[00:16] hi spiv [00:28] Morning poolie. [00:28] hi there [00:29] your lp haproxy change landed [00:29] i don't know if you need to qa it or something [05:58] morning all [05:59] morning jam et al. [07:12] hi jam [07:34] vila: can you try to message me again, I might have gotten ping to work [07:35] jam: pingeling [07:37] jam: so, does it work ? [07:38] hi jam, vila [08:02] Is there a newer CIA client than http://samba.org/~jelmer/bzr/cia_bzr.py ? I'm quite impressed with the notifications coming into #gnome-shell from a CIA bot. [08:07] good morning Bazaar [08:09] _o/ [08:18] no, it doesn't work [08:18] hi Riddell [08:18] vila: it doesn't highlight lines with my name in them, either [08:18] it apparently doesn't think my username is actually my username === jam is now known as jam-other [08:19] what a pity [08:19] jam-other: test [08:19] vila: try with a different name [08:19] jam-other: like that ? [08:19] yeah, doesn't work either [08:19] try "jam1" [08:19] yep [08:19] it still thinks that is my name... ugh [08:19] jam1: you mean *this* works ? [08:19] when I logged in a long time ago, it had to fall back to that [08:19] vila: yep [08:20] hmm, nice bug [08:20] and it doesn't seem to be following "/nick" [08:20] I'm going to log back in [08:20] brb [08:20] ok, I think it is working now [08:20] hopefully I don't lose my name again [08:20] jam: pingeling [08:21] vila: got it [08:38] jam: regarding bug #786980, [08:38] Launchpad bug 786980 in Bazaar "bzr: ERROR: bzrlib.errors.ReadOnlyError: A write attempt was made in a read only transaction" [High,In progress] https://launchpad.net/bugs/786980 [08:38] yessir [08:39] vila: ?? [08:39] jam: an interesting point is that the code involved is only triggered if bound_location is an url that will lead to the right branch [08:39] i.e. only small variations can trigger it (including the missing final slash) [08:40] so I've added tests for + and ~ but we already have coverage for urls that differs in a way that get_transport will point to a different branch (or no branch at all) [08:41] I'm merging 2.3 to 2.4 to trunk right now and will make my proposal when lp:bzr is up-to-date [08:41] vila: so you're saying we have coverage if URL mangling would cause us to not find the path? [08:42] yeah, because get_master_branch is called higher in the stack [08:42] so bound_location errors are also caught before we reach this code [08:43] sure, but %7E stuff will still open the right branch [08:43] it also explains why we didn't get more bug reports, the window is pretty small [08:43] jam: yup, hence the added tests [08:43] jam: but even %7E requires a manual user intervention [08:43] vila: we used to put one way into config files [08:43] and then changed which one we saved [08:43] jam: indeed [08:44] That may not be what is happening here [08:44] so the fix should cope with that [08:44] but *this* case could be that we used to keep the path read from the config file as Transport.base [08:44] and we started normalizing the path, etc. [08:44] oh, yeah, sorry isunderstanding [08:44] jam: yes, that may be the explanation, .base is now normalized so we shouldn't run into it anymore [08:45] s/that may be/that *is*/ but I don't feel like validating that by searching the history for the relevant change [10:11] maxb: is there some special stuff going on currently on jubany ? [10:11] Not that I'm aware of [10:12] mac_threads == 1 seems... wird [10:12] weird [10:12] ghaa [10:12] maxb: any objection about setting it to 8 again ? [10:12] hrm [10:13] I wonder why it's set to 1 [10:13] hehe, me too, hence my question [10:13] Regardless, we should set it back to 8 since no reason not to has been advertised [10:13] http://webnumbr.com/ubuntu-package-importer-conccurrency [10:14] done [10:15] maxb: once the queue is down again, I'll retry enigmail and similar (NoFinalPath) [10:15] Ah, when did your fix land? [10:15] hehe, when 2.4b4 did (or am I confused ?) [10:15] great [10:17] by the way, since we moved to 2.4b4, a number of packages seem to have started failing with