Commit Graph

1794 Commits

Author SHA1 Message Date
Kouhei Sutou
9c457ccb7a Add missing $(EXT) 2018-08-13 14:17:54 +09:00
Yann Collet
b3692db46d
Merge pull request #531 from lz4/dev
Preparing v1.8.2
2018-05-07 14:38:45 -07:00
Yann Collet
bf6fd938e5
Merge pull request #537 from lz4/xpHCmf2
Speed optimization for optimal parser
2018-05-07 13:06:43 -07:00
Yann Collet
ba1c7148a5 renamed variable for clarity 2018-05-07 12:14:26 -07:00
Yann Collet
abb1f70e17
Merge pull request #538 from lz4/frameTestError
Fix frametest error
2018-05-07 11:33:53 -07:00
Yann Collet
200b2960d5 fixed minor conversion warning 2018-05-06 18:26:14 -07:00
Yann Collet
24b9c485db small PA optimization
which measurably improves speed
on levels 9+
2018-05-06 16:53:33 -07:00
Yann Collet
d7b6c726ed small extDict : fixed side-effect
don't fix dictionaries of size 0.
setting dictEnd == source triggers prefix mode,
thus removing possibility to use CDict.
2018-05-05 19:59:00 -07:00
Yann Collet
af12733467 fixed frametest error
The error can be reproduced using following command :
./frametest -v -i100000000 -s1659 -t31096808

It's actually a bug in the stream LZ4 API,
when starting a new stream
and providing a first chunk to complete with size < MINMATCH.
In which case, the chunk becomes a dictionary.
No hash was generated and stored,
but the chunk is accessible as default position 0 points to dictStart,
and position 0 is still within MAX_DISTANCE.
Then, next attempt to read 32-bits from position 0 fails.

The issue would have been mitigated by starting from index 64 KB,
effectively eliminating position 0 as too far away.

The proper fix is to eliminate such "dictionary" as too small.
Which is what this patch does.
2018-05-05 18:24:11 -07:00
Yann Collet
cdb0275b7f lz4hc: fixed PA / SC parameter order
also :
reserved PA for levels 9+ (instead of 8+).
In most cases, speed is lower, and compression benefit is not worth.
2018-05-05 14:32:57 -07:00
Yann Collet
a4e918d7a6 lz4hc: SC only enabled for opt parser
the trade off is not good for regular HC parser :
compression is a little bit better, but speed cost is too large in comparison.
2018-05-05 14:25:37 -07:00
Yann Collet
d097bf93f8 fixed SC.opt integration with regular HC parser
Only enabled when searching forward.

note : it slighly improves compression ratio,
but measurably decreases speed.
Trade-off to analyse.
2018-05-05 13:46:45 -07:00
Yann Collet
fa89a9e18b lz4hc: fixed performance issue
when combining both PA and CS optimizations
2018-05-05 13:31:03 -07:00
Yann Collet
9699ba5ddf integrated chain swapper into HC match finder
slower than expected
Pattern analyzer and Chain Swapper
work slower when both activated.
Reasons unclear.
2018-05-04 19:13:33 -07:00
Yann Collet
a39fa9c5fc
Merge pull request #536 from terrelln/make-install
Fix make install
2018-05-04 16:18:46 -07:00
Nick Terrell
3e7fa1b14d Attempt to fix travis 2018-05-04 14:33:59 -07:00
Nick Terrell
a7cc0b590a Fix make install
* Uninstall didn't remove the pkg-config correctly.
* Fix `mandir`
* Allow overriding either upper- or lower-case location variables, but
  always use the lower case variables.
* Add test case that ensures overriding both upper- and lower-case
  variables is the same, and that the directory is empty after uninstall.
2018-05-04 13:50:23 -07:00
Yann Collet
434ace7244 implemented search accelerator
greatly improves speed compared to non-accelerated,
especially for slower files.

On my laptop, -b12 :
```
calgary.tar :  4.3 MB/s =>  9.0 MB/s
enwik7      : 10.2 MB/s => 13.3 MB/s
silesia.tar :  4.0 MB/s =>  8.7 MB/s
```

Note : this is the simplified version,
without handling dictionaries, external buffer, nor pattern analyzer.
Current `dev` branch on these samples gives :
```
calgary.tar :  4.2 MB/s
enwik7      :  9.7 MB/s
silesia.tar :  3.5 MB/s
```

interestingly, it's slower,
presumably due to handling of dictionaries.
2018-05-03 16:31:41 -07:00
Yann Collet
d358e33faa Added CDict speed graph to be used for release statement 2018-05-03 16:01:24 -07:00
Yann Collet
e00ba49cde updated API documentation 2018-05-03 15:40:01 -07:00
Yann Collet
dc42707107 created LZ4HC_FindLongestMatch()
simplified match finder
only searching forward and within current buffer,
for easier testing of optimizations.
2018-05-03 15:38:32 -07:00
Yann Collet
f3e84ffd41
Merge pull request #529 from felixhandte/lz4f-fast-reset-for-streaming-only
LZ4F: Only Reset the LZ4_stream_t when Init'ing a Streaming Block
2018-05-03 15:37:51 -07:00
Yann Collet
2b6c4f3d63
Merge pull request #530 from lz4/lz4fRingBuffer
Random lz4f clarifications
2018-05-03 12:55:04 -07:00
Yann Collet
ffbff1f360 Merge branch 'dev' into lz4fRingBuffer 2018-05-03 11:54:57 -07:00
Yann Collet
95607a749b
Merge pull request #528 from lz4/complexShortcut
Faster decoding speed
2018-05-03 11:35:50 -07:00
Cyan4973
2e2c9f6ff3 fix comments / indentation
as requested by @terrelln
2018-05-03 07:56:33 -07:00
W. Felix Handte
5406c2e479 Only Reset the LZ4 Stream when Init'ing a Streaming Block 2018-05-03 00:03:20 -04:00
Yann Collet
1e130d23e3 updated NEWS in preparation for v1.8.2 2018-05-02 16:52:33 -07:00
Yann Collet
79af95a5b3 Merge branch 'lz4fRingBuffer' of github.com:Cyan4973/lz4 into lz4fRingBuffer 2018-05-02 16:43:51 -07:00
Yann Collet
ba168ee5c7 updated benchmark for v1.8.2 2018-05-02 16:43:05 -07:00
Yann Collet
543223851f updated benchmark for v1.8.2 2018-05-02 16:41:15 -07:00
Yann Collet
c25eb16666 random lz4f clarifications
the initial intention was to update lz4f ring buffer strategy,
but lz4f doesn't use ring buffer.
Instead, it uses the destination buffer as much as possible,
and merely copies just what's required to preserve history
into its own buffer, at the end.
Pretty efficient.

This patch just clarifies a few comments and add some assert().
It's built on top of #528.

It also updates doc.
2018-05-02 16:05:42 -07:00
Yann Collet
858d12e3e1 Merge branch 'dev' into lz4fRingBuffer 2018-05-02 14:24:24 -07:00
Yann Collet
85be6b8f6d increased nbAttempts for lz4 -12
shaves one more kilobyte from silesia.tar
2018-05-02 14:22:35 -07:00
Yann Collet
999a8488f6 removed test that might be optimized away
under UB rule "no overflow on int"
2018-05-02 13:57:33 -07:00
Yann Collet
93cf628a08 introduce LZ4_decoderRingBufferSize()
fuzzer : fix and robustify ring buffer tests
2018-05-02 13:01:04 -07:00
Yann Collet
1a191b3f8d simplify shortcut 2018-05-02 10:33:12 -07:00
Yann Collet
0114b63b40 Merge branch 'dev' into complexShortcut 2018-05-02 10:08:30 -07:00
Yann Collet
5fc7e0b336
Merge pull request #525 from lz4/testDecMerge
added a test case for LZ4_decompress_fast_usingDict  …
2018-05-02 10:06:07 -07:00
Yann Collet
bd470ccd38
Merge pull request #521 from lz4/BD_deterministic
fix lz4hc -BD non-determinism
2018-04-30 20:40:34 -07:00
Cyan4973
6a7d501fed renamed variable for clarity
lowLimit -> lowestMatchIndex
2018-04-30 18:56:16 -07:00
Cyan4973
1949bf11e3 added visual test dir to .gitignore 2018-04-30 18:50:56 -07:00
Yann Collet
8c574990a9 lz4hc changed variable
to reduce confusion
dictLowLimit => dictStart
2018-04-30 16:08:16 -07:00
Yann Collet
4c696613a0 clarified streaming decompression function
restrictions for ring buffer
2018-04-30 15:55:33 -07:00
Yann Collet
90374271c2
Merge pull request #527 from svpv/fastDec
lz4.c: two-stage shortcut for LZ4_decompress_generic
2018-04-30 15:32:37 -07:00
Yann Collet
c32e0319a5
Merge pull request #523 from svpv/makeV1
lib/Makefile: show commands with V=1
2018-04-29 08:56:44 -07:00
Cyan4973
724bbcd960 Merge branch 'dev' of github.com:lz4/lz4 into dev 2018-04-29 08:47:08 -07:00
Cyan4973
e28ae0af07 updated NEWS for v1.8.2
mentioning work from @svpv
2018-04-29 08:46:39 -07:00
Yann Collet
99e5376ef7
Merge pull request #526 from svpv/makeV1
lib/Makefile: show commands with V=1
2018-04-29 08:45:16 -07:00
Cyan4973
5a2501a90d added a test case for LZ4_decompress_fast_usingDict
with a separated dictionary
since a joined dictionary is now detected as prefix64K.

Also : fixed a minor warning under msys
2018-04-29 07:42:24 -07:00