UInt32 ppp = GetVal(res, 5) + 6;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   130
            res = RG->GetRnd();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   131
            if (ppp > 30)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   132
              continue;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   133
            rep0 = /* (1 << ppp) +*/  GetVal(res, ppp);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   134
            res = RG->GetRnd();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   135
          }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   136
          while (rep0 >= pos);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   137
          rep0++;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   138
        }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   139
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   140
        for (UInt32 i = 0; i < len && pos < BufferSize; i++, pos++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   141
          Buffer[pos] = Buffer[pos - rep0];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   142
      }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   143
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   144
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   145
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   146
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   147
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   148
class CBenchmarkInStream: 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   149
  public ISequentialInStream,
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   150
  public CMyUnknownImp
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   151
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   152
  const Byte *Data;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   153
  size_t Pos;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   154
  size_t Size;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   155
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   156
  MY_UNKNOWN_IMP
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   157
  void Init(const Byte *data, size_t size)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   158
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   159
    Data = data;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   160
    Size = size;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   161
    Pos = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   162
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   163
  STDMETHOD(Read)(void *data, UInt32 size, UInt32 *processedSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   164
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   165
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   166
STDMETHODIMP CBenchmarkInStream::Read(void *data, UInt32 size, UInt32 *processedSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   167
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   168
  size_t remain = Size - Pos;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   169
  UInt32 kMaxBlockSize = (1 << 20);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   170
  if (size > kMaxBlockSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   171
    size = kMaxBlockSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   172
  if (size > remain)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   173
    size = (UInt32)remain;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   174
  for (UInt32 i = 0; i < size; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   175
    ((Byte *)data)[i] = Data[Pos + i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   176
  Pos += size;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   177
  if(processedSize != NULL)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   178
    *processedSize = size;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   179
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   180
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   181
  
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   182
class CBenchmarkOutStream: 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   183
  public ISequentialOutStream,
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   184
  public CBenchBuffer,
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   185
  public CMyUnknownImp
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   186
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   187
  // bool _overflow;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   188
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   189
  UInt32 Pos;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   190
  // CBenchmarkOutStream(): _overflow(false) {} 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   191
  void Init() 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   192
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   193
    // _overflow = false;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   194
    Pos = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   195
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   196
  MY_UNKNOWN_IMP
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   197
  STDMETHOD(Write)(const void *data, UInt32 size, UInt32 *processedSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   198
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   199
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   200
STDMETHODIMP CBenchmarkOutStream::Write(const void *data, UInt32 size, UInt32 *processedSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   201
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   202
  size_t curSize = BufferSize - Pos;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   203
  if (curSize > size)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   204
    curSize = size;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   205
  memcpy(Buffer + Pos, data, curSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   206
  Pos += (UInt32)curSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   207
  if(processedSize != NULL)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   208
    *processedSize = (UInt32)curSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   209
  if (curSize != size)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   210
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   211
    // _overflow = true;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   212
    return E_FAIL;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   213
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   214
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   215
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   216
  
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   217
class CCrcOutStream: 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   218
  public ISequentialOutStream,
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   219
  public CMyUnknownImp
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   220
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   221
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   222
  UInt32 Crc;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   223
  MY_UNKNOWN_IMP
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   224
  void Init() { Crc = CRC_INIT_VAL; }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   225
  STDMETHOD(Write)(const void *data, UInt32 size, UInt32 *processedSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   226
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   227
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   228
STDMETHODIMP CCrcOutStream::Write(const void *data, UInt32 size, UInt32 *processedSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   229
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   230
  Crc = CrcUpdate(Crc, data, size);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   231
  if (processedSize != NULL)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   232
    *processedSize = size;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   233
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   234
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   235
  
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   236
static UInt64 GetTimeCount()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   237
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   238
  #ifdef USE_POSIX_TIME
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   239
  #ifdef USE_POSIX_TIME2
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   240
  timeval v;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   241
  if (gettimeofday(&v, 0) == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   242
    return (UInt64)(v.tv_sec) * 1000000 + v.tv_usec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   243
  return (UInt64)time(NULL) * 1000000;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   244
  #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   245
