=== popey8 is now known as popey [06:06] morning [07:14] morning [07:16] PR snapd#10094 opened: overlord/snapstate: skip catalog refresh when snappy testing is enabled [07:16] pstolowski: hey [07:43] good morning === popey8 is now known as popey [07:55] zyga: hey [07:56] mborzecki there was some spam last evening [07:56] but I think it got sorted out [07:56] on one of the high-profile bugs [08:00] zyga: did you manage to get that landscape thing sorted out? [08:04] mborzecki: yes I did, I have several machines now [08:04] mborzecki: some small issue remains but we're trying to sort that one as well [08:04] zyga: just a couple? not a 100s like before? :) [08:05] mborzecki: I seem to have access to employee subscriptions [08:05] mborzecki: and can see hardware stats as well [08:06] yay [08:59] hmm refreshed the github review page and all fonts are bitmap, looks like they request helvetica now? [09:19] and a false alarm, fontconfig update moved the config files to /usr/share/fontconfig/conf.avail but not all symlinks in /etc/fonts/conf.d were updated [09:45] hmm is cla-check broken everywhere or is it just me and my email address? [09:47] hmm must be me, dang [09:59] so, apparently it is https://bugs.launchpad.net/launchpad/+bug/1921727 [09:59] Bug #1921727: Error 503 when trying to iterate over lp.people[team].participants [10:00] oh [10:00] looks like a recent regression [10:01] Yeah, except that nothing relevant changed :( [10:03] is there a server-side oops with a backtrace? [10:05] mhm [10:05] zyga: yes [10:07] it's a slow query that really shouldn't be particularly slow and that performs reasonably on staging, so needs a fair bit of digging [10:41] Looks like a query tipped over the edge into using a completely inappropriate plan, perhaps because the team is large [10:41] I don't think it was a change in Launchpad, but maybe just adding more people to the team [10:44] Staging's PostgreSQL does things in a reasonable order; production's PostgreSQL (when querying for that particular team) takes the index of all user accounts sorted by names and goes through that looking for people who are in the relevant team, which is, uh, not as fast [10:45] yeah sounds very ineffective [10:58] cjwatson nice analysis! [11:12] PR snapd#10095 opened: tests: add simple sanity check for systemctl show --property=UnitFileState for unknown service === alan_g_ is now known as alan_g [11:43] hm refresh endpoint is rate limited too? https://paste.ubuntu.com/p/9WVn6cjD36/ [11:57] PR snapd#10091 closed: overlord/snapstate, overlord/ifacestate: move late security profile removal to ifacestate [11:57] pedronis: thanks! [12:07] pedronis: what's the purpose of refreshCandidates() inside `func (r *refreshHints) refresh()` ? we are not modifying any local state are we? [12:27] cjwatson: thanks for updating the bug report and looking into it [12:32] PR snapd#10096 opened: gadget: add support for kernel command line provided by the gadget [13:37] * pstolowski lunch [13:42] PR snapd#9940 closed: boot: cmd/snap-bootstrap: handle a candidate recovery system v2 [13:42] PR snapd#10087 closed: release: 2.49.2 [13:42] PR snapd#10094 closed: overlord/snapstate: skip catalog refresh when snappy testing is enabled [13:42] mborzecki: ^ [13:47] pedronis: can you land #10089 and #10073 ? the failures there are all unrelated [13:47] Bug #10089: please don't depend on libstdc++5-dev [13:47] PR #10089: o/configstate/configcore: simple refactors in preparation for new function [13:47] PR #10073: o/devicestate/devicemgr: register install-device hook, run if present in install [13:47] Bug #10073: kernel-package: new changes from Debian require merging [13:51] hey cachio if you have time, I've seen ubuntu-20.04-64:tests/main/interfaces-desktop-document-portal fail because of a leftover dbus session daemon, can you try porting that test to use tests.session for systems that support using the user session (`tests.session has-session-systemd-and-dbus`)? I think that would make the leftover dbus session daemon go away on i.e. ubuntu-20.04 [13:51] ijohnson, hey [13:51] sure [13:51] on systems which don't support session systemd probably it's fine to just keep testing as-is with `su -l -c ...` and accept that maybe sometimes the test there will fail [13:52] over time those systems will fall out of support [13:52] thanks cachio [14:26] pedronis: thanks! [14:27] PR snapd#10090 closed: interfaces/apparmor: followup cleanups, comments and tweaks === jdstrand_ is now known as jdstrand [17:03] PR snapd#10097 opened: tests: use tests.session tool on interfaces-desktop-document-portal test === ijohnson is now known as ijohnson|lunch === ijohnson|lunch is now known as ijohnson [17:29] cachio: a comment on that pr [17:29] ijohnson, nice, thanks for reviewing [17:44] PR snapcraft#3479 closed: Use has-signed-canonical-cla GitHub Action [17:44] PR snapcraft#3481 closed: tests: stub job to get autokpgtest for edge [18:04] PR snapcraft#3446 closed: extensions: suppress realpath noise on headless systems [19:06] ijohnson, pr updated [19:13] cachio thanks I'll take a look in a bit [19:13] tx [19:43] PR snapd#10098 opened: tests: updated restore phase of spread tests - part 1 [23:19] PR snapcraft#3489 closed: errors: introduce details_from_called_process_error() helper