ci: More flatpak ci fixes #824
Conversation
|
Yup, makes sense! Though there are still some failures it seems. 😢 |
Not sure what happened there. Seems like the web server couldn't get started? All the other failures are variations of: |
|
It looks like this got a lot closer...we possibly got OOM killed? bot, retest this please |
|
bot, retest this please |
I was seeing this when trying to run flatpak's tests in ostree's CI: ostreedev/ostree#824 The race here is that the python process can still be writing to the output while sed is reading it, and hence we'll find a difference on the next line. Fix this by making a tmp copy of the file, which then both sed and cmp will read consistently. I'm not *entirely* sure this will fix the problem as I couldn't easily reproduce the race locally, but I believe it at least fixes *a* race.
|
I think this is flatpak/flatpak#772 |
|
bot, retest this please |
I was seeing this when trying to run flatpak's tests in ostree's CI: ostreedev/ostree#824 The race here is that the python process can still be writing to the output while sed is reading it, and hence we'll find a difference on the next line. Fix this by making a tmp copy of the file, which then both sed and cmp will read consistently. I'm not *entirely* sure this will fix the problem as I couldn't easily reproduce the race locally, but I believe it at least fixes *a* race.
|
Ah yep, RAM was the problem. This should be GTG. |
ci/flatpak.sh
Outdated
| dnf builddep -y ostree flatpak | ||
| dnf install -y flatpak && rpm -e flatpak | ||
| # From the current flatpak builds | ||
| dnf install -y gcc sudo which attr fuse gjs parallel clang libubsan gnome-desktop-testing 'pkgconfig(fuse)' 'pkgconfig(gio-unix-2.0)' 'pkgconfig(gobject-introspection-1.0)' 'pkgconfig(json-glib-1.0)' 'pkgconfig(libarchive)' 'pkgconfig(libelf)' 'pkgconfig(libsoup-2.4)' 'pkgconfig(ostree-1)' 'pkgconfig(polkit-gobject-1)' 'pkgconfig(libseccomp)' 'pkgconfig(xau)' 'bubblewrap' docbook-dtds docbook-style-xsl intltool libattr-devel libcap-devel libdwarf-devel elfutils systemd ostree gpgme-devel /usr/bin/xmlto /usr/bin/xsltproc redhat-rpm-config /usr/bin/update-mime-database /usr/bin/update-desktop-database /usr/bin/gtk-update-icon-cache |
There was a problem hiding this comment.
How about we just fetch it from the YAML file after we clone it, e.g. something like:
python3 -c 'import yaml; y = list(yaml.load_all(open(".redhat-ci.yml")))[0]; print("\n".join(y["packages"]))' | xargs dnf install -y
?
There was a problem hiding this comment.
OK, done (though I had to use NUL rather than newline separated to make e.g. pkgconfig(gobject-introspection-1.0) >= 1.40.0 work).
That said the reason I didn't do this is I think at least for flatpak we should switch upstream to do privcontainer rather than host as we are here - and that circles back to enhancing rhci for this, or prototyping it outside perhaps.
We need our `make install` to override the ostree RPM, so do it all in one txn. This sort of thing is where a more rigorous model like rdgo/gcontinuous use becomes better, but we'll hack it with shell for now.
|
OK, rebased and squashed. |
|
Nice! |
|
☀️ Test successful - status-atomicjenkins |
We need our
make installto override the ostree RPM, so do it all in one txn.This sort of thing is where a more rigorous model like rdgo/gcontinuous use
becomes better, but we'll hack it with shell for now.