[01:42] doctor who event tonight! [01:42] \o/ [02:29] so the daily from monday installer crashed before I even started [02:42] This it? https://www.kubuntuforums.net/showthread.php?68163-Be-Wily-and-Enjoy-the-Progress-of-Kubuntu-15-10&p=379554&viewfull=1#post379554 [02:42] I grab the iso and threw it in a VM [02:42] the first line was about OEM config [02:43] maybe some changes sitter did at randa [03:52] mparillo: here is the long from a build today (9/16/15) https://notes.kde.org/p/DailyInstallerCrash [03:52] *log [03:57] I got it to start with running it from konsole [04:48] ahoneybun: UCC nominations happening [04:48] you interested? [07:16] yofel: [07:16] Started 10 hr ago [07:16] Build has been executing for 10 hr [07:19] wow [07:38] hm, where i can set two-click-open-folder behavior? [07:39] because I've always had two-click to open, after vivid->wily upgrade, it is one-click-to-open [07:43] oh, "mouse settings" [07:44] vip: this chan is for development [07:44] support in #kubuntu [08:17] sitter: kill it, clivejo wanted to upload applications, but I'm not sure if he actually got to do it [08:19] killed it [08:19] you really must not let CI pause over night [08:19] it basically backs everything up by a day [08:19] oh ok, sorry [08:46] http://anonscm.debian.org/cgit/pkg-kde/applications/akonadi.git/log/?h=kubuntu_wily_archive [08:46] someone needs to fix that script [09:00] hm... I did take a look at that a long time ago, seems like dch is still being run incorrectly o.O [09:01] and the push semantics are still wrong in that one. --nopush should be --push and the default be the other way around [09:08] this patch is getting on my nerves [09:11] Riddell: the staging failed towards the end, I have no idea what happened it [09:13] got up to ktnef and it went haywire [09:32] clivejo: I'll try it then [09:32] clivejo: did you get anywhere with frameworks? [09:33] think I fixed khtml with yofel's help [09:34] Riddell: can I retry it again? [09:35] it I remove ~/src/applications and restart it? [09:35] wasnt a good idea doing it in evening time on this internet connection [09:35] clivejo: just start from where it broke [09:36] clivejo: remove the lines in package-name-lists/applications-wily up to ktnef [09:36] was just going to ask was that possible [09:37] if I leave ktnef there and let it try again? [09:37] yes [09:37] maybe rm -r ~/src/applications/ktnef [09:37] could have been internet related [09:37] done [09:38] yeah same crazyness [09:39] https://paste.kde.org/ptnfmdaet [09:40] it has a blank epoch [09:40] couldnt see that last night [09:43] Riddell: ^ [09:47] where does it pick up the epoch from? [09:48] Riddell, clivejo: kindly note that gpgmepp needs a nwe upload [10:20] clivejo: the no-epoch file lists packages which are new so don't need and shouldn't get an epoch [10:21] I dont understand? [10:21] clivejo: in kubuntu-automation there's a file no epoch [10:21] no-epoch [10:21] it lists some pacakges which don't need an epoch [10:21] ah but ktnef does need an epoch [10:22] so that should be fixed, just remove it from the no-epoch file [10:22] I cant find a no-epoch file [10:23] ah [10:23] packaging-exceptions.json [10:25] ah right yes that's where it ended up [10:25] ok, picking up where I left off [10:25] FFS [10:25] I HATE gpg-agent [10:28] * yofel only just realized what happened to akonadi [10:28] wtf [10:29] what happened? [10:29] http://anonscm.debian.org/cgit/pkg-kde/applications/akonadi.git/log/?h=kubuntu_wily_archive [10:29] a) that shouldn't be there 3 times, b) the version is wrong [10:29] I totally didn't see the version when sitter pointed that out earlier [10:30] throw away that script already and build something in jenkins :P [10:30] I have first class version handling classes you know ;) [10:30] yeah, but that's ruby [10:31] which is probably why my code doesn't come up with wrong changelog entries :P [10:31] how on earth did that happen? [10:32] where is that version in the changelog entry even coming from o.O [10:32] that shouldn't be there [10:33] clivejo: are you overriding the changelog message? [10:34] ah man, I have a typo in the changelog message [10:35] ./staging-upload -d wily -v 15.08.1 -m "New upstream release (15.18.1)" -r applications -t /home/clivejo/src/applications/ [10:35] you don't usually put the version in there, if anything, it should say "New upstream bugfix release" [10:36] one more reason to move this into a jenkins :P [10:37] less opportunity for human error [10:37] * sitter is under the impression that the KCI build