[05:08] hi? [05:08] i wonder if pqm is just not sending error messages at all [05:08] ooh, or maybe there is a size limit [07:34] hi all ! [07:55] hi there vila [07:55] i liked your option expansion patch [07:56] * vila happily shudders :) [07:57] the HPSS glitch found by jelmer is a serious (and depressing) issue which was well hidden behind a @property that took me off-guard [07:58] hopefully it can addressed with some additional store classes (jelmer said he was looking at it) [07:58] s/can/can be/ [08:01] ha ! pylzma fmp finally landed, poolie, what was the issue ? [08:02] i think it made every test fail in the original version [08:03] and that was so much bad news the mta dropped it [08:03] hehe [09:08] morning all [09:12] mgz: hey ! [09:15] hi everyone [09:15] i have a question about chmod and bazaar [09:16] afaik when i change the file permissions bazaar lists that as a change [09:16] is there any way to avoid this? to tell bazaar to ignore file permissions changes [09:16] ? [09:17] bzr only tracks the excecutable bit [09:18] you can either just not worry about bzr telling you that has changed, or er... [09:19] o/ mgz [09:19] there were some bugs involving windows/cygwin where the excecutability was changing all the time, but most things shouldn't be changing it [09:20] * mgz waves at vila and poolie [09:23] pqm failure on bzrlib.tests.test_test_server.TestTCPServerInAThread.test_handle_request_closes_if_it_doesnt_process(TestingTCPServer) :-) :-/ :-( [09:24] so new buildd stuff is almost but not actually deployde [09:25] it's teasing everyone :) [09:30] mgz: yeah i just found that bug [09:33] deni: ah, is that what you were running into? have you got the bug number and done +affectsmeto as well? [09:35] incidentally my affects count across dupes is now live [09:35] the data is a bit more representative [09:37] that's good, had wondered why it didn't seem to do that. [09:42] mgz: i'm still investigating what's actually going on [09:44] deni: do report when you find out. [10:03] pfew, finally, sanity restored in mail inboxes [10:43] mgz: fixed the issue...it was an error on out part...seems unrelated to that bug [10:44] deni: cool. I'm still curious about the details :) [10:58] argh, transports are a mess. [11:01] mgz: as far as path encodings are concerned you mean ? [11:02] the url handling is all over the place. [11:02] they all do their own things with string parsing, and randomly add and remove trailing forward slashes [11:03] yeah :) [11:03] some tests may also enforce some bugs here [11:03] must resist the urge to rewrite, and just fix the current bug first [11:03] well, bugs, practices at least [11:04] certainly bugs too. [11:11] okay, done PathFilteringTransport, LocalTransport, ConnectedTransport... next failure is MemoryTransport [11:11] what's your current bug ? [11:12] bug 842233 / landing jelmer's rmbranch-colo mp [11:12] Launchpad bug 842233 in Bazaar "InvalidURL from tests for local_path_from_url on windows" [Medium,In progress] https://launchpad.net/bugs/842233 [11:13] basically the logic is bogus but happens to mostly work [11:14] yeah, there is a pattern there: one feature implemented in many different places, mostly tested instead of one feature implemented in a single place, fully tested [11:15] it's really just existing debt from all the different quirks to url handling [11:16] in some ways it's better to land a change that needs it cleaned up and break things, so it's clearer what the requirements for the refactoring are [11:17] same thing :) The debt grows because it's in many different places and the short term gains reinforce the pattern [11:29] mgz: what can i tell...user error! :) naturally i jumped to conclusions and blamed bzr instead of my collegue :) [11:30] deni: no worries, but understanding what happened and why bzr presented a confusing output is what we're trying to understand ;) [11:30] bah, mgz syndrom... understanding at both start and end of the sentence... [11:30] :D [11:32] slangasek: ping, I like to make sure I understand when do-po-merges is used, do you have a few minutes for me ? [11:32] vila: he chmoded the whole directory manually, and that is fine, i mean it's fine that bzr tracks that/ [11:33] i was thinking that it had something to do with the windows clients when bzr showed him that he has over 2000 filed modified :) [11:33] deni: ha ok, great, that all I wanted to know [11:33] :D [11:33] sorry for the hastle [11:34] still i would be interested to know is there a feature planned that would disable file permission changes? [11:34] *disable the tracking of [11:36] not that I know of [11:37] from a high level, bzr tracks only the executable bit because it's the only that really makes sense from a VCS point of view [11:37] and has a meaning on all platforms [11:37] the other bits are the job of install scripts generally not the VCS [11:38] i have a lot of .bat and .sh files.....windows doesnt care about +x on bat files [11:38] afaik [11:38] but yeah, i get your point [11:42] okay, lame fix now done. [11:42] I just need to go out briefly, will propose when I return. [12:08] Riddell: ping [12:10] Riddell: does 'po4a' rings a bell ? What's the relationship with a 'doc/po4a' dir and the 'po' dir in the same package ? [12:27] vila: nope, don't think I've herad of it [12:27] heard [12:28] k, thanks [12:28] seems like po4a has tools for converting other file formats into po [14:52] ah....the latest xkcd....so true [14:52] :) === beuno is now known as beuno-lunch === beuno-lunch is now known as beuno [21:37] Launchpad just emailed me that a build failed with "bzr: ERROR: no such option: --allow-fallback-to-native" - https://launchpadlibrarian.net/85424088/buildlog.txt.gz [22:18] Is it appropriate to discuess bzr-builder issues here? [22:34] is there a way to 'undo' a bzr unshelve? aka move the state back to how it was before that command? [22:35] i apparently had a modified file, unshelved, it failed to merge *some* parts into that file but merged others and now im quite lost as to what was what [22:43] Leon_Nardella: yes [22:43] PsyberS: unfortunately not [22:44] merges with conflicts should make a .OTHER file, not sure if unshelve also does [22:45] wgz: it doesnt, that i can see but it did leave the old shelf backup file .~1~ there, so at least i can see what the shelved version looked like (and the current) and figure it out manually [22:45] wgz, I have a recipe in Launchpad that won't build. In the build log there's an error message saying that {git-commit-id}, which I'd like to use in the deb-version, couldn't be expanded. Any ideas? [22:49] ah, I see you filed bug 892358 [22:49] Launchpad bug 892358 in bzr-builder "Bzr-builder doesn't expand {git-commit}" [Undecided,New] https://launchpad.net/bugs/892358 [22:50] bug 891880 also seems relevent [22:50] Launchpad bug 891880 in Bazaar "mentions new substitution variables even when recipe format doesn't support them" [Medium,Triaged] https://launchpad.net/bugs/891880 [22:51] * Leon_Nardella taking a look at 891880 [22:51] Yup, seems relevant. [22:51] I tried changing the format to 0.4, but Lauchpad didn't accept it. [22:52] also this thread on the udd list: [22:52] [22:52] particularly [22:52] so, I think the state from this morning, was poolie said the new builders are nearly but not quite deployed now [22:53] and I guess your error stems from that [22:53] Hmm.. Guess I should just relax and wait a bit? [22:53] so... with luck it'll work on monday :D [22:53] Cool. :) [22:53] right, dinner for me. [22:53] See ya