=== ezri is now known as rww [10:32] Good morning. [10:34] morning lordievader [10:35] Hey clivejo, how are you? [10:35] good thanks, and you? [10:36] Doing okay here :) [11:02] yofel: would you ping me when you are free [11:02] I think I have done something bad to kio [11:31] clivejo: what's the problem? [11:32] I dont know :( [11:32] Ive confused myself again [11:32] clivejo: you can use other lib package as a tempalte [11:33] *template [11:33] I tried to fix the unstable branch [11:33] and realised there was no stable branch [11:38] this one package is breaking about 100 others :( [11:39] http://kci.pangea.pub/job/xenial_unstable_kio/ [11:47] clivejo: show me control file [11:48] http://anonscm.debian.org/cgit/pkg-kde/frameworks/kio.git/tree/debian/control?h=kubuntu_unstable [11:55] looks fine [11:56] KCI-W :: Patch fix_kfreebsd_build in VCS but not listed in debian/series file. [11:56] its commented out [11:56] should I remove that patch totally? [11:57] I dont know what that lib does, so the description is just copied [11:57] its new for 15.19.0 [11:57] which is available in depot now [12:03] clivejo: this patch is not needed in ubuntu [12:04] grrr [12:04] what have I done! [12:04] m? [12:05] I seem to have merged with master somewhere [12:05] and maybe everything is broken because of two new packages libkf5kiogui5 and libkf5kiogui-dev [12:05] http://anonscm.debian.org/cgit/pkg-kde/frameworks/kio.git/tree/debian/changelog?h=kubuntu_unstable [12:06] how did that first changelog entry get there [12:06] Hiyas all [12:07] this is xenial branch changelog - http://anonscm.debian.org/cgit/pkg-kde/frameworks/kio.git/tree/debian/changelog?h=kubuntu_xenial_archive [12:12] UNRELEASED is not what we want here :) [12:13] Im gonna go clear my head [12:13] so confused [12:53] hi folks. there's pyqt5-dev-tools, which installs pyuic5 and pyrcc5, and which depend on python3, but there's no similar packages for developers working with python2 (those scripts would work under both python2 and python3 i think). [12:54] i'm packaging my python2/pyqt5 app in a docker container, and this forces me to pull in python3 just to be able to run pyuic5 and pyrcc5 as part of my packaging. [12:55] shouldn't there really be a python-pyqt5-dev-tools (for python2) and then the current pyqt5-dev-tools should be renamed to python3-pyqt5-dev-tools? [12:57] (as you can see, pyqt5-dev-tools is currently the only package that provides these scripts: http://packages.ubuntu.com/search?searchon=contents&keywords=pyuic5&mode=exactfilename&suite=wily&arch=any , leaving python2 devs having to install python3) [12:58] in other distros which doesn't use such aggressive package splitting, the scripts are installed by both the "python2 pyqt5" package and the "python3 pyqt5" package (with the scripts renamed to avoid conflicts). [13:05] estan: better be to address that to Debian Python Modules Team as they're original maintainers of this package [13:15] oh lord http://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-PRIME-Sync-V3 [13:24] blaze: ah yes. [14:47] blaze do you know how to do debian merges? [14:49] clivejo: you don't have access to their git, right? [14:49] I do [14:49] Ive done from kmine to konsole on this list [14:50] http://qa.kubuntu.co.uk/ppa-status/applications/build_status_15.12.1_xenial.html [14:50] Id like someone to check them over [14:50] oh, better wait for Philip [15:54] estan: Our original plan was not to provide PyQt5 for python2 at all since we figured people would be porting to python3 faster than it turned out to be the case. Please file a bug in the Debian bug tracker and we can take a look at it. [16:10] ScottK: ah. alright . [16:11] would love to go python3. what's holding us back for this project is VTK, which we don't want to build ourself. but 7.0 is in the process of being packaged, and supports Python 3. [16:12] just hope the debian-science-maintainers will make a python3 version of that package. [16:13] (e.g. "python3-vtk7") [16:13] If upstream already supports it, it's not usually that hard. [16:15] yea. i e-mailed the guy who's working on it and he seemed homeful they'll have it done at the same time as the other vtk7 packages. [16:16] but packaging that is going to take a while, so we need to workg against VTK 6.2 for now , which means python2. [16:21] *hopeful. [16:53] ScottK: alright, i filed https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=814023 [16:53] Debian bug 814023 in pyqt5-dev-tools "pyqt5-dev-tools should be offered in Python 2 and 3 variants" [Normal,Open] [17:08] Thanks. [18:51] anyone here? [18:53] Ive archived a Trello card and I dont know how I did that! [18:59] lol [19:02] not having a good day today [19:02] should have never got out of