Yann Collet
b1f76e5990
more clearly identify error when there is no input
2025-01-22 09:49:08 -08:00
Yann Collet
f5a0b6fa8f
Merge pull request #1004 from Cyan4973/usan_fix
...
attempt to silence a recent ubsan warning with code reorganization
2025-01-22 09:34:22 -08:00
Yann Collet
71e797bd8c
add XXH_NOESCAPE
2025-01-21 20:55:25 -08:00
Yann Collet
c3e587a1ca
attempt to silence a recent ubsan warning with code reorganization
...
apparently, UBSAN uses the prototype of the inner inlined function
instead of the outer shell of the function.
Let's see if one indirection level is enough to fix this.
2025-01-21 17:28:25 -08:00
Yann Collet
1bf7bc4af2
Merge pull request #993 from walterdejong/xxhsum-help
...
Xxhsum improved help
2025-01-21 11:51:35 -08:00
Yann Collet
d5422ce901
Merge pull request #997 from Cyan4973/nostreamtest
...
fix nostreamtest
2025-01-21 11:51:18 -08:00
Yann Collet
51fa4ef151
Merge pull request #996 from 24bit-xjkp/dev
...
feat(xxhash3): Support LASX instruction set and refactor LSX implement
2025-01-17 09:57:12 -08:00
Yann Collet
26f622122f
fix nostreamtest
2025-01-13 16:50:45 -08:00
24bit-xjkp
7d6bd4ea09
feat(xsum): Display which mode is complied in the welcome message for LoongArch
...
1. Display the mode which is used as below:
"loongarch64 + lasx" -> LoongArch64 platform with LoongArch Advanced SIMD Extension
"loongarch64 + lsx" -> LoongArch64 platform with LoongArch SIMD Extension
"loongarch64" -> LoongArch64 platform, use scalar implement
2. Align the define in xxhash.h
2025-01-13 11:30:39 +08:00
24bit-xjkp
63e083c539
feat(xxhash3): Support LASX instruction set and refactor LSX implement
...
1. Use __lsx_vmul_d dircetly instead of using 2 32-bit multiply to emulate a 64-bit multiply.
2. Add LASX support.
2025-01-12 19:33:18 +08:00
Yann Collet
36cd8bfe01
Merge pull request #994 from sunpoet/dev
...
Update MANDIR for FreeBSD
2025-01-05 06:58:04 -08:00
Po-Chuan Hsieh
867e8752f3
Update MANDIR for FreeBSD
...
FreeBSD uses share/man/ rather than man/ now.
Reference: https://cgit.freebsd.org/ports/tree/devel/xxhash/Makefile#n17
2025-01-04 19:35:30 +08:00
Walter de Jong
f90089f90b
Revert "help is basically lying about omitting filename"
...
This reverts commit 1c54e8abc3 .
2025-01-02 10:15:56 +01:00
Walter de Jong
1c54e8abc3
help is basically lying about omitting filename
2025-01-02 09:09:38 +01:00
Walter de Jong
6d39e97f25
cleanup newline
2025-01-02 08:56:04 +01:00
Walter de Jong
5920111153
consistency typo
2025-01-02 08:55:26 +01:00
Walter de Jong
540433885f
align help message
2025-01-02 08:54:53 +01:00
Walter de Jong
7d49d13b41
help message shows option --files-from
2025-01-02 08:51:38 +01:00
Yann Collet
1ac318b264
Merge pull request #990 from Cyan4973/xxh3_finalize
...
Extract finalization from XXH3 hashLong
2024-12-27 17:58:57 -08:00
easyaspi314
a8d6ff0e13
Extract finalization from XXH3 hashLong
2024-12-27 17:37:17 -08:00
Yann Collet
abae92fe33
Merge pull request #989 from Cyan4973/xxh_leg_refactor
...
Modernize XXH32/XXH64
2024-12-27 17:36:42 -08:00
Yann Collet
4dd799c922
fix static analyzer warnings
...
they are suspicious that some tests on unsigned values might wrap around.
They should not, but let's write these branches differently to also cover these cases.
2024-12-27 17:15:50 -08:00
easyaspi314
bca2d275b2
Modernize XXH32/XXH64
...
