You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=277905 a patch is proposed to remove the check for the --version-script link flag in lib/meson.build because it's breaking the new lld version. That patch seems to have no adverse affect on the Linux and FreeBSD builds, but it looks like it fails on OSX. e.g., https://github.com/hdholm/jose/actions/runs/8402521186/job/23011995537 and I could easily be missing something about which symbols are getting exported. I'm frankly a little out of my depth here. Of course the FreeBSD build could just patch that check out (which is almost certain to happen in the short term) but, of course, I'd rather see a more proper fix here. It's tempting to submit a PR to add host_machine.system() == 'freebsd' logic to the check and be done with it, or the inverse and simply check for OSX. But I feel like there's likely a better meson fix for this that actually accounts for the linker difference. I'm happy to write the PR, if someone can give me insight into the "right" answer here.
The text was updated successfully, but these errors were encountered:
Hello @hdholm. IMHO, I see no issue on doing this configuration conditional to freebsd, as it is done in meson.build for license, for example.
Ideally, this should be fixed by using an option that is available in all architectures, but, as OSX is being used on its latest version, and it fails, I don't see a cleaner way of proceeding.
Please, go ahead on creating the PR and, if we figure out another solution, we can proceed with it later.
In https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=277905 a patch is proposed to remove the check for the --version-script link flag in lib/meson.build because it's breaking the new lld version. That patch seems to have no adverse affect on the Linux and FreeBSD builds, but it looks like it fails on OSX. e.g., https://github.com/hdholm/jose/actions/runs/8402521186/job/23011995537 and I could easily be missing something about which symbols are getting exported. I'm frankly a little out of my depth here. Of course the FreeBSD build could just patch that check out (which is almost certain to happen in the short term) but, of course, I'd rather see a more proper fix here. It's tempting to submit a PR to add host_machine.system() == 'freebsd' logic to the check and be done with it, or the inverse and simply check for OSX. But I feel like there's likely a better meson fix for this that actually accounts for the linker difference. I'm happy to write the PR, if someone can give me insight into the "right" answer here.
The text was updated successfully, but these errors were encountered: