6502/lzsa
1
0
Fork 0
mirror of https://github.com/emmanuel-marty/lzsa.git synced 2024-12-29 00:31:54 +00:00
Code Issues Projects Releases Wiki Activity

Small improvement to merging large matches

Browse Source
This commit is contained in:
Emmanuel Marty 2019-11-11 18:41:08 +01:00 committed by GitHub
parent df9690a949
commit 53b2013b73
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 57 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
src/shrink_block_v1.c
View File

@ -327,24 +327,33 @@ static int lzsa_optimize_command_count_v1(lzsa_compressor *pCompressor, const un
}
}
if ((i + pMatch->length) < nEndOffset && pMatch->length >= LCP_MAX &&
pBestMatch[i + pMatch->length].offset &&
if ((i + pMatch->length) < nEndOffset && pMatch->offset > 0 && pMatch->length >= MIN_MATCH_SIZE_V1 &&
pBestMatch[i + pMatch->length].offset > 0 &&
pBestMatch[i + pMatch->length].length >= MIN_MATCH_SIZE_V1 &&
pBestMatch[i + pMatch->length].offset <= pMatch->length &&
(pMatch->length + pBestMatch[i + pMatch->length].length) >= LEAVE_ALONE_MATCH_SIZE &&
(pMatch->length + pBestMatch[i + pMatch->length].length) <= MAX_VARLEN &&
(i + pMatch->length) > pMatch->offset &&
(i - pMatch->offset + pMatch->length + pBestMatch[i + pMatch->length].length) < nEndOffset &&
(i + pMatch->length) > pBestMatch[i + pMatch->length].offset &&
(i + pMatch->length + pBestMatch[i + pMatch->length].length) < nEndOffset &&
!memcmp(pInWindow + i - pMatch->offset + pMatch->length,
pInWindow + i + pMatch->length - pBestMatch[i + pMatch->length].offset,
pBestMatch[i + pMatch->length].length)) {
int nMatchLen = pMatch->length;
/* Join */
int nCurPartialSize = lzsa_get_match_varlen_size_v1(pMatch->length - MIN_MATCH_SIZE_V1);
nCurPartialSize += 8 /* token */ + lzsa_get_literals_varlen_size_v1(0) + ((pBestMatch[i + pMatch->length].offset <= 256) ? 8 : 16) /* match offset */ + lzsa_get_match_varlen_size_v1(pBestMatch[i + pMatch->length].length - MIN_MATCH_SIZE_V1);
pMatch->length += pBestMatch[i + nMatchLen].length;
pBestMatch[i + nMatchLen].offset = 0;
pBestMatch[i + nMatchLen].length = -1;
continue;
int nReducedPartialSize = lzsa_get_match_varlen_size_v1(pMatch->length + pBestMatch[i + pMatch->length].length - MIN_MATCH_SIZE_V1);
if (nCurPartialSize >= nReducedPartialSize) {
int nMatchLen = pMatch->length;
/* Join */
pMatch->length += pBestMatch[i + nMatchLen].length;
pBestMatch[i + nMatchLen].offset = 0;
pBestMatch[i + nMatchLen].length = -1;
continue;
}
}
i += pMatch->length;

48
src/shrink_block_v2.c
View File

@ -611,25 +611,51 @@ static int lzsa_optimize_command_count_v2(lzsa_compressor *pCompressor, const un
}
}
if ((i + pMatch->length) < nEndOffset && pMatch->length >= LCP_MAX &&
pBestMatch[i + pMatch->length].offset &&
if ((i + pMatch->length) < nEndOffset && pMatch->offset > 0 && pMatch->length >= MIN_MATCH_SIZE_V2 &&
pBestMatch[i + pMatch->length].offset > 0 &&
pBestMatch[i + pMatch->length].length >= MIN_MATCH_SIZE_V2 &&
pBestMatch[i + pMatch->length].offset <= pMatch->length &&
(pMatch->length + pBestMatch[i + pMatch->length].length) >= LEAVE_ALONE_MATCH_SIZE &&
(pMatch->length + pBestMatch[i + pMatch->length].length) <= MAX_VARLEN &&
(i + pMatch->length) > pMatch->offset &&
(i - pMatch->offset + pMatch->length + pBestMatch[i + pMatch->length].length) < nEndOffset &&
(i + pMatch->length) > pBestMatch[i + pMatch->length].offset &&
(i + pMatch->length + pBestMatch[i + pMatch->length].length) < nEndOffset &&
!memcmp(pInWindow + i - pMatch->offset + pMatch->length,
pInWindow + i + pMatch->length - pBestMatch[i + pMatch->length].offset,
pBestMatch[i + pMatch->length].length)) {
int nMatchLen = pMatch->length;
/* Join */
int nNextIndex = i + pMatch->length;
int nNextLiterals = 0;
pMatch->length += pBestMatch[i + nMatchLen].length;
pBestMatch[i + nMatchLen].offset = 0;
pBestMatch[i + nMatchLen].length = -1;
nDidReduce = 1;
continue;
while (nNextIndex < nEndOffset && pBestMatch[nNextIndex].length < MIN_MATCH_SIZE_V2) {
nNextLiterals++;
nNextIndex++;
}
int nCurPartialSize = lzsa_get_match_varlen_size_v2(pMatch->length - MIN_MATCH_SIZE_V2);
nCurPartialSize += 8 /* token */ + lzsa_get_literals_varlen_size_v2(0) + lzsa_get_match_varlen_size_v2(pBestMatch[i + pMatch->length].length - MIN_MATCH_SIZE_V2);
if (pBestMatch[i + pMatch->length].offset != pMatch->offset)
nCurPartialSize += (pBestMatch[i + pMatch->length].offset <= 32) ? 4 : ((pBestMatch[i + pMatch->length].offset <= 512) ? 8 : ((pBestMatch[i + pMatch->length].offset <= (8192 + 512)) ? 12 : 16));
if (pBestMatch[nNextIndex].offset != pBestMatch[i + pMatch->length].offset)
nCurPartialSize += (pBestMatch[nNextIndex].offset <= 32) ? 4 : ((pBestMatch[nNextIndex].offset <= 512) ? 8 : ((pBestMatch[nNextIndex].offset <= (8192 + 512)) ? 12 : 16));
int nReducedPartialSize = lzsa_get_match_varlen_size_v2(pMatch->length + pBestMatch[i + pMatch->length].length - MIN_MATCH_SIZE_V2);
if (pBestMatch[nNextIndex].offset != pMatch->offset)
nReducedPartialSize += (pBestMatch[nNextIndex].offset <= 32) ? 4 : ((pBestMatch[nNextIndex].offset <= 512) ? 8 : ((pBestMatch[nNextIndex].offset <= (8192 + 512)) ? 12 : 16));
if (nCurPartialSize >= nReducedPartialSize) {
int nMatchLen = pMatch->length;
/* Join */
pMatch->length += pBestMatch[i + nMatchLen].length;
pBestMatch[i + nMatchLen].offset = 0;
pBestMatch[i + nMatchLen].length = -1;
nDidReduce = 1;
continue;
}
}
nPrevRepMatchOffset = nRepMatchOffset;

2
src/shrink_context.h
View File

@ -41,7 +41,7 @@ extern "C" {
#define LCP_BITS 14
#define TAG_BITS 3
#define LCP_MAX (1U<<(LCP_BITS - TAG_BITS - 1))
#define LCP_MAX ((1U<<(LCP_BITS - TAG_BITS)) - 1)
#define LCP_AND_TAG_MAX (1U<<(LCP_BITS - 1))
#define LCP_SHIFT (31-LCP_BITS)
#define LCP_MASK (((1U<<LCP_BITS) - 1) << LCP_SHIFT)