I wanted to write down where I've got to in digging into the rust oe-selftest failure... much of this is probably repeating things which have already been said/are in the ticket (https://bugzilla.yoctoproject.org/show_bug.cgi?id=15275), but it feels like it needs more discussion than ticketing...
As far as I can tell what we've tripped over (based on my reading of the code), is that upstream rust only runs their test suite on nightlies, not on releases and the test suite is now using features that are only available in nightlies and are explicitly disabled in releases (either through build time configuration, or run time checks). 1.71.0 is where it all breaks, the changes which are introduced in that extend the testing to cover the bootstrap process, which then makes unpicking those changes all but impossible. I suspect we could patch the build so that the run time checks which we trip over could be bypassed, but it feels like that would move us to a long running world of pain which is going to need people who really know what they're doing with a rust codebase (which is definitely not me!) We're now 4 releases behind on rust, with a couple of point releases mixed into that. There's really only two ways forward I can see: find someone to unpick this with upstream and make testing in release a thing that they test, or drop the rust test suite from oe-selftest. -- Alex Kiernan
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#191216): https://lists.openembedded.org/g/openembedded-core/message/191216 Mute This Topic: https://lists.openembedded.org/mt/102786409/21656 Group Owner: openembedded-core+ow...@lists.openembedded.org Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