  return time(NULL);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   246
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   247
  #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   248
  /*
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   249
  LARGE_INTEGER value;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   250
  if (::QueryPerformanceCounter(&value))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   251
    return value.QuadPart;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   252
  */
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   253
  return GetTickCount();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   254
  #endif 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   255
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   256
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   257
static UInt64 GetFreq()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   258
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   259
  #ifdef USE_POSIX_TIME
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   260
  #ifdef USE_POSIX_TIME2
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   261
  return 1000000;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   262
  #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   263
  return 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   264
  #endif 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   265
  #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   266
  /*
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   267
  LARGE_INTEGER value;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   268
  if (::QueryPerformanceFrequency(&value))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   269
    return value.QuadPart;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   270
  */
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   271
  return 1000;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   272
  #endif 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   273
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   274
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   275
#ifndef USE_POSIX_TIME
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   276
static inline UInt64 GetTime64(const FILETIME &t) { return ((UInt64)t.dwHighDateTime << 32) | t.dwLowDateTime; }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   277
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   278
static UInt64 GetUserTime()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   279
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   280
  #ifdef USE_POSIX_TIME
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   281
  return clock();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   282
  #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   283
  FILETIME creationTime, exitTime, kernelTime, userTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   284
  if (::GetProcessTimes(::GetCurrentProcess(), &creationTime, &exitTime, &kernelTime, &userTime) != 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   285
    return GetTime64(userTime) + GetTime64(kernelTime);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   286
  return (UInt64)GetTickCount() * 10000;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   287
  #endif 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   288
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   289
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   290
static UInt64 GetUserFreq()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   291
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   292
  #ifdef USE_POSIX_TIME
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   293
  return CLOCKS_PER_SEC;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   294
  #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   295
  return 10000000;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   296
  #endif 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   297
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   298
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   299
class CBenchProgressStatus
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   300
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   301
  #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   302
  NWindows::NSynchronization::CCriticalSection CS;  
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   303
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   304
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   305
  HRESULT Res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   306
  bool EncodeMode;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   307
  void SetResult(HRESULT res) 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   308
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   309
    #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   310
    NWindows::NSynchronization::CCriticalSectionLock lock(CS);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   311
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   312
    Res = res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   313
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   314
  HRESULT GetResult()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   315
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   316
    #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   317
    NWindows::NSynchronization::CCriticalSectionLock lock(CS);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   318
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   319
    return Res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   320
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   321
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   322
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   323
class CBenchProgressInfo:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   324
  public ICompressProgressInfo,
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   325
  public CMyUnknownImp
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   326
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   327
public:
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   328
  CBenchProgressStatus *Status;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   329
  CBenchInfo BenchInfo;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   330
  HRESULT Res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   331
  IBenchCallback *callback;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   332
  CBenchProgressInfo(): callback(0) {}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   333
  MY_UNKNOWN_IMP
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   334
  STDMETHOD(SetRatioInfo)(const UInt64 *inSize, const UInt64 *outSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   335
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   336
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   337
void SetStartTime(CBenchInfo &bi)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   338
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   339
  bi.GlobalFreq = GetFreq();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   340
  bi.UserFreq = GetUserFreq();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   341
  bi.GlobalTime = ::GetTimeCount();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   342
  bi.UserTime = ::GetUserTime();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   343
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   344
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   345
void SetFinishTime(const CBenchInfo &biStart, CBenchInfo &dest)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   346
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   347
  dest.GlobalFreq = GetFreq();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   348
  dest.UserFreq = GetUserFreq();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   349
  dest.GlobalTime = ::GetTimeCount() - biStart.GlobalTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   350
  dest.UserTime = ::GetUserTime() - biStart.UserTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   351
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   352
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   353
STDMETHODIMP CBenchProgressInfo::SetRatioInfo(const UInt64 *inSize, const UInt64 *outSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   354
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   355
  HRESULT res = Status->GetResult();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   356
  if (res != S_OK)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   357
    return res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   358
  if (!callback)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   359
    return res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   360
  CBenchInfo info = BenchInfo;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   361
  SetFinishTime(BenchInfo, info);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   362
