[01:11] my theory about kalzium is right, it's the latest avogadro what broke it [01:13] clivejo: if you are still awake just ignore kalzium, for now I have a solution but I want to discuss it with yofel tomorrow [04:19] clivejo: print-manager and kde4libs ftbfs'es fixed in my merge requests, now the only thing not building yet in apps is kalzium. as I said we could discuss the solution for that tomorrow [04:23] this jenkins thing is pretty crazy [04:25] ? [04:26] the way it works I mean [04:26] seems to be widely adopted [04:26] KDE uses it as well [04:27] off to bed, my plan is to fix something tomorrow in Kubuntu unstable land as I haven't done much lately, suggestions welcome :P [04:27] o/ [04:27] sweet dreams, tsimonq2 [04:28] I know just hard to understand valorie [04:28] heh, there I agree [04:28] which is why I don't even try [04:28] :-) [04:29] I do want to try [04:29] but next cycle lol [04:33] I try to get the slideshow working with yofe first valorie lol [04:33] that's awesome [04:33] thank you for stepping up [04:34] TRY [04:34] lol [04:34] I got the python script to at least start but not sure what it means [04:40] maybe #python can help? [04:41] it's dealing with our installer [04:42] I've changed and added a few things to get it to start with Qt5 [04:42] but still no images yet [04:42] wait what [04:43] that's weird [04:43] it works in test-slideshow.py [04:44] http://imgur.com/a/l6AS4 [04:44] valorie: ^ [04:46] nice [04:47] no clue if it works for real [04:48] my branch is on LP if any MOTU or whatever wants to test it [04:50] I want to add a few screenshots [04:50] be sure to post it to the list [04:51] maybe there are more python-lovers there [04:56] http://imgur.com/a/l6AS4 weird yofel [04:58] you notice that says "16.04", right? [04:58] yep fixed [04:58] I'm thinking of changing the text anyway [04:58] somehting like [04:59] "A community driven Ubuntu based distro that integrates the Plasma desktop" [04:59] bit of a rip from Ubuntu Mate really [05:01] How about A community produced Ubuntu flavor with the Plasma desktop and KDE applications [05:05] thats better [05:20] I [05:20] I'm 75% sure the slideshow will work now [05:22] \o/ [05:44] installing some updates and rebooting, wish me luck [05:44] clivejo: ^ [05:44] another no-drama update [05:44] :-) [05:45] hmmm, about system says Plasma 5.6.3 [05:45] but the loading screen says 5.7.2 [05:46] mine said 5.7.2 I'll see after reboot [05:47] yofel: getting a lot of output from fakeroot as per https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=830912 [05:47] Debian bug 830912 in fakeroot "fakeroot complains about missing acl_* symbols" [Normal,Open] [05:47] right, you have the staging PPAs [05:48] well that is nice to see lol [05:50] I have all staging [05:59] Rebooting now [06:02] and I'm back [06:02] with no issues [06:03] valorie: still says 5.7.2 in About System [06:11] git-buildpackage-ppa -d yakkety -y 16.10 -s 51 [06:11] Traceback (most recent call last): [06:11] File "/home/ubuntu/kubuntu/kubuntu-automation/git-buildpackage-ppa", line 63, in [06:11] link_upstream_tarball(release_type) [06:11] NameError: name 'release_type' is not defined [06:12] ^^ yofel anyone? [06:33] I guess the fakeroot output is result of new glibc 7hrs ago! [06:46] newest blocker to clearer visibility on transition in yakkety is pillow synced from Debian + a 1000 packages queue for each architecture thanks to new glibc [07:00] Mirv: when is yakkety release? October? Maybe we can just about get everything migrated by THEN! grrrr. lol [07:02] acheronuk: October, Final Freeze is on Oct 6th, maybe by then! [07:02] it's not hopeless but I've lost most hope ;) I'd at least want to even see the situation of update_output.txt when everything would be "stable" for a brief moment. [07:03] because it's so interesting that eg Laney claims WebP would have been the last blocker, but it's hard to validate since there's "something" happening all the time. [07:03] and indeed I'd have more Qt updates pending myself, but I've not published them because that again would case 1-2 days delay at minimum [07:06] well I have to take their word for it on the blockers, but too many false dawns already for me to have confidence [07:09] ok. KA is still a bit busted, even after reverting a commit [07:09] Successfully signed dsc and changes files [07:09] Traceback (most recent call last): [07:09] File "/home/neo/kubuntu/kubuntu-automation/git-buildpackage-ppa", line 88, in [07:09] upload_area_dir = readConfigFile("areas.json")["upload-area"] [07:09] NameError: name 'readConfigFile' is not defined [07:09] Still get that ^^^^ [07:11] santa_: were those KA changes tested in, or require anything specific to work? [07:11] as it's pretty busted here [07:24] on my local copy of kubuntu-automation I have reverted the whole of https://git.launchpad.net/~kubuntu-packagers/+git/kubuntu-automation/commit/?id=757851f458d6bb6c5fa45a31b9ab7f6d05a1cb07 [07:56] yofel et al : those fakeroot errors/messages seem to just be noise as far as a build goes, but they confuse the hell out of the ppa QA page. i.e. goes red [08:05] pbuilder run of okular changes produces them in quantity [08:10] !info snap-confine [08:10] snap-confine (source: snap-confine): Support executable to apply confinement for snappy apps. In component main, is optional. Version 1.0.38-3 (yakkety), package size 21 kB, installed size 77 kB [08:38] * acheronuk gives up [08:38] back sometime in the afternoon probably [09:59] santa_: git-buildpackage-ppa is assuming that the clone of the kubuntu packaging for example of okular has the folders /okular/git/ , where the packaging is in the git folder [10:00] santa_: whereas, much of the time we just 'git clone kubuntu:okular', so there is no git folder [10:01] santa_: hence git-buildpackage-ppa fails as it it looking for the package name one folder too low in the tree [10:03] santa: so needs so logic to sort that depending on whether the immediate parent folder is 'git' or something else? [10:09] i.e upstream_name = os.path.basename(os.path.split(os.getcwd())[0]) [10:10] needs to be changed to take account of whether the git clone is in a directory with is the package name itself, or in a 'git' subdirectory of that [10:13] FWIW, you could actually parse that out of the VCS-Git field [10:14] as that's not any different from looking at the folder name in any common case [10:14] or from the watch file [10:15] which is probably even more accurate (would even catch kdelibs) [10:15] but has an annoying syntax [10:19] maybe git-buildpackage-ppa should be moved out of automation into kubuntu-dev-tools [10:20] that's a required piece for the automation, but it's not supposed to be *exclusive* to the automation, but usable with any kind of git based package [10:29] that reminds me. on kubuntu-dev-tools, I think the setup.rb script now needs a small tweak to fix build/install [10:31] could be, I haven't tried to build that in ages [10:36] who merged santa_'s PR? [10:39] never mind [10:39] * clivejo is going crazy === santa is now known as Guest50613 [10:50] clivejo: well, it *did* work for the most common use case, and he said he would fix it up soon - hasn't happened yet :/ [10:54] yofel: sorry I was talking about the apps PR [10:55] oh [10:58] but I was looking at the wrong webpage [10:58] I thought someone had merged it, but not done an upload to PPA [10:58] and was getting on my high horse [10:58] is there a way to find out who actually merged the request in? [10:58] git only seems to record the person who submitted it [10:59] did I leave the channel just then? [10:59] not unless launchpad tracks who pushed the merge [11:00] not this one [11:00] (channel) [11:00] strange [11:00] must have lost internet connection [11:01] [11:58:19] <-- clivejo (clivejo@kde/community/clivej) has quit (No Ping reply in 180 seconds.) [11:01] [11:59:21] --> clivejo (clivejo@kde/community/clivej) has joined #kubuntu-devel [11:01] o.O [11:01] your bouncer connection might've died [11:01] that looked like my bouncer dropped [11:04] * clivejo volunteers yofel to do a lightning talk at QtCon [11:04] oh yeah, I saw that mail. Not going to happen ^^ [11:04] LOL [11:04] I just wanted to see your reaction :P [11:08] anyone else finding that LP randomly ignores some upolads? === Guest50613 is now known as santa_ [11:19] good "morning" everyone [11:19] o/ Hey santa [11:20] hi santa_ [11:20] acheronuk: so this is your only problem https://paste.kde.org/p8oewochl right? do you have any other issues? [11:21] santa_: Im having the same problem [11:21] sorry for the inconvenience by the way, I didn't expected that use case but I will come up with a fix soon [11:21] and also even with the directory cloned into the corrent package name, there are some exceptions [11:22] to be honest I would prefer something like what yofel