Created attachment 256873 [details] v0 (apply via 'git am') user@hostname: [?:0] /usr/ports/benchmarks/lzbench $ lzbench | head -1 lzbench 1.9 (64-bit FreeBSD) Intel(R) Core(TM) i7-7820HQ CPU @ 2.90GHz
Created attachment 256874 [details] 'poudriere bulk -t' log (14.2-R amd64)
Noticed we don't have '-r' option (recurse directories). Looks like here are problematic lines of code [0] which tries to detect system capabilities: #elif (defined(__APPLE__) && defined(__MACH__)) || \ ((defined(__unix__) || defined(__unix) || defined(__midipix__)) && defined(_POSIX_C_SOURCE) && (_POSIX_C_SOURCE >= 200112L)) /* snprintf, opendir */ # define UTIL_HAS_CREATEFILELIST # include <dirent.h> /* opendir, readdir */ I guess it would be nice to create a local patch or upstream PR, however I don't know how to work with _POSIX_C_SOURCE neither have GitHub account. [0]: https://github.com/inikep/lzbench/blob/cbeaa92b137558ae26df428c8a2cae7598a9c6e8/_lzbench/util.h#L352
Maintainer timeout.
I guess we should be updating the port to the latest version 2.0.1 now that it had been released. I'm not sure about the "dual GPLv2 GPLv3" licensing: this benchmark includes not only GPL'ed code. I've opted out of defining the LICENSE knobs originally because omitting it was easier than making it exhaustive. :-) > Noticed we don't have '-r' option (recurse directories). Yeah, we need a patch for this, no big deal. > Created attachment 256874 [details] 'poudriere bulk -t' log (14.2-R amd64) Please don't attach build logs unless being asked. Hardly anyone would ever look at them and they just clutter Bugzilla and eat disk space.
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=7cc244f0d1e7fff69b971cee8e0b18ed257809a4 commit 7cc244f0d1e7fff69b971cee8e0b18ed257809a4 Author: Alexey Dokuchaev <danfe@FreeBSD.org> AuthorDate: 2025-03-01 06:03:00 +0000 Commit: Alexey Dokuchaev <danfe@FreeBSD.org> CommitDate: 2025-03-01 06:03:00 +0000 benchmarks/lzbench: the port had been updated to version 2.0.1 Adjust preprocessor logic for the fact that FreeBSD has proper directory(3) and nanosleep(2) support. PR: 284231 Reported by: portscout benchmarks/lzbench/Makefile | 5 +++-- benchmarks/lzbench/distinfo | 6 +++--- .../lzbench/files/patch-__lzbench_util.h (new) | 20 ++++++++++++++++++++ 3 files changed, 26 insertions(+), 5 deletions(-)
(In reply to Alexey Dokuchaev from comment #4) > I'm not sure about the "dual GPLv2 GPLv3" licensing This is exactly what upstream declares [0]. In relevant pull request [1], MySQL has been mentioned as an example of software which includes lots of other software with different licenses, and it appears we just set 'main' license in ports in such cases: $ find /usr/ports/databases/ -maxdepth 1 -type d -name mysql\*server -exec make -C {} -V LICENSE \; GPLv2 GPLv2 GPLv2 [0]: https://github.com/inikep/lzbench/blob/master/LICENSE [1]: https://github.com/inikep/lzbench/pull/158