  if (Status->EncodeMode)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   363
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   364
    info.UnpackSize = *inSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   365
    info.PackSize = *outSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   366
    res = callback->SetEncodeResult(info, false);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   367
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   368
  else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   369
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   370
    info.PackSize = BenchInfo.PackSize + *inSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   371
    info.UnpackSize = BenchInfo.UnpackSize + *outSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   372
    res = callback->SetDecodeResult(info, false);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   373
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   374
  if (res != S_OK)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   375
    Status->SetResult(res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   376
  return res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   377
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   378
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   379
static const int kSubBits = 8;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   380
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   381
static UInt32 GetLogSize(UInt32 size)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   382
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   383
  for (int i = kSubBits; i < 32; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   384
    for (UInt32 j = 0; j < (1 << kSubBits); j++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   385
      if (size <= (((UInt32)1) << i) + (j << (i - kSubBits)))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   386
        return (i << kSubBits) + j;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   387
  return (32 << kSubBits);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   388
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   389
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   390
static void NormalizeVals(UInt64 &v1, UInt64 &v2)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   391
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   392
  while (v1 > 1000000)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   393
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   394
    v1 >>= 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   395
    v2 >>= 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   396
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   397
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   398
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   399
UInt64 GetUsage(const CBenchInfo &info)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   400
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   401
  UInt64 userTime = info.UserTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   402
  UInt64 userFreq = info.UserFreq;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   403
  UInt64 globalTime = info.GlobalTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   404
  UInt64 globalFreq = info.GlobalFreq;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   405
  NormalizeVals(userTime, userFreq);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   406
  NormalizeVals(globalFreq, globalTime);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   407
  if (userFreq == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   408
    userFreq = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   409
  if (globalTime == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   410
    globalTime = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   411
  return userTime * globalFreq * 1000000 / userFreq / globalTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   412
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   413
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   414
UInt64 GetRatingPerUsage(const CBenchInfo &info, UInt64 rating)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   415
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   416
  UInt64 userTime = info.UserTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   417
  UInt64 userFreq = info.UserFreq;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   418
  UInt64 globalTime = info.GlobalTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   419
  UInt64 globalFreq = info.GlobalFreq;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   420
  NormalizeVals(userFreq, userTime);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   421
  NormalizeVals(globalTime, globalFreq);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   422
  if (globalFreq == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   423
    globalFreq = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   424
  if (userTime == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   425
    userTime = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   426
  return userFreq * globalTime / globalFreq *  rating / userTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   427
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   428
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   429
static UInt64 MyMultDiv64(UInt64 value, UInt64 elapsedTime, UInt64 freq)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   430
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   431
  UInt64 elTime = elapsedTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   432
  NormalizeVals(freq, elTime);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   433
  if (elTime == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   434
    elTime = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   435
  return value * freq / elTime;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   436
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   437
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   438
UInt64 GetCompressRating(UInt32 dictionarySize, UInt64 elapsedTime, UInt64 freq, UInt64 size)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   439
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   440
  UInt64 t = GetLogSize(dictionarySize) - (kBenchMinDicLogSize << kSubBits);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   441
  // UInt64 numCommandsForOne = 1000 + ((t * t * 7) >> (2 * kSubBits)); // AMD K8
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   442
  UInt64 numCommandsForOne = 870 + ((t * t * 5) >> (2 * kSubBits)); // Intel Core2
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   443
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   444
  UInt64 numCommands = (UInt64)(size) * numCommandsForOne;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   445
  return MyMultDiv64(numCommands, elapsedTime, freq);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   446
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   447
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   448
UInt64 GetDecompressRating(UInt64 elapsedTime, UInt64 freq, UInt64 outSize, UInt64 inSize, UInt32 numIterations)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   449
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   450
  // UInt64 numCommands = (inSize * 216 + outSize * 14) * numIterations; // AMD K8
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   451
  UInt64 numCommands = (inSize * 220 + outSize * 8) * numIterations; // Intel Core2
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   452
  return MyMultDiv64(numCommands, elapsedTime, freq);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   453
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   454
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   455
#ifdef EXTERNAL_LZMA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   456
typedef UInt32 (WINAPI * CreateObjectPointer)(const GUID *clsID, 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   457
    const GUID *interfaceID, void **outObject);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   458
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   459
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   460
struct CEncoderInfo;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   461
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   462
struct CEncoderInfo
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   463
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   464
  #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   465
  NWindows::CThread thread[2];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   466
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   467
  CMyComPtr<ICompressCoder> encoder;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   468
  CBenchProgressInfo *progressInfoSpec[2];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   469
  CMyComPtr<ICompressProgressInfo> progressInfo[2];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   470