talked about [11:22] by picking up the package name via the changelog? [11:23] kde4libs is one that throws errors [11:57] technically, only the watch files know the *real* upstream name, which is exactly why the script used uscan. [11:57] Now you can make it only work for kde - and support depot, or you leave it as it is and have it fail in some situations [11:58] so maybe there should be a generic git-buildpackage-ppa in kubuntu-dev-tools using uscan, and git-buildpackage-kde-ppa in automation which is tuned to work for kde in all possible cases [12:24] santa_: it's the only one I've bumped up against so far, but I'm only just getting used to using the tools, so you never know... [12:25] clivejo: I merged nothing [12:25] clivejo, acheronuk, yofel: possible fix here https://code.launchpad.net/~panfaust/+git/kubuntu-automation/+ref/work3 [12:26] note that the work3 branch has other stuff we discussed here [12:27] what I didn't implemented yet are the overrides for the versions maps, but I hope to work on that soon [12:28] yofel: thanks for the watch idea btw, I was thinking about the copyright files or that new metadata, but right now I think the watch file is the best option [12:29] there isn'n any "combo breaker" in frameworks/plasma/apps if I'm not mistaken [12:29] I will test it better after lunch anyway [12:32] Hiyas all [12:34] yofel: the kubuntu-dev-tools change was trivial http://paste.ubuntu.com/23049303/ [12:34] but ir did allow me to install them, rather than run in place [12:34] *it [12:37] acheronuk: committed, thanks [12:39] santa_: and the script now also works if the download location is not download.kde.org? i.e. if upstream_name is set to nothing? [13:02] yofel: if that's the question, with the fix in work3 you must have at least one download.kde.org line in debian/watch [13:03] yofel: if the question is if it works with depot.kde.org, not completely yet [13:13] yofel: hmm, just forget that last thing I wrote, if I'm not mistaken gbp-ppa should get the tarball from depot if there isn't already downloaded [13:14] * if it's not already downloaded [13:14] that was done in 84de65dc9ffd15e84a5e6e5ab07a1346e9431d3b which is already on the master branch [13:15] what it's not done yet is: falling back to uscan/download.kde.org if you don't have you ssh key on depot [13:32] santa_: my point is, if your lib cannot generate a valid download location, please fall back to uscan [13:32] so that the script works like it did before [13:34] santa_: and I probably should read all your messages before replying ^^ [13:34] santa_: so yeah, that sounds right, thanks [14:31] santa_ yofel: whats going on here - https://launchpadlibrarian.net/278484268/buildlog_ubuntu-yakkety-amd64.krfb_4%3A16.04.3-0ubuntu1~ubuntu16.10~ppa51_BUILDING.txt.gz [14:31] dlsym(acl_get_file): /usr/lib/x86_64-linux-gnu/libfakeroot/libfakeroot-sysv.so: undefined symbol: acl_get_file [14:32] there are a number of amd64 builds showing as red due to that same thing [14:32] Rik found that this morning as well [14:32] glibc fallout I believe [14:33] the release team is also really enjoying the transition output right now [14:42] I think that's something out of our "jurisdiction" [14:42] ftr, so far it doesn't happen in my build box [14:43] yofel: ack, to your uscan proposal for gbp-ppa, I wil try to patch that soon [14:44] but, before anything there's a couple of topics I would like to discuss with you guys [14:44] first and most important one is the kalzium ftbfs [14:44] the situation is the following [14:45] kalzium build depends on avogadro [14:45] avogadro build depends on eigen [14:45] eigen has 2 versions in yakkety: eigen2 and eigen3 [14:46] doko uploaded recently a new avogadro to yakkety [14:46] this new avogadro was imported from debian experimental [14:46] this new avogadro includes a patch from debian to make it build against eigen3 and not eigen2 [14:47] this patch is making kalzium FTBFS [14:48] if you build kalzium against an avogadro without the eigen3 patch it doesn't fail [14:48] ↑ those are facts [14:48] now, my proposed solution for this: [14:49] I think doko's intention was *just* fixing the avogadro build failure against GCC 6 [14:50] I also think including an avogadro with debian's eigen3 patch is a bit premature for ubuntu [14:50] so ... [14:51] my proposal is disabling the eigen3 patch from avogadro [14:51] I have a package here doing that