queue is growing rather than shrinking [10:37] yofel: how do I fix this? [10:37] clivejo: fix the changelog in each package by hand [10:38] or write a script for it [10:38] bash for loops may help [10:38] bash never helps. [10:38] well, he has to delete the old messages [10:38] I'd rewrite the entry TBH [10:38] which requires python [10:38] you can do that? [10:38] for asdf in */; do cd ${asdf}; sed s,15.18.1,15.08.1, debian/changelog; cd ..; done [10:38] ah [10:39] Riddell: that won't fix the message being there 3 times [10:39] FWIW its not the first time we have duplicated messages [10:39] ach when that happens I just leave it [10:39] so perhaps we should just not care :P [10:39] happens quite a lot [10:39] I don't see what's wrong with the regex for tha though (other than it being a regex) [10:39] * clivejo is extremely sorry [10:39] right, and I find that seriously annoying [10:40] clivejo: well, the multiple messages isn't exactly your fault [10:40] it's a combination of you re-running the script, the script pushing immediately and the don't-add-changelog-twice regex not doing it's job [10:40] clivejo: want me to upload frameworks? [10:41] I think once I argued that it shouldn't do the push right away anyway [10:41] Riddell: have they been tested? [10:41] clivejo: nope, I'll do that first of course [10:41] !testers [10:41] testers is Help is needed in #kubuntu-devel. Please ping Riddell, yofel, soee, Tm_T, shadeslayer, BluesKaj, James147, Quintasan, lordievader, shrini, tester56, parad1se, mamarley, alket, SourBlues, sgclark, neo31, vip, mparillo for information [10:41] ideally the script would create an integration branch for the specific version as to avoid push clashes [10:41] then push that instead [10:41] frameworks http://qa.kubuntu.co.uk/ppa-status/frameworks/build_status_5.14.0_wily.html testers! [10:41] once uploading to archvie the branches can then be merged in one go [10:42] good idea actually, that could at least be easily reverted [10:43] so what do I need to do then? [10:43] that bash loop that Riddell posted looks pretty sane, run that [10:44] well test it first :) [10:44] but how does that fix debian git changelogs? [10:45] you'll need to adjust the path for that, and add git commit and push [10:45] To test, add to wily: https://launchpad.net/~kubuntu-ppa/+archive/ubuntu/staging-frameworks/+packages ? [10:45] assuming you still have the temp folder you worked on till now [10:46] the src? [10:46] src/applications/ [10:46] mparillo: yep [10:46] right, /home/clivejo/src/applications/ [10:48] so grab a backup of that folder, just in case? [10:50] yeah, do that for testing [10:50] ok Ive got ~/backup/src/applications [10:53] should I remove the upload and manual folders? [10:54] Riddell, yofel, clivejo: random note, the commit for fixing the changelog message needs to be commited NOCI [10:54] Funny, krunner crashed as muon was applying the updates to frameworks. I did not even have krunner open. [10:55] Riddell: The testing approach is adding the staging-frameworks ppa and upgrade on wily? [10:55] oh great, the parentheses in the message break the regex, that's why the message is there multiple times [10:56] hurray to regexp [10:56] lol [10:57] :( [10:57] My Wily is still non functional :( guess I need to reinstall [10:57] I thought I was being good being more detailed in commit and changelogs [10:58] heh, well, you found a bug in the script ^^ [10:59] pure stupidity on my part [11:00] not really, when I wrote that check, I didn't think someone would do that because nobody usually does that [11:00] you were just the first one to do it ^^ [11:01] yofel: write a test! [11:01] that's actually a good idea..... [11:04] huh [11:04] ok I run that for loop Riddell posted [11:05] but run it on git/debian [11:05] some have worked, others not [11:06] if I used that script to run nano on the changelog instead and just fix the file manually? [11:06] actually, thinking about it, this is a bit crazy: re.match(".*New upstream release (15.18.1).*", re.match(pattern, changelog).group(0), re.DOTALL) [11:06] re.match(pattern, changelog).group(0).find("New upstream release (15.18.1)") might be less error-prone [11:10] OK Frameworks updates are complete. Re-booting. [11:10] thelionroars: want to test frameworks? [11:13] yofel: looks equally mental :P [11:14] sitter: well, I'm reducing the regex matching from 2 to 1. re.match(pattern, changelog).group(0) gives the full first changelog entry [11:14] shadeslayer: please merge your docker branch with logging additions I just did to kci/mgmt/docker.rb [11:14] also needs a test [11:21] ok, so I made a test folder [11:22] removed manual and upload folders [11:22] run for asdf in */; do cd ${asdf}; sed -i 's/15.18.1/15.08.1/g' git/debian/changelog; cd ..; done [11:22] I added the frameworks PPA to Muon Update Manager, applied the updates (random crashes); restarted (had to hard power off); and powered-on. Wireless, konversation, and rekonq seem to work. How do I know I really have the frameworks update completed? [11:23] the run the same again only running nano, manually editting the repeat lines from the changelogs [11:27] the slideshow font hinting is god awful [11:35] ahoneybun: https://goo.gl/photos/F53FVYuoUQdmxubb9 [11:48] ahoneybun: what happened to your muon slideshow update? I forget [12:00] Riddell: it has not been approved [12:01] ahoneybun: where is it again? [12:01] * Riddell pauses CI for frameworks upload [12:01] https://launchpad.net/~aaronhoneycutt/+archive/ubuntu/muon-discover [12:02] Riddell: NOCI? [12:02] git -a -m "NOCI" [12:03] ? [12:03] -m "uploading to archive\nNOCI" [12:07] Riddell: ^ [12:08] sitter: but what's your point? [12:08] you don't need to pause CI with that [12:09] good to know === rdieter_work is now known as rdieter [12:11] also you don't cause pointless builds [12:11] thinking about it, shouldn't staging upload use that too.......... [12:12] yofel: staging uploads increase the version [12:12] then why do we pause CI for that again? [12:13] so builds are run in order [12:13] ah ok [12:13] since the sources are processed in an arbitrary order it might trigger plasma-framework before ECM which might then fail [12:14] with pause you'd block all builds and then release them all at once letting jenkins build them accoridng to the dep tree [12:14] fancy version of the upload tiers we had 6 years ago [12:27] sitter: I pushed one of the fixes, does it look ok? http://anonscm.debian.org/cgit/pkg-kde/applications/akonadi.git/commit/?h=kubuntu_stable [12:29] clivejo: that NOCI went wrong xD [12:29] Riddell: Upgraded to the framework staging ppa, works fine. [12:30] * clivejo cries [12:31] sitter: whats gone wrong now? [12:31] apparently -m doesn't like nonprintable characters [12:32] lordievader: lovely thanks :) [12:32] mparillo: how was it for you? [12:32] [12:38] bash never helps. [12:32] to quote myself here [12:35] lots of lessons today! [12:36] clivejo: so just pause it? [12:36] that solves nothing [12:37] only queues 300 more builds [12:38] not that it matters because Riddell pushed without noci anyway [12:38] so it's busy until saturday anyway [12:38] also failing [12:38] I pushed one as test [12:39] can I revert the commit locally and fix it? [12:40] fix the message I mean [12:49] Hey all [13:21] Riddell: i have installed Framewroks 5.14, during unpacking packages kdeinit5 and krunner crashed but they start again [13:21] Riddell: but all update was just fine [13:36] but annoyingly there seems to be a problem in kservice [14:03] ahoneybun: kubuntu-settings uploaded! [14:33] yofel: did you say you were applying a fix so that the script wont add dupes like this again? [14:49] clivejo: yes he did === dgm816` is now known as dgm816 [15:21] Riddell: would you add http://qa.kubuntu.co.uk/ppa-status/applications/build_status_15.08.1_wily.html to the topic please [15:22] clivejo: I expect you can [15:22] have you tried? [15:22] hello Mandarina [15:23] me? [15:23] I dont have the pwer [15:23] Riddell: Hi! [15:24] Mandarina: want to become an elite kubuntu ninja? === Riddell changed the topic of #kubuntu-devel to: Kubuntu - Friendly computing | https://trello.com/kubuntu | http://qa.kubuntu.co.uk/ | apps 15.08 http://qa.kubuntu.co.uk/ppa-status/applications/build_status_15.08.1_wily.html [15:25] Riddell: I'm not very sneaky. [15:30] Mandarina: that's ok, we do everything in the open :) [15:31] Riddell: kde4libs is in manual, but cant see in in the depot download? [15:31] or has it been renamed kdelibs-4.14.12.tar.xz [15:33] it always is kdelibs, we rename it in debian/kubuntu [15:33] oh wait, I remember talk about this before [15:37] so I bump the changelog to kde4libs (4:4.14.12-1ubuntu1~) UNRELEASED; urgency=medium [15:37] push to debian git and ppa-build? [15:48] clivejo: 4:4.14.12-0ubuntu1 in changelog [15:48] Riddell: when I try to "git-buildpackage-ppa" kopete I get an error - Error: Distribution is not "UNRELEASED". [15:49] yes keep it as UNRELEASED [15:50] I didnt change it [15:51] regarding kde4libs - tar: /home/clivejo/src/applications/manual/kde4libs/build-area/kde4libs_4.14.11.orig.tar.gz: Cannot open: No such file or directory [15:52] why is it looking for 4.14.11 when changeslog says .12 [15:52] git-buildpackage is also spooky voodoo that likes to do that for some reason [15:52] but it should also get the .12 version [15:53] its there [15:53] I have to modifty the watch file [15:53] the watch file is looking for kde4libs but the filename on depot of kdelibs [15:54] uscan is showing the new version and getting it [15:54] ah [15:54] then do fix that please [15:55] does git-buildpackage-ppa still go back to debian git? [16:03] Riddell: in your changeset http://anonscm.debian.org/cgit/pkg-kde/applications/kde4libs.git/commit/?h=kubuntu_wily_archive&id=2b026e149f0191a257483af8cb5419ece90f1a3c [16:03] you set the distribuation to wily [16:03] clivejo: la la la [16:04] how do I fix it? [16:04] clivejo: what's wrong with it? [16:05] clivejo: just add a new changelog entry with dch [16:05] shouldnt it be UNRELEASED;? [16:08] on your new entry yes [16:08] Ninja voodoo [16:10] ok I think thats fixed [16:10] and uploaded to PPA [16:14] kde-baseapps is the same [16:14] why is this? [16:17] clivejo: they're in maintinance mode, no new features only bugfixes, so they don't get a new version number only the bugfix version changes [16:18] but how did they get the release as wily in debian git? [16:20] clivejo: the staging-upload script that you used should add the new changelog entries with UNRELEASED [16:20] clivejo: then when it's all built and tested the next script is ubuntu-archive-upload [16:20] which sets the changelog to wily [16:21] it passed over these [16:21] put them into manual [16:26] oh it probably needs packaging-exceptions.json update first, I'll add that to README [16:30] Riddell: have to go for a bit, will you be online later? [16:31] clivejo: I'm out for the next 3 hours, probably back about 9 [16:31] unless I'm not [16:31] LOL [16:31] ok [16:31] have fun on the farm :) [16:31] indeed [18:29] thanks Riddell also need to update this page desc: https://launchpad.net/ubuntu/wily/+source/kubuntu-settings [18:29] remove active and netbootk [18:29] *netbook [18:30] Riddell yofel: Having problems with kmailtransport, can someone look at the error message and let me know whats wrong - https://paste.kde.org/pgph2lscn [18:30] ahoneybun: that's extracted from the package [18:30] so it seems we still ship that.. [18:30] it's not in the package [18:30] so it can't be [18:31] unless it is in a file I did not edit/know about [18:32] they might be NBS.. [18:32] yeah [18:32] kubuntu-settings | 1:15.10ubuntu1 | wily/universe | source [18:32] kubuntu-settings-netbook | 1:15.04ubuntu2 | wily/universe | all [18:32] needs a binary removal request I think.. [18:33] mm [18:35] clivejo: that looks very wrong [18:35] dpkg-source: info: local changes detected, the modified files are: [18:35] kmailtransport-15.08.1/libkf5mailtransport5.symbols [18:35] yofel: http://anonscm.debian.org/cgit/pkg-kde/applications/kmailtransport.git/tree/?h=kubuntu_wily_archive&id=abaf052b86e579b56e816c48df774d2e14abd574 [18:35] symbol file in wrong folder? [18:35] why isnt the symbols file in debian/ [18:35] whoever put it there messed up? [18:35] how do I find out who did that? [18:36] http://anonscm.debian.org/cgit/pkg-kde/applications/kmailtransport.git/commit/libkf5mailtransport5.symbols?h=kubuntu_wily_archive&id=abaf052b86e579b56e816c48df774d2e14abd574 [18:36] :/ [18:39] just git mv the file [18:40] yofel: will you ask him, I think Ive damaged enough things today :( [18:40] just move the file, you can't really make unbuildable any worse :P [18:40] want bet [18:40] no [18:40] Im having a very bad day [18:40] ^^ [18:47] kmix has a problem with a patch [18:48] https://paste.kde.org/pujddxv7k [18:50] drop the patch, upstream_ shows that it's a cherry-pick from kde git, and it's now included in the release [18:51] yeah kmix no longer shows up in the system tray, it's been replaced by that pulseaudio volume ctrl which does nothing since i don't use pulseaudio. I have kmix set to show ithe system tray, but it doesn't. [18:54] should I physically delete the file and series entry or just comment out the series entry? [18:54] git rm the file and remove it from series [18:54] ok [18:55] and say in the changelog that you're dropping , applied upstream [18:55] sorry, which patch kmix_showeverywhere.diff or upstream_fix-kmixui.rc-install-location.diff ? [18:56] I was talking about the upstream_ one, wasn't I? [18:56] and the log shows that it was the patch that it failed on [18:56] sorry, misread [18:57] we usually prefix integration patches with kubuntu_ and upstream fixes with upstream_ and add DEP-3 headers for documentation [18:58] also always put the filename in the changelog when you modify patches, so it's easy to track modifications to them later on [18:59] esp. when adding/removing them [19:02] yofel: http://anonscm.debian.org/cgit/pkg-kde/applications/kmix.git/commit/?h=kubuntu_wily_archive&id=e5fe876d059d6a4090b30ba56651b5e51b90b52a hows that? [19:02] good :) [19:03] learned a lot today [19:03] after that stupid typo I made last night [19:04] I think thats all the apps uploaded bar kde-l10n [19:05] that shouldn't even be there.. I guess it can't keep files and folders apart when listing the packages :/ [19:07] it didnt do gpgmepp, killbots, syndication for some reason, they all still version 15.08.0 [19:08] are they in manual? [19:08] nope [19:08] o.O [19:08] do you still have the script outoput? [19:09] Ive fixed all the manual ones bar kde-l10n [19:09] or rather, can you re-run the script on those packages? [19:09] oh wiat [19:09] they are in applications [19:10] but didnt get uploaded [19:10] empty sources? [19:10] dunno, if they're not in manual then something happened that the script has no error handling for :/ [19:10] yeah strange [19:11] might have been my internet [19:11] killbots just dput fine [19:12] hm, tarball fetch failures are handled though [19:12] are dput errors caught? [19:13] no, that's not part of the script [19:13] but if the upload failed there will be no .upload file [19:13] right, Ive reloaded those to LP [19:13] if the upload file was missing then it was an upload failure [19:15] rejected by LP [19:16] huh, why? [19:16] syndication_15.08.1-0ubuntu1~ubuntu15.10~ppa1.dsc: Version older than that in the archive. 15.08.1-0ubuntu1~ubuntu15.10~ppa1 <= 4:15.08.0-0ubuntu1~ubuntu15.10~ppa1 [19:17] urgh, the old packages are still in the PPA [19:17] you'll have to delete those by hand [19:18] if I upload PPA2? [19:18] no, no matter what you upload, the epoch will conflict [19:18] you'll have to delete the packages from the PPA in the LP UI, then later upload again === rdieter is now known as rdieter_work [19:36] yay the installer did not crash on first boot [19:37] Riddell: do you know if sitter is going to make anymore changes in the installer or is it frozen? [19:38] we can do bugfixes, but we're past UI freeze, so no visible changes anymore [19:38] alright cool [19:38] Riddell: did you moved 5.14 to wily ? [19:38] thanks yofel [19:42] redshift won't show unless installed with muon-discover [19:43] according to a report [19:44] what's happening with apt ? [20:07] yofel: its seems to be a epoch issue [20:07] I said that it's an epoch issue... [20:08] you can't upload the packages because the epoch of the packages in the PPA is higher [20:08] didnt spot that [20:08] that's why you first need to manually delete them [20:09] should I not add the epoch to this one? [20:09] no, those packages should not have an epoch [20:09] isnt syndication 4:15.08.0 in backports? [20:10] dunno, I certainly hope not [20:11] that's why you should not backport the staging packages :/ [20:12] it's not, 15.08 was never backported [20:13] so the only people that will have upgrade issues are the wily staging PPA testers [20:13] acceptable collateral damage [20:13] http://imgur.com/cS1Denm [20:16] who is apachelogger? [20:16] sitter [20:17] I dont feel comfortable deleting packages uploaded by Riddell and sitter [20:17] Ill leave those few packages until another time [20:18] just delete them if they're 15.08.0 [20:18] if you want, I can also completely wipe the whole PPA if you want to re-upload everything [20:19] there are 15.07.90 versions of gpgmepp [20:19] the only thing I'm trying to teach you is that epochs are a very scary thing and that you shouldn't use them unless absolutely necessary [20:20] that's even older, kill it [20:20] I know, thats why I dont want to touch anything [20:21] thinking about it, our epoch handling is completely wrong [20:21] listing packages without an epoch means that any unknown, new packages automatically get an epoch [20:21] which is the opposite of the intended behavior [20:23] but why would Riddell and sitter upload versions with an epoch? [20:23] when it doesnt need it? [20:23] read what I just said [20:24] our script didn't know those packages, so it just added an epoch because they weren't in the no-epoch list [20:24] ohhh [20:24] but new sources are not supposed to have an epoch [20:24] thats dangerious [20:24]