Created attachment 574836 [details] sci-mathematics:pari-2.9.4-r1:20190430-090529.log >>> Completed installing pari-2.9.4-r1 into /var/tmp/portage/sci-mathematics/pari-2.9.4-r1/image/ * Final size of build directory: 105264 KiB (102.7 MiB) * Final size of installed tree: 37472 KiB ( 36.5 MiB) * QA Notice: The following shared libraries lack a SONAME * /usr/lib/libpari.so.2.9.4 strip: hppa2.0-unknown-linux-gnu-strip --strip-unneeded -R .comment -R .GCC.command.line -R .note.gnu.gold-version usr/lib/libpari.so.2.9.4 usr/bin/gp-2.9
Can you please check that this still affects the v2.11.2 that I just committed? (It probably does). We have a very suspicious line in the ebuild, use hppa && \ mymake=DLLD\="${EPREFIX}"/usr/bin/gcc\ DLLDFLAGS\=-shared\ -Wl, soname=\$\(LIBPARI_SONAME\)\ -lm that might even be causing this problem. Does the build succeed for you on hppa without that line? Does deleting the line fix the soname issue?
Can anyone from hppa@ check this?
Would love to help but need help helping.
Interestingly, I don't even see the message at all now.
Created attachment 782612 [details] build.log
It is possible that the underlying problem was fixed upstream in the meantime. I always suspected the use hppa && \ mymake... line, but it was just a guess. (Is it still needed?)
(In reply to Michael Orlitzky from comment #6) > It is possible that the underlying problem was fixed upstream in the > meantime. I always suspected the > > use hppa && \ > mymake... > > line, but it was just a guess. (Is it still needed?) Trying. While we wait: - https://github.com/gentoo/gentoo-historical-2/commit/10663e580016203a42e5815d74048c9dc866b249 ???? - https://github.com/gentoo/gentoo-historical-2/commit/eb7905e41ed0043c06e31b7795a106bb3c8ce30a - https://github.com/gentoo/gentoo-historical-2/commit/18117ff9e466f9d9a11b21c9eb96d46076d72862
I lack the patience to look further but I'm guessing this got restored accidentally tbh (copied from wrong ebuild as a base by somebody at some point)?
(In reply to Sam James from comment #8) > I lack the patience to look further but I'm guessing this got restored > accidentally tbh (copied from wrong ebuild as a base by somebody at some > point)? As good a guess as any. It's certainly not *correct*, so if the build succeeds without it I'll ditch it in an -r1.
(In reply to Sam James from comment #8) > I lack the patience to look further but I'm guessing this got restored > accidentally tbh (copied from wrong ebuild as a base by somebody at some > point)? Quite possibly me. I may have copied the hppa stuff from the pari 2.3 ebuild before it was removed, while sage was using pari-2.4 - one of these development "release". I may not have noticed when it was dropped from the main tree.
Don't worry about it, it happens :) Yes, we're all good. Built fine w/ the line dropped.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=d1d77480299ec0900a8602adbb8469c22132c84c commit d1d77480299ec0900a8602adbb8469c22132c84c Author: Michael Orlitzky <mjo@gentoo.org> AuthorDate: 2022-06-04 13:03:59 +0000 Commit: Michael Orlitzky <mjo@gentoo.org> CommitDate: 2022-06-07 14:40:53 +0000 sci-mathematics/pari: remove obsolete hppa workaround in an -r1. Bug: https://bugs.gentoo.org/684814 Package-Manager: Portage-3.0.30, Repoman-3.0.3 Signed-off-by: Michael Orlitzky <mjo@gentoo.org> sci-mathematics/pari/pari-2.13.4-r1.ebuild | 117 +++++++++++++++++++++++++++++ 1 file changed, 117 insertions(+)
Thanks sam, marking fixed. The line will phase out soon enough.