  UInt32 NumIterations;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   471
  #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   472
  size_t AllocaSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   473
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   474
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   475
  struct CDecoderInfo
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   476
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   477
    CEncoderInfo *Encoder;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   478
    UInt32 DecoderIndex;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   479
    #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   480
    size_t AllocaSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   481
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   482
    bool CallbackMode;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   483
  };
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   484
  CDecoderInfo decodersInfo[2];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   485
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   486
  CMyComPtr<ICompressCoder> decoders[2];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   487
  HRESULT Results[2];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   488
  CBenchmarkOutStream *outStreamSpec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   489
  CMyComPtr<ISequentialOutStream> outStream;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   490
  IBenchCallback *callback;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   491
  UInt32 crc;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   492
  UInt32 kBufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   493
  UInt32 compressedSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   494
  CBenchRandomGenerator rg;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   495
  CBenchmarkOutStream *propStreamSpec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   496
  CMyComPtr<ISequentialOutStream> propStream;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   497
  HRESULT Init(UInt32 dictionarySize, UInt32 numThreads, CBaseRandomGenerator *rg);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   498
  HRESULT Encode();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   499
  HRESULT Decode(UInt32 decoderIndex);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   500
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   501
  CEncoderInfo(): outStreamSpec(0), callback(0), propStreamSpec(0) {}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   502
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   503
  #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   504
  static THREAD_FUNC_DECL EncodeThreadFunction(void *param)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   505
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   506
    CEncoderInfo *encoder = (CEncoderInfo *)param;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   507
    #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   508
    alloca(encoder->AllocaSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   509
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   510
    HRESULT res = encoder->Encode();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   511
    encoder->Results[0] = res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   512
    if (res != S_OK)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   513
      encoder->progressInfoSpec[0]->Status->SetResult(res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   514
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   515
    return 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   516
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   517
  static THREAD_FUNC_DECL DecodeThreadFunction(void *param)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   518
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   519
    CDecoderInfo *decoder = (CDecoderInfo *)param;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   520
    #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   521
    alloca(decoder->AllocaSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   522
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   523
    CEncoderInfo *encoder = decoder->Encoder;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   524
    encoder->Results[decoder->DecoderIndex] = encoder->Decode(decoder->DecoderIndex);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   525
    return 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   526
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   527
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   528
  HRESULT CreateEncoderThread()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   529
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   530
    return thread[0].Create(EncodeThreadFunction, this);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   531
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   532
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   533
  HRESULT CreateDecoderThread(int index, bool callbackMode
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   534
      #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   535
      , size_t allocaSize
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   536
      #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   537
      )
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   538
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   539
    CDecoderInfo &decoder = decodersInfo[index];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   540
    decoder.DecoderIndex = index;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   541
    decoder.Encoder = this;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   542
    #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   543
    decoder.AllocaSize = allocaSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   544
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   545
    decoder.CallbackMode = callbackMode;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   546
    return thread[index].Create(DecodeThreadFunction, &decoder);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   547
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   548
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   549
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   550
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   551
HRESULT CEncoderInfo::Init(UInt32 dictionarySize, UInt32 numThreads, CBaseRandomGenerator *rgLoc)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   552
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   553
  rg.Set(rgLoc);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   554
  kBufferSize = dictionarySize + kAdditionalSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   555
  UInt32 kCompressedBufferSize = (kBufferSize / 2) + kCompressedAdditionalSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   556
  if (!rg.Alloc(kBufferSize))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   557
    return E_OUTOFMEMORY;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   558
  rg.Generate();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   559
  crc = CrcCalc(rg.Buffer, rg.BufferSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   560
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   561
  outStreamSpec = new CBenchmarkOutStream;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   562
  if (!outStreamSpec->Alloc(kCompressedBufferSize))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   563
    return E_OUTOFMEMORY;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   564
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   565
  outStream = outStreamSpec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   566
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   567
  propStreamSpec = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   568
  if (!propStream)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   569
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   570
    propStreamSpec = new CBenchmarkOutStream;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   571
    propStream = propStreamSpec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   572
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   573
  if (!propStreamSpec->Alloc(kMaxLzmaPropSize))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   574
    return E_OUTOFMEMORY;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   575
  propStreamSpec->Init();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   576
  
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   577
  PROPID propIDs[] = 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   578
  { 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   579
    NCoderPropID::kDictionarySize, 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   580
    NCoderPropID::kMultiThread
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   581
  };
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   582
