On Mon, Mar 13, 2023 at 6:38 PM Alexander Kanavin <[email protected]>
wrote:

> On Mon, 13 Mar 2023 at 18:28, Martin Jansa <[email protected]> wrote:
> >
> > On Mon, Mar 13, 2023 at 6:26 PM Alexander Kanavin <
> [email protected]> wrote:
> >>
> >> Tests in distrodata that contact upstream servers had to be disabled
> >> because it's not possible to tell what is a real failure, and what is
> >> a transient upstream issue (with expired certificates or other network
> >> issues). To ensure upstream version checks don't regress, I run them
> >> manually every now and then.
> >
> >
> > Thanks, can you confirm that git-submodule-test is currently failing for
> you (and probably was for long time) and
> > that binutils is often reporting failure?
>
> I run the check (basically 'devtool check-upgrade-status' with a
> manual review of what it prints) without meta-selftest included into
> the build,




> but I can try that if you want.
>

No need, "without meta-selftest included" explains why you haven't fixed it
already. I was reading "I run them
manually every now and then." as that you're triggering some special job on
AB manually, but with meta-selftest and
through OESelftestTestCase.test_checkpkg.

binutils is flaky, yes. The server is contaced mutliple times for
> various flavours of it, and some kind of silly rate limiting kicks in.
> If you run a standaline 'devtool check-upgrade-status binutils', it
> will work.
>

Yes, that's what I was seeing yesterday as well, but then it was failing
quite reliably in OESelftestTestCase.test_checkpkg until I've added some
debug into it.

Cheers,
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#178463): 
https://lists.openembedded.org/g/openembedded-core/message/178463
Mute This Topic: https://lists.openembedded.org/mt/97578950/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to