Search Criteria
Package Details: isa-l_crypto 2.26-1
Package Actions
| Git Clone URL: | https://aur.archlinux.org/isa-l_crypto.git (read-only, click to copy) |
|---|---|
| Package Base: | isa-l_crypto |
| Description: | A collection of optimized low-level functions targeting storage applications |
| Upstream URL: | https://github.com/intel/isa-l_crypto |
| Licenses: | BSD-3-Clause |
| Submitter: | afsio |
| Maintainer: | fft |
| Last Packager: | fft |
| Votes: | 1 |
| Popularity: | 0.000000 |
| First Submitted: | 2017-10-12 00:15 (UTC) |
| Last Updated: | 2026-03-08 13:05 (UTC) |
Dependencies (4)
- autoconf (autoconf-gitAUR) (make)
- automake (automake-gitAUR) (make)
- make (make-gitAUR, make-staticAUR) (make)
- nasm (nasm-gitAUR) (make)
Latest Comments
pobrn commented on 2025-12-17 12:57 (UTC)
@fft Tests run successfully now, thanks.
fft commented on 2025-12-16 16:42 (UTC)
Agreed, package updated. This issue doesn't reproduced for me (un)fortunately, so I can't check, whether this helps. @pobrn inform, whether all ok now, please.
pobrn commented on 2025-12-06 22:15 (UTC)
@fft But the default makepkg.conf enables LTO, so in my opinion one could reasonably expect that all(?) AUR packages build with that successfully.
In any case please see my new comment on the upstream issue: https://github.com/intel/isa-l_crypto/issues/165#issuecomment-3621275482 I believe this is a case of undefined behaviour. So instead of disabling LTO, it would be nice if
-fno-strict-aliasingcould be added here until an upstream fix.fft commented on 2025-11-28 07:32 (UTC)
@pobrn, tests work for me both with and without LTO and upstream doesn't force to use LTO also. So I think, that PKGBUILD is not a right place for fixing issue.
As I understand, this -lto=auto was added from makepkg.conf, right? May be you can override it as temp workaround? (I mean something like
makepkg CFLAGS="-flto=auto -O3 -g")pobrn commented on 2025-11-25 14:49 (UTC)
At least on my machine LTO causes the test failure ( https://github.com/intel/isa-l_crypto/issues/165 ). Could it be disabled until an upstream fix is found?
travisghansen commented on 2025-07-11 16:46 (UTC)
I am guessing it may be connected to specific cpu feature perhaps? I am on an amd machine..
fft commented on 2025-07-03 19:14 (UTC)
@travisghansen, I can't reproduce this. Tried to run this test with valgrind and/or clang. This doesn't look like packaging problem, but let's try to investigate it. Whether this is reproducible behaviour? If yes, whether data in 'ref:' string is always the same?
travisghansen commented on 2025-06-30 23:00 (UTC)
Getting this error trying to install..
fft commented on 2025-01-27 16:53 (UTC)
Hello, thank you for packaging. v2.25 released.
libisal_crypto.pc is buggy (same issue as for https://aur.archlinux.org/packages/isa-l#comment-1008968). Can you fix this also?