bed! [19:05] relax, just relax [19:06] think Ill go back to kolab [19:10] dont think I can break anything there! [19:18] With a laptop, there's no need to leave bed to cause problems ... [19:19] I dont like working in bed :P [19:22] ScottK would you mind checking my mergers? [19:22] Sorry, I don't know enough about the state of Kubuntu stuff to know if they're right. [19:22] If you have a specific question, I can try to answer it. [19:23] do you know why maxy has only updated some apps to 15.12.1 and others to 15.12.0? [19:25] ScottK: for example - http://anonscm.debian.org/cgit/pkg-kde/applications/zeroconf-ioslave.git/commit/ [19:26] http://download.kde.org/stable/applications/15.12.1/src/zeroconf-ioslave-15.12.1.tar.xz [19:26] Usually it's because there were no changes in the point release. [19:26] yet otheres are - http://anonscm.debian.org/cgit/pkg-kde/applications/umbrello.git/commit [19:27] oh right [19:27] so we package it regardless of changes? [19:28] Typically, but I'm not sure it makes sense. [19:28] clivejo: perhaps you accidentally hit c with a card selected and archived the card? http://webapps.stackexchange.com/questions/24271/how-to-undo-card-archive-on-trello [19:29] I think Kubuntu needs to decide if they want to sync from Debian when they can and how it impacts the VCS strategy. [19:29] god knows what I did! [19:29] Personally, I think you would be better off to sync as much as you can. [19:29] ScottK: is there a way to auto sync [19:30] If there's no Ubuntu diff, it's automatic from Debian Unstable prior to Debian Import Freeze. [19:30] After that or if there's a diff, then it's not automatic. [19:31] It's trivial for an developer to do. [19:31] Also it needs to be done manually from Experimental. [19:32] mparillo: can you see the card again, in backlog? [19:34] There are maybe a dozen archived cards. Or did you mean you already restored it to backlog? which card? [19:34] Debian Merges? [19:34] I think I restored it [19:34] I see that at the top [19:34] yup [19:34] ah good [19:50] estan: We looked into it and it turns out it's not easy to do what you want without breaking building of existing packages. Given your situation, you might want to look into equivs as a mechanism to avoid python3 for now. It's a gross hack, but it may be just the gross hack you need. [19:57] ScottK: aha. even making python recommended instead of required is problematic? (see my followup correspondence with Dimitri). if so, then i'll look at equivs (hadn't heard of it, thanks). [19:57] It is. I've been reading the bug. [19:57] okay. yea i can see how things would break. [19:58] the "clean" solution imho would be to have separate packages, e.g. pyqt5-dev-tools -> python3-pyqt5-dev-tools and then a new python2-pyqt5-dev-tools, but Dimitri seemed not so amused with that idea (don't blame him :)). [19:59] i mean, if you're going to split these commands into a separate packages, then separate packages for python2 and 3 would be the logical conclusion. [19:59] *a separate package. [20:00] Right, but then duplicating the C++ parts would be weird. [20:01] ah yes. can't really give those pythonX- prefixes :/ [20:05] What we discussed was making two new binary packages for the python bits that they existing one would depend on. [20:05] That would work, but it's excessive for a three line wrapper. [20:05] yea. i can understand that. [20:06] it's just, to me, having to install python3 just to run a binary (!) is excessive, heh. [20:06] and i guess changing the dependency to python2 is out of the question? that has some logic too, it's the lowest version of python that is actually required by the scripts. [20:06] (that's what e.g. Arch does: https://www.archlinux.org/packages/extra/x86_64/pyqt5-common/) [20:07] No. Arch is fundamentally insane when it comes to Python. [20:07] They made /usr/bin/python point to a python3 version and that's nuts. [20:08] It's entirely their fault it was necessary to invent /usr/bin/python2 to work around their idiocy. [20:08] yea. i've heard all about that. [20:08] (anyway, i was wrong. they don't even register a direct dependency on any python for that package.) [20:10] but alright. i'll live with this until we can move our project to python3, or i'll look at equivs. [20:11] should be able to switch as soon as a python3-vtk7 is created (in the works). === Blizzzek is now known as Blizzz [20:13] should i close the bug with a reference to this conversation, or wait for some e-mail feedback? [20:14] (yea, equivs seems like a hack, but it would actually be OK. i'd only have it in the Dockerfile, with a big fat comment.) [20:21] Yep. That's why I suggested it. [20:22] Easier if you close it. [20:39] alright, closed, thanks for looking into it. [20:56] No problem.