http://gpul.grupos.udc.es/tritemio/pool/main/a/avogadro/avogadro_1.1.1-1~exp3.1ubuntu1.dsc [14:52] kalzium builds fine against that package [14:52] yofel, clivejo: ↑ what do you think? [14:58] who maintains it? [14:58] avogadro? [14:58] yes [14:59] avogadro is usually imported from debian unstable but doko uploaded a fixed version imported from debian experimental [15:20] But noone here will have permissions to remove the patch [15:25] this lack of permissions of everyone is going to kill this project [15:25] anyway. I think the best approach would be mailing doko about the issue [15:26] explaining what I explained above [15:51] clivejo: yofel: fakeroot update https://launchpad.net/ubuntu/+source/fakeroot/1.21-1ubuntu2 [15:52] allright, so eigen2 doesn't build with gcc 6 [15:52] hence why doko included the eigen3 patch from debian [15:53] yofel was there a way you could see if the fix worked? [15:54] ahoneybun: no, still at work. I'll be home in ~4h [15:54] oh no I just wanted to know if there is a way to test it unless we have to push it to the main [15:58] I think for a test, you can build your modified package, install it in a yakkety live session and run the installer. It should pick the slideshow up as long as it's installed [15:59] do kalzuim devs have any plans/fixs? [16:00] doubt it, but you never know [16:02] yofel: santa_ cmakelists.txt metions Eigen3? https://quickgit.kde.org/?p=kalzium.git&a=blob&h=050f6610dc99caba3f40df06bdf7d2714161ad1a&hb=ad24b0f296504a22f0c54d17eb3f6b3252eca143&f=CMakeLists.txt [16:10] ok. so kalzium has build deps on libeigen3-dev, but FTBFS if avogadro is built with v3 instead of v2. that is nuts [16:17] mm [16:17] MM [16:17] :P [16:18] I just wonder how to build it [16:18] * tsimonq2 throws ahoneybun at a giant iPhone-sized cake [16:18] how is it giant and iphone size? [16:19] ahoneybun: this big: https://www.youtube.com/watch?v=uNferthlGeQ [16:19] :P [16:19] oh [16:19] I want a piece :P [16:20] bbl [16:20] ahoneybun: don't want to fight today? :P [16:20] o/ yofel [16:21] not really just work up a bit agao [16:21] *aog [16:21] ago [16:21] switching kalzuim build deps to libeigen2-dev lets it build, but with the 'compound viewer' usr/lib/libcompoundviewer.so.4 absent from the result [16:28] trying to build it now yofel [16:37] yofel: building kalzium with libeigen2-dev you lose the molecular editor then libeigen2-dev http://i.imgur.com/d79XpLJ.png [16:39] I've only really used the isoptopes and Nuclear Calculator on that, so not sure how much losing that would matter to a chemist. lol === mgolden is now known as mgolden_ [18:06] stupid debsign [18:09] ahoneybun: PEBKAC [18:09] :P [18:09] ? [18:10] ahoneybun: Problem Exists Between Keyboard And Chair [18:10] is there a way to skip it? [18:10] what command are you running? [18:16] debuild [18:17] it's for my personal use atm so don't need to sign it [18:17] * tsimonq2 does debuild -us -uc (-S) [18:18] mm just running debuild again [18:18] I think I put in the wrong PW tbh [18:29] testing now yofel [18:55] acheronuk: the result f your experiment is interesting, however I think I have now the actual solution [19:04] mm virtualbox is not working for me [19:04] I can't get my usb to show up [19:05] virt-manager ftw [19:06] santa_ yep, building that way was intended as a last resort option [19:22] mm might have broken something [19:30] tsimonq2: your going to hit the FF trying to geth LXQt in no? [19:33] mm this is not working... [19:34] ahoneybun: LXQt is already in the archive... [19:35] oh did not know [19:35] sudo apt install lubuntu-qt-desktop works fine [19:35] just waiting on an image... [19:35] this script is looking for a dir that is not there [19:35] * acheronuk tries sudo apt install lubuntu-qt-desktop [19:36] acheronuk: stop stop stop [19:36] s/apt/apt-get/ [19:36] acheronuk: if you have plasma, it doesn't like things [19:36] acheronuk: and no, use apt, apt is better than apt-get [19:36] * tsimonq2 said that intentionally :P [19:36] acheronuk: VM ftw! \o/ [19:37] acheronuk: and things need to be tweaked a LOT [19:37] tsimonq2: I was seeing what it wanted to do. Was not going to actually say yes to the install [19:37] mm there is a arm ubiqity installer? [19:38] right Ubuntu Mate is on arm [19:38] oh k acheronuk [19:38] clear [19:38] is there a tasksel target for netinst? [19:39] yofel: I'm 99% it is not fixed [19:39] need to fix the kde_ui.py [19:39] tsimonq2: not that I could have, anyway http://paste.ubuntu.com/23050423/ [19:40] * ahoneybun wonders why LXQt does not use the Breeze icon set [19:40] yeahhh acheronuk [19:40] * tsimonq2 throws ahoneybun down a hole [19:40] :P [19:41] it just looks better then anything GTK has [19:41] GNOME I mean [19:41] * acheronuk wonders how many packages are stuck in proposed now? 1,000? 