BzrCheckError [10:18] ha crap, my fix is not included in 2.4b4 :-/ [10:20] maxb: I didn't follow who/what/when/why the upgrade was done, could we ask for 2.4b5 with good changes of success ? [10:20] The upgrade occurred as a side effect of jubany hardy->lucid [10:21] To get 2.4b5, we should speak to jelmer since he did the lucid-cat upload for 2.4b4 [10:22] ooooh, I see, so that's probably a side-effect of lp itself upgrading to 2.4b4 [10:22] jelmer: ^ [10:26] What was the bzr version before the lucid upgrade? [10:27] If it's https://bugs.launchpad.net/udd/+bug/806348, it may be that we now have a bzr version that fetches tags by default, which AFAICT is more likely to hit the corner case that triggers that error. [10:27] Ubuntu bug 806348 in Ubuntu Distributed Development "BzrCheckError: Internal check failed: Cannot add revision(s) to repository: missing chk node(s) for id_to_entry maps" [High,In progress] [10:28] Anyway, I'm battling a cold today so I should probably log off. [10:34] spiv: Before you go... [10:34] Would you be ok with me doing the appendrevisionsonly udd rollback? [10:38] I think the tag fetching is indeed the issue [10:42] hi vila, maxb [10:43] it's a side-effect of the builders almost upgrading to 2.4b4 [10:43] (Launchpad has its own bzr copy, the builders use the bzr from lucid-cat) [10:43] ha, thanks for confirming [10:43] jelmer: so how about 2.4b5 ? :D [10:44] vila: Happy to backport 2.4b5 as well, though I'd prefer to get it into oneiric first [10:44] jelmer: fair enough [10:45] jelmer: just let me know when it's deployed on jubany so I can retry the affected packages [10:45] vila: ok [10:52] jam: Thanks for proxying that RT for me :-) By the way, do you need /srv/package-import.canonical.com/new/test-scripts/ any more? (I tried to mv it into probably_obsolete, but it's owned by jameinel not pkg_import, so I couldn't) [10:52] don't you have sudo rights? [10:52] anyway, not specifically [10:54] I can sudo to pkg_import, but the file ownership is such that only jameinel or root can move it [10:54] removed [10:54] Thanks. [11:14] maxb: go ahead [11:14] acknowledged [11:50] * jelmer -> doctor, back in ~1.5 hrs [13:18] jam: https://code.launchpad.net/~vila/bzr/786980-url-aliases/+merge/68069 is up for review [13:31] vila: approved [13:32] \o/ [13:32] jam: thanks ;) === zyga is now known as zyga-afk [14:09] jelmer: what did the doctor say ? [14:11] vila: Salut! [14:12] vila: she took most of the stitches out :) [14:12] \o/ [14:12] (added bonus for 'she', they are more careful ;) [14:17] Careful? Nah, you just gotta do it quick. You grab one stick with a pair of pliers, see, and... === shadeslayer_ is now known as shadeslayer === jpds_ is now known as jpds === beuno is now known as beuno-lunch [16:41] * maxb unsubscribed ~bzr-core from lp:bzr/2.4 and subscribed ~bzr-codereview-subscribers instead [16:58] * jelmer waves to maxb [17:14] * vila fixes the wiki page used by the losas to create the 2.4 branch so in 6 months maxb won't have to tweak the subscriptions again [17:14] It seems immensely silly that we need LOSAs for this [17:16] maxb: DRY :D [17:16] maxb: at least it's every 6 months now rather than every month.. === beuno-lunch is now known as beuno === zyga-afk is now known as zyga === deryck is now known as deryck[lunch] === deryck[lunch] is now known as deryck === yofel_ is now known as yofel === tomaw_ is now known as tomaw [20:26] lifeless, in theory of course we can detect someone else has started doing the replacemen [20:26] in practice there are quite a few different possible traces [20:29] maybe we should have initialize(threaded=True) [20:43] poolie_: that sounds like a possible compromise [20:43] poolie_: my concerns about single-threaded overhead are perhaps unwarranted [20:43] poolie_: changing it and testing should be pretty straight forward [20:44] by booting loggerhead and running several clients? [20:47] 'bzr st' or similar [20:47] I meant assessing the impact of a simple mutex solution [20:48] right === medberry is now known as med_out