- Fix the state structs to use unsigned char, update names
- Extract the algorithm steps into inline subroutines
- Fix a theoretical integer overflow bug
- Reroll XXH64 on 32-bit (it is going to run like crap anyway)
2024-12-27 16:28:26 -08:00
Yann Collet
86f6400a2a
updated CHANGELOG for v0.8.3
2024-12-27 02:11:10 -08:00
Yann Collet
2d70d4c00c
Merge pull request #988 from Cyan4973/cmake310
...
update cmake minimum version to 3.10
2024-12-26 22:58:43 -08:00
Yann Collet
43d8f7987d
update cmake minimum version test on github actions to 3.10
2024-12-26 22:38:18 -08:00
Yann Collet
81a5efc7e5
update cmake minimum version to 3.10
2024-12-26 22:27:10 -08:00
Yann Collet
67bba310c7
Merge pull request #987 from Cyan4973/libxxh_dispatch
...
LIBXXH_DISPATCH: control runtime vector dispatch in `libxxhash`
2024-12-26 20:06:57 -08:00
Yann Collet
9ff12f6797
control runtime dispatch in the library via separate variable LIBXXH_DISPATCH
...
disabled by default (in contrast with `DISPATCH`, which is enabled by default, for `xxhsum`).
2024-12-26 18:51:36 -08:00
Yann Collet
8aa8944b27
Merge pull request #983 from Cyan4973/mingw64_og
...
Added -Og test to mingw on Github Actions
2024-12-26 17:58:46 -08:00
Yann Collet
574aabad87
document XXH_NO_INLINE_HINTS as being useful for -Og
...
some compiler versions may require it.
2024-12-26 17:06:27 -08:00
Yann Collet
5dd119795e
fix -Og test, by adding -DXXH_NO_INLINE_HINTS
2024-12-26 17:01:02 -08:00
Yann Collet
71891a0bda
Added -Og test to mingw GA test
2024-12-26 16:53:08 -08:00
Yann Collet
cf9e2dc589
Merge pull request #986 from Cyan4973/autovec_x86
...
enable by default runtime detection of vector extension on x86/64 target
2024-12-26 16:47:48 -08:00
Yann Collet
5c6dd83b42
removed the XXH_X86DISPATCH_ALLOW_AVX build variable
...
if the program is compiled with, for example, -mavx2,
then the resulting binary will simply require this instruction set to run.
2024-12-26 16:24:09 -08:00
Yann Collet
5a8de81527
make XXH_VECTOR values more accessible
...
via #include "xxhash.h"
to fix double-include issues in xsum_arch.c
2024-12-26 16:10:04 -08:00
Yann Collet
cfd77237e6
updated doc
2024-12-26 16:02:40 -08:00
Yann Collet
877d9e5092
enable runtime detection of vector extension on x86/64 target
2024-12-26 15:25:51 -08:00
Yann Collet
6fe4f191b6
Merge pull request #985 from Cyan4973/autovec_rtdiag
...
xxhsum with auto-dispatch displays the locally detected vector extension
2024-12-26 11:36:23 -08:00
Yann Collet
4b9ca605cf
fix mingw32 test
...
the file statement just changed
2024-12-26 03:36:50 -08:00
Yann Collet
f9ea49e107
ensure that the file test in mingw ci test displays its outcome
2024-12-26 03:25:13 -08:00
Yann Collet
6bdecb5f88
fix newline minor warning
2024-12-26 02:29:15 -08:00
Yann Collet
dbcc17f87e
fix cmake recipe
2024-12-26 02:26:02 -08:00
Yann Collet
ad392603bc
fixed empty translation unit
2024-12-26 02:21:46 -08:00
Yann Collet
6c666adcdc
xxhsum with auto-dispatch displays the locally detected vector extension
...
using command -V
ex: compiled as 64-bit x86 autoVec (AVX2 detected) little endian with GCC 13.3.0
2024-12-26 01:55:05 -08:00
Yann Collet
2bf8313b93
Merge pull request #903 from Cyan4973/notFullyInlined
...
Ensure XXH_INLINE_ALL fully inlines all XXH32_ and XXH64_ symbols
2024-12-05 08:22:07 -08:00
Yann Collet
c5a6a14676
only enforce inlining of XXH64_finalize() when XXH_INLINE_ALL is enabled
2024-12-05 00:44:07 -05:00
Yann Collet
e876423360
fixed XXH64_finalize() not being inlined
...
even though `XXH_INLINE_ALL` is set
on `gcc-11`.
2024-12-04 11:13:54 -05:00
Yann Collet
61509f29b4
added a test that can catch if any XXH64_ symbol is not inlined
...
even though XXH_INLINE_ALL is set
2024-12-04 11:13:54 -05:00
Yann Collet
52a6f91461
Merge pull request #978 from Cyan4973/filelist
...
added --filelist command
2024-12-04 07:33:44 -08:00