2d9332eb21
Fix Types
2018-11-12 12:52:31 -08:00
4127de5fa6
Switch Enum to Only Non-Negative Values, Update Comments
2018-11-12 12:47:47 -08:00
596f7d1256
Fix #1412 : Perform Signed Comparison When Setting Attach Dict Param
2018-11-12 12:07:57 -08:00
7b0c551bff
Merge pull request #1411 from facebook/prefetch_dict
...
Improves decompression speed when using cold dictionary
2018-11-09 11:31:35 -08:00
1b4a9c518b
Merge pull request #1410 from facebook/prefetch_dec
...
improve long-range decoder speed
2018-11-08 18:41:58 -08:00
483759a3de
Improves decompression speed when using cold dictionary
...
by triggering the prefetching decoder path
(which used to be dedicated to long-range offsets only).
Figures on my laptop :
no content prefetch : ~300 MB/s (for reference)
full content prefetch : ~325 MB/s (before this patch)
new prefetch path : ~375 MB/s (after this patch)
The benchmark speed is already significant,
but another side-effect is that this version
prefetch less data into memory,
since it only prefetches what's needed, instead of the full dictionary.
This is supposed to help highly active environments
such as active databases,
that can't be properly measured in benchmark environment (too clean).
Also :
fixed the largeNbDict test program
which was working improperly when setting nbBlocks > nbFiles.
2018-11-08 17:00:23 -08:00
a801ba643c
Merge pull request #1409 from terrelln/tsan-dict
...
Add TSAN dictionary tests
2018-11-08 13:43:01 -08:00
20fb9e7f36
reduced assertion strength
...
one limit case can apparently be generated during fuzzer tests
2018-11-08 12:57:34 -08:00
f40c6b4395
added a simple test case for decode-only benchmark mode
...
as it was broken and no one noticed
2018-11-08 12:51:42 -08:00
9126da5b5c
improve long-range decoder speed
...
on enwik9 at level 22 (which is almost a worst case scenario),
speed improves by +7% on my laptop (415 -> 445 MB/s)
2018-11-08 12:47:46 -08:00
8bed4012bd
fixed decompression-only benchmark
2018-11-08 12:36:39 -08:00
a40b6958c9
Run TSAN fuzztest
2018-11-08 10:59:01 -08:00
103d1ee7a4
Add multithreaded dictbuilder tests to fuzzer.c
2018-11-08 10:58:51 -08:00
a8daa2d683
Signal before unlocking in pool.c
2018-11-08 10:45:53 -08:00
5b598bb203
Merge pull request #1408 from szreder/dev
...
Prevent deadlock on malloc() failure.
2018-11-08 10:43:40 -08:00
5c5c476338
Prevent deadlock on malloc() failure.
2018-11-08 10:29:31 +01:00
f6eb12084d
Merge pull request #1405 from pitrou/allow_cmake_debug_builds
...
Allow creating debug builds with CMake
2018-11-06 13:58:50 -08:00
2186fbcf67
Allow creating debug builds with CMake
...
Only set the build type to Release if the user didn't select any.
See discussion in https://github.com/facebook/zstd/pull/1278#issuecomment-436327923
and below.
2018-11-06 21:24:29 +01:00
e0701d3c5d
Merge pull request #1404 from facebook/T36302429
...
fixed T36302471
2018-11-06 11:53:20 -08:00
af892e5f1c
Merge pull request #1403 from facebook/assert_opt
...
just add an assert() in ZSTD_insertBtAndGetAllMatches()
2018-11-05 17:57:30 -08:00
3e5cdf1b6a
fixed T36302429
2018-11-05 17:50:30 -08:00
2caa995558
just add an assert() in ZSTD_insertBtAndGetAllMatches()
...
to express a condition on ll0 .
May help static analyzer as in #1397
2018-11-05 17:13:32 -08:00
3a90229616
Merge pull request #1395 from facebook/decompressblock
...
created zstd_decompress_block module
2018-10-29 16:28:09 -07:00
2a0c9e0806
Merge pull request #1398 from bmwiedemann/pgo
...
Fix building zstd-pgo
2018-10-29 16:00:26 -07:00
acd75a1448
fixed a second memset() on NULL
...
not sure why it only triggers now,
this code has been around for a while.
