Created attachment 233166 [details] devel/liberasurecode: update to 1.6.3, take over maintainership - take over maintainership of this unmaintained port - add a DOCS option - add a non-default SIMD option for adapting the binary to the CPU - patch out -Werror more thoroughly Changelog: https://github.com/openstack/liberasurecode/blob/master/ChangeLog Tested with Poudriere on armv7 arm64 i386 amd64 FreeBSD 13. The test suite passes on all of these targets, though I have not tested with JErasure (it's not in ports).
Committed. Thanks.
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=9d4a2bfa624891b509f0ee0fdfcbbc396fbc1bd9 commit 9d4a2bfa624891b509f0ee0fdfcbbc396fbc1bd9 Author: Robert Clausecker <fuz@fuz.su> AuthorDate: 2022-04-12 13:01:45 +0000 Commit: Tobias C. Berner <tcberner@FreeBSD.org> CommitDate: 2022-04-12 16:00:27 +0000 devel/liberasurecode: update to 1.6.3 - take over maintainership of this unmaintained port - add a DOCS option - add a non-default SIMD option for adapting the binary to the CPU - patch out -Werror more thoroughly Changelog: https://github.com/openstack/liberasurecode/blob/master/ChangeLog PR: 263241 devel/liberasurecode/Makefile | 17 +++++++++++++---- devel/liberasurecode/distinfo | 6 +++--- devel/liberasurecode/files/patch-Makefile.am (new) | 11 +++++++++++ devel/liberasurecode/files/patch-doc_Makefile.am (new) | 10 ++++++++++ devel/liberasurecode/files/patch-src_Makefile.am (gone) | 11 ----------- devel/liberasurecode/files/patch-test_Makefile.am | 6 +++--- ...st_builtin_rs__vand_liberasurecode__rs__vand__test.c | 6 +++--- devel/liberasurecode/pkg-plist | 3 +-- 8 files changed, 44 insertions(+), 26 deletions(-)
(In reply to Robert Clausecker from comment #0) > - patch out -Werror more thoroughly But why? Is there are warnings, ain't it better to fix them instead?
(In reply to Alexey Dokuchaev from comment #3) It is policy to remove -Werror as warnings can randomly appear despite there being nothing wrong with the code. Patching out pointless warnings incurs a significant maintainer load and can often not be done because warnings are e.g. caused by the use of deprecated interfaces. And getting the build to be warning free across all supported architectures and OS versions is often a ridiculous amount of work. Also there's always the chance that a well-meant patch to remove a warning actually introduces novel bugs that do not exist upstream. Recall for example the Debian fiasco [1] when such a patch removed the number of keys OpenSSL would generate to just 65536 possibilities. [1]: https://www.schneier.com/blog/archives/2008/05/random_number_b.html