  const int kNumProps = sizeof(propIDs) / sizeof(propIDs[0]);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   583
  PROPVARIANT properties[kNumProps];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   584
  properties[0].vt = VT_UI4;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   585
  properties[0].ulVal = (UInt32)dictionarySize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   586
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   587
  properties[1].vt = VT_BOOL;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   588
  properties[1].boolVal = (numThreads > 1) ? VARIANT_TRUE : VARIANT_FALSE;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   589
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   590
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   591
    CMyComPtr<ICompressSetCoderProperties> setCoderProperties;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   592
    RINOK(encoder.QueryInterface(IID_ICompressSetCoderProperties, &setCoderProperties));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   593
    if (!setCoderProperties)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   594
      return E_FAIL;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   595
    RINOK(setCoderProperties->SetCoderProperties(propIDs, properties, kNumProps));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   596
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   597
    CMyComPtr<ICompressWriteCoderProperties> writeCoderProperties;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   598
    encoder.QueryInterface(IID_ICompressWriteCoderProperties, &writeCoderProperties);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   599
    if (writeCoderProperties)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   600
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   601
      RINOK(writeCoderProperties->WriteCoderProperties(propStream));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   602
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   603
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   604
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   605
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   606
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   607
HRESULT CEncoderInfo::Encode()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   608
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   609
  CBenchmarkInStream *inStreamSpec = new CBenchmarkInStream;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   610
  CMyComPtr<ISequentialInStream> inStream = inStreamSpec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   611
  inStreamSpec->Init(rg.Buffer, rg.BufferSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   612
  outStreamSpec->Init();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   613
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   614
  RINOK(encoder->Code(inStream, outStream, 0, 0, progressInfo[0]));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   615
  compressedSize = outStreamSpec->Pos;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   616
  encoder.Release();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   617
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   618
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   619
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   620
HRESULT CEncoderInfo::Decode(UInt32 decoderIndex)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   621
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   622
  CBenchmarkInStream *inStreamSpec = new CBenchmarkInStream;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   623
  CMyComPtr<ISequentialInStream> inStream = inStreamSpec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   624
  CMyComPtr<ICompressCoder> &decoder = decoders[decoderIndex];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   625
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   626
  CMyComPtr<ICompressSetDecoderProperties2> compressSetDecoderProperties;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   627
  decoder.QueryInterface(IID_ICompressSetDecoderProperties2, &compressSetDecoderProperties);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   628
  if (!compressSetDecoderProperties)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   629
    return E_FAIL;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   630
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   631
  CCrcOutStream *crcOutStreamSpec = new CCrcOutStream;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   632
  CMyComPtr<ISequentialOutStream> crcOutStream = crcOutStreamSpec;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   633
    
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   634
  CBenchProgressInfo *pi = progressInfoSpec[decoderIndex];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   635
  pi->BenchInfo.UnpackSize = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   636
  pi->BenchInfo.PackSize = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   637
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   638
  for (UInt32 j = 0; j < NumIterations; j++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   639
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   640
    inStreamSpec->Init(outStreamSpec->Buffer, compressedSize);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   641
    crcOutStreamSpec->Init();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   642
    
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   643
    RINOK(compressSetDecoderProperties->SetDecoderProperties2(propStreamSpec->Buffer, propStreamSpec->Pos));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   644
    UInt64 outSize = kBufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   645
    RINOK(decoder->Code(inStream, crcOutStream, 0, &outSize, progressInfo[decoderIndex]));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   646
    if (CRC_GET_DIGEST(crcOutStreamSpec->Crc) != crc)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   647
      return S_FALSE;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   648
    pi->BenchInfo.UnpackSize += kBufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   649
    pi->BenchInfo.PackSize += compressedSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   650
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   651
  decoder.Release();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   652
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   653
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   654
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   655
static const UInt32 kNumThreadsMax = (1 << 16);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   656
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   657
struct CBenchEncoders
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   658
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   659
  CEncoderInfo *encoders;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   660
  CBenchEncoders(UInt32 num): encoders(0) { encoders = new CEncoderInfo[num]; }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   661
  ~CBenchEncoders() { delete []encoders; }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   662
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   663
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   664
HRESULT LzmaBench(
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   665
  #ifdef EXTERNAL_LZMA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   666
  CCodecs *codecs,
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   667
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   668
  UInt32 numThreads, UInt32 dictionarySize, IBenchCallback *callback)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   669
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   670
  UInt32 numEncoderThreads = 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   671
    #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   672
    (numThreads > 1 ? numThreads / 2 : 1);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   673
    #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   674
    1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   675
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   676
  UInt32 numSubDecoderThreads = 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   677
    #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   678
    (numThreads > 1 ? 2 : 1);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   679
    #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   680
    1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   681
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   682
  if (dictionarySize < (1 << kBenchMinDicLogSize) || numThreads < 1 || numEncoderThreads > kNumThreadsMax)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   683
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   684
    return E_INVALIDARG;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   685
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   686
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   687
  CBenchEncoders encodersSpec(numEncoderThreads);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   688
  CEncoderInfo *encoders = encodersSpec.encoders;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   689
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   690
  #ifdef EXTERNAL_LZMA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   691
  UString name = L"LZMA";
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   692
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   693
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   694
  UInt32 i;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   695
  for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   696
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   697
    CEncoderInfo &encoder = encoders[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   698
    encoder.callback = (i == 0) ? callback : 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   699
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   700
    #ifdef EXTERNAL_LZMA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   701
    RINOK(codecs->CreateCoder(name, true, encoder.encoder));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   702
    #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   703
    encoder.encoder = new NCompress::NLZMA::CEncoder;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   704
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   705
    for (UInt32 j = 0; j < numSubDecoderThreads; j++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   706
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   707
      #ifdef EXTERNAL_LZMA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   708
      RINOK(codecs->CreateCoder(name, false, encoder.decoders[j]));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   709
      #else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   710
      encoder.decoders[j] = new NCompress::NLZMA::CDecoder;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   711
      #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   712
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   713
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   714
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   715
  CBaseRandomGenerator rg;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   716
  rg.Init();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   717
  for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   718
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   719
    RINOK(encoders[i].Init(dictionarySize, numThreads, &rg));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   720
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   721
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   722
  CBenchProgressStatus status;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   723
  status.Res = S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   724
  status.EncodeMode = true;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   725
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   726
  for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   727
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   728
    CEncoderInfo &encoder = encoders[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   729
    for (int j = 0; j < 2; j++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   730
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   731
      encoder.progressInfo[j] = encoder.progressInfoSpec[j] = new CBenchProgressInfo;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   732
      encoder.progressInfoSpec[j]->Status = &status;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   733
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   734
    if (i == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   735
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   736
      encoder.progressInfoSpec[0]->callback = callback;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   737
      encoder.progressInfoSpec[0]->BenchInfo.NumIterations = numEncoderThreads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   738
      SetStartTime(encoder.progressInfoSpec[0]->BenchInfo);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   739
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   740
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   741
    #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   742
    if (numEncoderThreads > 1)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   743
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   744
      #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   745
      encoder.AllocaSize = (i * 16 * 21) & 0x7FF;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   746
      #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   747
      RINOK(encoder.CreateEncoderThread())
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   748
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   749
    else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   750
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   751
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   752
      RINOK(encoder.Encode());
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   753
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   754
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   755
  #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   756
  if (numEncoderThreads > 1)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   757
    for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   758
      encoders[i].thread[0].Wait();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   759
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   760
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   761
  RINOK(status.Res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   762
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   763
  CBenchInfo info;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   764
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   765
  SetFinishTime(encoders[0].progressInfoSpec[0]->BenchInfo, info);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   766
  info.UnpackSize = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   767
  info.PackSize = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   768
  info.NumIterations = 1; // progressInfoSpec->NumIterations;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   769
  for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   770
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   771
    CEncoderInfo &encoder = encoders[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   772
    info.UnpackSize += encoder.kBufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   773
    info.PackSize += encoder.compressedSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   774
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   775
  RINOK(callback->SetEncodeResult(info, true));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   776
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   777
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   778
  status.Res = S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   779
  status.EncodeMode = false;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   780
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   781
  UInt32 numDecoderThreads = numEncoderThreads * numSubDecoderThreads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   782
  for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   783
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   784
    CEncoderInfo &encoder = encoders[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   785
    encoder.NumIterations = 2 + kUncompressMinBlockSize / encoder.kBufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   786
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   787
    if (i == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   788
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   789
      encoder.progressInfoSpec[0]->callback = callback;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   790
      encoder.progressInfoSpec[0]->BenchInfo.NumIterations = numDecoderThreads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   791
      SetStartTime(encoder.progressInfoSpec[0]->BenchInfo);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   792
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   793
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   794
    #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   795
    if (numDecoderThreads > 1)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   796
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   797
      for (UInt32 j = 0; j < numSubDecoderThreads; j++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   798
      {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   799
        size_t allocaSize = ((i * numSubDecoderThreads + j) * 16 * 21) & 0x7FF;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   800
