Hello @ppc, We have a patch (pari-2.9.4-ppc.patch) for sci-mathematics/pari that modifies an ugly architecture/OS detection script by adding, linux) arch=`uname -m` case $arch in + ppc) arch=powerpc;; So if the OS is "linux", then it will mangle the arch from "ppc" to "powerpc". ...However, in two other places in that same script, they do darwin*) arch=`uname -p` if test "$arch" = powerpc; then arch=ppc; fi ;; and aix) arch=`uname -p`; case $arch in powerpc) arch=ppc;;esac;; So it looks like "ppc" is what they want the canonical arch name to be anyway? IDK. The reason I'm second-guessing this now is because I'm trying to reduce the pile of patches we have, either by sending them upstream, working around them in the ebuild, or simply dropping them. Can someone please build-test pari (the test suite should pass) on ppc WITHOUT pari-2.9.4-ppc.patch, and let me know what happens? If it breaks I'd like to make a report upstream. If not, git rm.
Pretty please @ppc? Otherwise I'll eventually have to just cross my fingers and try removing it, letting users do the testing.
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=982b872bba85553067ed601aaf8f553994811801 commit 982b872bba85553067ed601aaf8f553994811801 Author: Michael Orlitzky <mjo@gentoo.org> AuthorDate: 2023-07-27 03:09:16 +0000 Commit: Michael Orlitzky <mjo@gentoo.org> CommitDate: 2023-07-27 03:36:32 +0000 sci-mathematics/pari: add 2.15.4 I've dropped pari-2.9.4-ppc.patch from this version to see if it's still necessary, and to report the problem upstream if so. Closes: https://bugs.gentoo.org/891887 Signed-off-by: Michael Orlitzky <mjo@gentoo.org> sci-mathematics/pari/Manifest | 1 + sci-mathematics/pari/pari-2.15.4.ebuild | 132 ++++++++++++++++++++++++++++++++ 2 files changed, 133 insertions(+)