| author | nemo |
| Mon, 10 Apr 2017 12:06:43 -0400 | |
| changeset 12213 | bb5522e88ab2 |
| permissions | -rw-r--r-- |
|
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 |
// Compress/RangeCoder/RangeCoderOpt.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
|
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 |
#ifndef __COMPRESS_RANGECODER_OPT_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 |
#define __COMPRESS_RANGECODER_OPT_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
|
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
|
6 |
#define RC_INIT_VAR \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
UInt32 range = rangeDecoder->Range; \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
UInt32 code = rangeDecoder->Code; |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
|
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
#define RC_FLUSH_VAR \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
rangeDecoder->Range = range; \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
rangeDecoder->Code = code; |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
|
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
#define RC_NORMALIZE \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
if (range < NCompress::NRangeCoder::kTopValue) \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
{ code = (code << 8) | rangeDecoder->Stream.ReadByte(); range <<= 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
|
17 |
|
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
#define RC_GETBIT2(numMoveBits, prob, mi, A0, 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
|
19 |
{ UInt32 bound = (range >> NCompress::NRangeCoder::kNumBitModelTotalBits) * prob; \
|
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
if (code < bound) \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
{ A0; range = bound; \
|
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
prob += (NCompress::NRangeCoder::kBitModelTotal - prob) >> numMoveBits; \ |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
mi <<= 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
|
24 |
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
|
25 |
{ A1; range -= bound; code -= bound; prob -= (prob) >> numMoveBits; \
|
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
mi = (mi + mi) + 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
|
27 |
RC_NORMALIZE |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
|
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
#define RC_GETBIT(numMoveBits, prob, mi) RC_GETBIT2(numMoveBits, prob, mi, ; , ;) |
|
bb5522e88ab2
bulk copy of latest physfs to our misc/libphysfs since this seems to fix an off-by-1 error reliably hit in readln read of 1 byte probably introduced in the addition of the buffered read. Whether this is 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 |
#endif |