        HRESULT res = encoder.CreateDecoderThread(j, (i == 0 && j == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   801
            #ifdef USE_ALLOCA
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   802
            , allocaSize
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   803
            #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   804
            );
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   805
        RINOK(res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   806
      }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   807
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   808
    else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   809
    #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   810
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   811
      RINOK(encoder.Decode(0));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   812
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   813
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   814
  #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   815
  HRESULT res = S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   816
  if (numDecoderThreads > 1)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   817
    for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   818
      for (UInt32 j = 0; j < numSubDecoderThreads; j++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   819
      {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   820
        CEncoderInfo &encoder = encoders[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   821
        encoder.thread[j].Wait();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   822
        if (encoder.Results[j] != S_OK)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   823
          res = encoder.Results[j];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   824
      }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   825
  RINOK(res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   826
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   827
  RINOK(status.Res);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   828
  SetFinishTime(encoders[0].progressInfoSpec[0]->BenchInfo, info);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   829
  info.UnpackSize = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   830
  info.PackSize = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   831
  info.NumIterations = numSubDecoderThreads * encoders[0].NumIterations;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   832
  for (i = 0; i < numEncoderThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   833
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   834
    CEncoderInfo &encoder = encoders[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   835
    info.UnpackSize += encoder.kBufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   836
    info.PackSize += encoder.compressedSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   837
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   838
  RINOK(callback->SetDecodeResult(info, false));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   839
  RINOK(callback->SetDecodeResult(info, true));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   840
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   841
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   842
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   843
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   844
inline UInt64 GetLZMAUsage(bool multiThread, UInt32 dictionary)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   845
{ 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   846
  UInt32 hs = dictionary - 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   847
  hs |= (hs >> 1);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   848
  hs |= (hs >> 2);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   849
  hs |= (hs >> 4);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   850
  hs |= (hs >> 8);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   851
  hs >>= 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   852
  hs |= 0xFFFF;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   853
  if (hs > (1 << 24))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   854
    hs >>= 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   855
  hs++;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   856
  return ((hs + (1 << 16)) + (UInt64)dictionary * 2) * 4 + (UInt64)dictionary * 3 / 2 + 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   857
      (1 << 20) + (multiThread ? (6 << 20) : 0);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   858
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   859
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   860
UInt64 GetBenchMemoryUsage(UInt32 numThreads, UInt32 dictionary)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   861
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   862
  const UInt32 kBufferSize = dictionary;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   863
  const UInt32 kCompressedBufferSize = (kBufferSize / 2);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   864
  UInt32 numSubThreads = (numThreads > 1) ? 2 : 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   865
  UInt32 numBigThreads = numThreads / numSubThreads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   866
  return (kBufferSize + kCompressedBufferSize +
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   867
    GetLZMAUsage((numThreads > 1), dictionary) + (2 << 20)) * numBigThreads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   868
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   869
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   870
static bool CrcBig(const void *data, UInt32 size, UInt32 numCycles, UInt32 crcBase)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   871
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   872
  for (UInt32 i = 0; i < numCycles; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   873
    if (CrcCalc(data, size) != crcBase)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   874
      return false;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   875
  return true;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   876
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   877
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   878
#ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   879
struct CCrcInfo
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   880
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   881
  NWindows::CThread Thread;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   882
  const Byte *Data;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   883
  UInt32 Size;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   884
  UInt32 NumCycles;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   885
  UInt32 Crc;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   886
  bool Res;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   887
  void Wait()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   888
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   889
    Thread.Wait();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   890
    Thread.Close();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   891
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   892
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   893
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   894
static THREAD_FUNC_DECL CrcThreadFunction(void *param)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   895
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   896
  CCrcInfo *p = (CCrcInfo *)param;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   897
  p->Res = CrcBig(p->Data, p->Size, p->NumCycles, p->Crc);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   898
  return 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   899
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   900
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   901
struct CCrcThreads
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   902
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   903
  UInt32 NumThreads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   904
  CCrcInfo *Items;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   905
  CCrcThreads(): Items(0), NumThreads(0) {}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   906
  void WaitAll()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   907
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   908
    for (UInt32 i = 0; i < NumThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   909
      Items[i].Wait();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   910
    NumThreads = 0;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   911
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   912
  ~CCrcThreads() 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   913
  { 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   914
    WaitAll();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   915
