On Fri, 2023-10-20 at 15:12 +0200, Alexander Kanavin wrote: > On Wed, 18 Oct 2023 at 21:39, Alexandre Belloni > <alexandre.bell...@bootlin.com> wrote: > > ERROR: Can't find a task we're supposed to have written out? (hash: > > e79d70b9c2cc72030c1ce822525510699a1eeb1ddf5986271d3217422244366a)? > > ERROR: Can't find a task we're supposed to have written out? (hash: > > 999250461834bf74545962a8577bc26854a3d08bb2a96140d5efe314e0ef2d63)? > > This is proving tricky to reproduce. It happens fairly reliably in > oe-selftest-armhost, but so far never on regular oe-selftest, or in > standalone local build on an arm host where I run only the selftests > that are failing. It seems like one of those dreaded selftest things > that fail only in particular execution order with other tests. > > The only idea I have right now is to run specially configured > oe-selftest-armhost which would run only a subset (e.g. just > sstatetests) and see if it can be narrowed down from there. > > I'm also tempted to set this aside for now as it has all the signs of > protracted, painful debugging slog, and push the esdk/sstate build > registry subject, so there's something to show for it soon.
I did idly wonder if we need to set the sig handler to not use hashequiv for these tests? I haven't checked if that would help though, I'm just making a wild guess. Cheers, Richard
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#189533): https://lists.openembedded.org/g/openembedded-core/message/189533 Mute This Topic: https://lists.openembedded.org/mt/102035446/21656 Group Owner: openembedded-core+ow...@lists.openembedded.org Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-