Introduced a new error code : dstBuffer_null,
I couldn't express anything even remotely similar with existing error codes set.
2018-10-29 15:03:57 -07:00
9c58098200
fixed memcpy() on NULL warning
...
memcpy(NULL, src, 0) is undefined behavior.
2018-10-29 13:57:37 -07:00
ea966c8fb1
Merge pull request #1396 from facebook/huf_refactor
...
refactor HUF_compress_internal for clarity
2018-10-29 13:06:45 -07:00
572279e73e
pgo: also optimize compression
...
and clean *.o so that linking does not fail from code
that is partially augmented with coverage generation instructions
2018-10-29 17:20:03 +01:00
a38ad53ace
zstd-pgo: do not make clean and zstd in parallel
...
otherwise, there might not be a ./zstd to run
2018-10-29 15:11:28 +01:00
fc20b3c441
added flag -Wc++-compat
...
for library and cli
2018-10-26 16:38:23 -07:00
4860311d99
renamed NEWS as CHANGELOG
...
to conform fb publication rules.
2018-10-26 15:57:13 -07:00
9bf00707c7
minor clarifications of history update rules
2018-10-26 15:51:51 -07:00
c3c7deb1e1
Merge pull request #1392 from coetry/dev
...
provide consistent spacing to enum field
2018-10-26 15:29:07 -07:00
0db109c041
Merge pull request #1394 from terrelln/circleci-packaging
...
Publish uncompressed, zstd, and gzipped packages
2018-10-26 15:28:44 -07:00
1866bd374a
Merge branch 'dev' into huf_Refactor
2018-10-26 15:25:01 -07:00
8d56f4baee
added a few comments for clarifications
2018-10-26 15:21:52 -07:00
450356b5af
Merge branch 'dev' into decompressblock
2018-10-26 15:03:43 -07:00
7d4960a5e8
Merge pull request #1390 from facebook/nullAsOutput
...
support decompressing an empty frame into NULL
2018-10-26 14:43:16 -07:00
3783720f70
vertically align code comment
2018-10-26 16:16:06 -05:00
7b74405150
refactor HUF_compress_internal for clarity
...
changed workspace parameter convention
to always provide workspaceSize,
so that size can be explicitly checked.
Also, use more enum to make the meaning of some parameters more explicit.
2018-10-26 13:21:37 -07:00
191e19aff8
Publish uncompressed, zstd, and gzipped packages
2018-10-26 13:06:53 -07:00
4ed9233db6
fixed VS2010 project
2018-10-26 11:25:40 -07:00
b832498aee
fixed : extraneous function prototype in fullbench.c
...
now provided through zstd_internal.h
2018-10-26 10:39:49 -07:00
26e34d8a73
provide consistent spacing to enum field
2018-10-25 18:45:20 -05:00
2b4914082e
created zstd_decompress_block module
...
isolate all logic associated with block decompression
into its own module.
zstd_decompress is still in charge
of context creation/destruction,
frames, headers, streaming, special blocks, etc.
Compressed blocks themselves are now handled within zstd_decompress_block .
2018-10-25 16:28:41 -07:00
e26126caca
Merge pull request #1388 from facebook/ddict
...
separate DDict logic into its own module
2018-10-25 10:20:40 -07:00
cb320a9fc0
added comment on public ddict functions
2018-10-24 16:50:03 -07:00
806a5c84e4
support decompressing an empty frame into NULL
...
fix #1385
decompressing into NULL was an automatic error.
It is now allowed, as long as the content of the frame is empty.
Seems to simplify things for `arrow`.
Maybe some other projects rely on this behavior ?
2018-10-24 16:34:35 -07:00
b7421f8e69
fix and refactored libzstd_partial_build.sh
...
make the test intention clearer
(easier to maintain)
also resist the existence of zstd_ddict.o within libzstd.a
2018-10-24 11:32:09 -07:00
debff3929b
fixed warnings in testpools
2018-10-24 10:36:06 -07:00
W. Felix Handte
Yann Collet
Yann Collet
Nick Terrell
Bartosz Szreder
Antoine Pitrou
Bernhard M. Wiedemann
Allen Hai