    delete []Items; 
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   916
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   917
};
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   918
#endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   919
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   920
static UInt32 CrcCalc1(const Byte *buf, UInt32 size)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   921
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   922
  UInt32 crc = CRC_INIT_VAL;;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   923
  for (UInt32 i = 0; i < size; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   924
    crc = CRC_UPDATE_BYTE(crc, buf[i]);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   925
  return CRC_GET_DIGEST(crc);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   926
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   927
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   928
static void RandGen(Byte *buf, UInt32 size, CBaseRandomGenerator &RG)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   929
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   930
  for (UInt32 i = 0; i < size; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   931
    buf[i] = (Byte)RG.GetRnd();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   932
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   933
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   934
static UInt32 RandGenCrc(Byte *buf, UInt32 size, CBaseRandomGenerator &RG)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   935
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   936
  RandGen(buf, size, RG);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   937
  return CrcCalc1(buf, size);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   938
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   939
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   940
bool CrcInternalTest()
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   941
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   942
  CBenchBuffer buffer;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   943
  const UInt32 kBufferSize0 = (1 << 8);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   944
  const UInt32 kBufferSize1 = (1 << 10);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   945
  const UInt32 kCheckSize = (1 << 5);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   946
  if (!buffer.Alloc(kBufferSize0 + kBufferSize1))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   947
    return false;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   948
  Byte *buf = buffer.Buffer;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   949
  UInt32 i;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   950
  for (i = 0; i < kBufferSize0; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   951
    buf[i] = (Byte)i;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   952
  UInt32 crc1 = CrcCalc1(buf, kBufferSize0);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   953
  if (crc1 != 0x29058C73)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   954
    return false;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   955
  CBaseRandomGenerator RG;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   956
  RandGen(buf + kBufferSize0, kBufferSize1, RG);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   957
  for (i = 0; i < kBufferSize0 + kBufferSize1 - kCheckSize; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   958
    for (UInt32 j = 0; j < kCheckSize; j++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   959
      if (CrcCalc1(buf + i, j) != CrcCalc(buf + i, j))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   960
        return false;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   961
  return true;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   962
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   963
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   964
HRESULT CrcBench(UInt32 numThreads, UInt32 bufferSize, UInt64 &speed)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   965
{
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   966
  if (numThreads == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   967
    numThreads = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   968
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   969
  CBenchBuffer buffer;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   970
  size_t totalSize = (size_t)bufferSize * numThreads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   971
  if (totalSize / numThreads != bufferSize)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   972
    return E_OUTOFMEMORY;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   973
  if (!buffer.Alloc(totalSize))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   974
    return E_OUTOFMEMORY;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   975
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   976
  Byte *buf = buffer.Buffer;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   977
  CBaseRandomGenerator RG;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   978
  UInt32 numCycles = ((UInt32)1 << 30) / ((bufferSize >> 2) + 1) + 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   979
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   980
  UInt64 timeVal;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   981
  #ifdef BENCH_MT
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   982
  CCrcThreads threads;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   983
  if (numThreads > 1)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   984
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   985
    threads.Items = new CCrcInfo[numThreads];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   986
    UInt32 i;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   987
    for (i = 0; i < numThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   988
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   989
      CCrcInfo &info = threads.Items[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   990
      Byte *data = buf + (size_t)bufferSize * i;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   991
      info.Data = data;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   992
      info.NumCycles = numCycles;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   993
      info.Size = bufferSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   994
      info.Crc = RandGenCrc(data, bufferSize, RG);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   995
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   996
    timeVal = GetTimeCount();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   997
    for (i = 0; i < numThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   998
    {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
   999
      CCrcInfo &info = threads.Items[i];
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1000
      RINOK(info.Thread.Create(CrcThreadFunction, &info));
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1001
      threads.NumThreads++;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1002
    }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1003
    threads.WaitAll();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1004
    for (i = 0; i < numThreads; i++)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1005
      if (!threads.Items[i].Res)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1006
        return S_FALSE;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1007
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1008
  else
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1009
  #endif
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1010
  {
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1011
    UInt32 crc = RandGenCrc(buf, bufferSize, RG);
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1012
    timeVal = GetTimeCount();
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1013
    if (!CrcBig(buf, bufferSize, numCycles, crc))
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1014
      return S_FALSE;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1015
  }
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1016
  timeVal = GetTimeCount() - timeVal;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1017
  if (timeVal == 0)
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1018
    timeVal = 1;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1019
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1020
  UInt64 size = (UInt64)numCycles * totalSize;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1021
  speed = MyMultDiv64(size, timeVal, GetFreq());
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1022
  return S_OK;
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1023
}
bb5522e88ab2 bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is excessive or whether libphysfs should even be maintained by us is another matter. But at least we shouldn't crash
nemo
parents:
diff changeset
  1024