2,000? [19:41] loooots probably [19:43] britney tries to migrate 656 (i.e. not counting 'not considered') [19:59] acheronuk: thanks for taking the time to do the experiment anyway, thanks to that first thing I checked when testing the fixes was that compoundviewer is still built [19:59] so ... [20:00] yofel, clivejo, acheronuk: I think this is the correct solution http://gpul.grupos.udc.es/tritemio/pool/main/a/avogadro/avogadro_1.1.1-1~exp3.1ubuntu1+santa3.dsc [20:00] I just fixed a couple of loose ends with avogadro's eigen3 port [20:00] forget about switching avogadro back to eigen2 because eigen2 doesn't build with GCC 6 [20:01] I'm doing now my last build tests, if everything goes well I would send doko a mail asking him to upload this fixed version, unless someone has a better alternative [20:03] * ahoneybun hangs his hat on fixing the slideshow... [20:04] I might be able to upload avogadro [20:04] I'm sorry yofel and valorie [20:04] it's so above me [20:05] np, take your time if you want. There's a reason why people are reluctant to tackle this ^^ [20:05] so really, thanks a lot for trying [20:06] The slideshow thing works preview wise in my repo [20:06] Just need fix in the KDE_ui.py [20:09] santa_: that actually works? [20:09] 0 +find_package(Eigen3 REQUIRED) [20:09] 1 include_directories(${EIGEN2_INCLUDE_DIR}) [20:09] The ubiquity package on lp is missing some files [20:09] I looked at the porting guide and decided I didn't know enough to do it [20:10] You need to use apt source ubiquity to get all of them [20:19] yofel: yep [20:20] but let me recheck the patch [20:24] yofel: well, I'm surprised it works but it does. anyway I have just uploaded a new version also fixing the include dir, once it's built I will retest kalziumç [20:24] thanks for pointing that out [20:39] yofel: the error is about wpa [20:49] is the new fakeroot in proposed? [20:50] yes [20:50] NC rebuild should fix those red ones failing on it then? [20:51] it should [20:52] * clivejo uploads okular ppa52 [21:18] how long has gcc6 been out? [21:20] not veery much, I think a few days. if with "being out" you mean being the default c/c++ compiler for yakkety [21:20] I mean been released? [21:21] surely they knew it would be default in yakkety and should have had it ready when toolchain was created? [21:22] well, we should have fixed our build failures in advance like we did with gcc 5, but we were busy with other stuff [21:24] it was available for testing builds for month, there's even a mail from doko in kubuntu-devel about it [21:24] * for months [21:25] clivejo: btw are you using the fixed gbp-ppa from my work3 branch? [21:25] if you find more issues let me know, so we can get them fixed [21:25] santa_: yes, I merged them in locally and working good [21:25] nice to know [21:26] I was getting rather fed up building packages manually! [21:26] git-buildpackage-ppa is very nice wee tool [21:35] yofel: http://gpul.grupos.udc.es/tritemio/pool/main/a/avogadro/avogadro_1.1.1-1~exp3.1ubuntu1+santa4.dsc [21:36] ↑ in this one I also fixed the include dir [21:36] it builds and doesn't make kalzium FTBFS [21:46] breaking mergers left, right and centre [21:48] yofel: about gbp-ppa retrieving the tarballs, it seems I already did what you suggested today; I made the following test: I commented out my depot config on ssh/config, tried to build kjs and the tarball was downloaded with uscan [21:49] great, thanks [21:50] about avogadro, are you going to upload it? [21:52] doing another testbuild just for the sake of it, then yes [21:53] great! you just have to strip the "+santa4" from the changelog [21:55] already did ;) [21:58] santa_: uploaded [21:59] it would be nice if you could paste the changes into a debian bug report [23:45] * clivejo is having internet problems [23:46] or weegie is down [23:46] not sure [23:46] anyone seeing this can you ping me? [23:46] clivejo: o/ [23:47] ah hi simon! [23:47] I saw all four messages from * clivejo is having internet problems [23:47] what does the t in your name stand for? [23:48] I have to tell you via PM