Hi, On Wed, 27 Nov 2024 22:02:41 +0100 Nicolas Schodet <n...@ni.fr.eu.org> wrote: > It seems that gcc-riscv64-unknown-elf still did not migrate to testing. > Is there something I can do to help? > > If I understand properly the problem, the code, or the GOT is too big > and relative offsets overflow. I have found a debian bug where they > found a solution to overcome this: https://bugs.debian.org/827651 this > change the code to allow loading values from big GOT.
instead of fixing the problem on mips64el, Keith seems to have decided to remove mips64el from the list of architectures that this package is built for: https://salsa.debian.org/debian/gcc-riscv64-unknown-elf/-/commit/d312586dfbdcc4f6d622ad6c78c28a0df1f671a3 But gcc-riscv64-unknown-elf 13.2.0-10+12 is still in testing for mips64el, so version 14 cannot migrate because it's (even though intentionally) missing a build on mips64el. You can see this reason in the excuses page: - Issues preventing migration: - missing build on mips64el https://qa.debian.org/excuses.php?package=gcc-riscv64-unknown-elf So there are two ways forward: 1. Keith files a bug with ftp.debian.org to remove the mips64el package of gcc-riscv64-unknown-elf from testing, thus declaring that the removal of mips64el from the package was indeed intentional. 2. Somebody fixes the mips64el problem for gcc-riscv64-unknown-elf and then re-introduces mips64el as an architecture that gcc-riscv64-unknown-elf is built for. In any case, this needs some action of Keith. I'm here as the picotool and pico-sdk maintainer which will also be removed from testing should this bug not get fixed. Thanks! cheers, josch
signature.asc
Description: signature
