| vertago1 | If I were to start trying to help with packaging, where on KCI would be the best place to look? senile? | 00:12 |
|---|---|---|
| frecel | hello | 00:57 |
| frecel | is anyone working on plasma 5.5.1 packages for kubuntu or aware of a ppa that has them? | 00:57 |
| lordievader | Good morning. | 08:52 |
| yofel | frecel: no, we're not working on them yet. Yes, we are working in their direction (currently on frameworks), but we're very shorthanded currently | 10:06 |
| mparillo | yofel: Suggestion: Since it seems that either you or sgclark seem to have to reply almost daily, perhaps load it into the channel topic? | 10:39 |
| yofel | might as well add that back, we did have that in the past | 10:40 |
| === yofel changed the topic of #kubuntu-devel to: Kubuntu - Friendly computing | Plasma 5.5.1: TODO Apps 15.08.3: W/WIP X/TODO FW 5.17 W/TODO X/WIP | https://trello.com/kubuntu | http://qa.kubuntu.co.uk/ | milestoned bugs: http://goo.gl/AOx7bx | ||
| === yofel changed the topic of #kubuntu-devel to: Kubuntu - Friendly computing | Plasma 5.5.1: TODO Apps 15.08.3: W/WIP X/TODO FW 5.17: W/TODO X/WIP | https://trello.com/kubuntu | http://qa.kubuntu.co.uk/ | milestoned bugs: http://goo.gl/AOx7bx | ||
| yofel | a bit cryptic, but those people that actually bother to read the topic should understand taht | 10:42 |
| yofel | actually | 10:43 |
| === yofel changed the topic of #kubuntu-devel to: Kubuntu - Friendly computing | Plasma 5.5.1: TODO Apps 15.08.3: W/WIP X/TODO FW 5.17: W/TODO X/WIP | https://trello.com/kubuntu | http://qa.kubuntu.co.uk/ | milestoned bugs: http://goo.gl/AOx7bx | No, plasma 5.5 isn't packaged yet | ||
| yofel | peace | 10:43 |
| mparillo | TY. Mind if I try to add the last bit to the support channel? | 10:45 |
| mparillo | Sorry, I guess I do not have the permissions anyway. | 10:46 |
| yofel | I would rather not have that over there, even if it means answering a couple more questions. Additional info in the already long topic there is really just for emergencies | 11:00 |
| yofel | ok, so finally took a look at those app "backports" and the PPA had a dep on the INDI ppa which it shouldn't have. People really need to stop adding random PPA deps | 11:17 |
| yofel | uh... staging-frameworks didn't depend on -proposed | 11:21 |
| BluesKaj | Hiyas all | 11:40 |
| santa_ | yofel: hi since I have been absent for a while: where do you need help (besides parallelinzing git-clone-all)? | 18:07 |
| santa_ | * parallelizing | 18:08 |
| vertago1 | I was looking at the merging failures and it looks like it would just take someone with git access to address a merge conflict: http://kci.pangea.pub/job/merger_breeze-icons/ | 18:18 |
| vertago1 | I was looking into how to help, but I don't know the git information to look into fixing the merge conflicts. | 18:18 |
| yofel | vertago1: git access is restricted and requires membership on the qt-kde team on alioth (debian). In the beginning we recommend sending patches for actual packaging issues (or pointing us to a fixed git branch somewhere) or doing coop packaging with someone or tackling small package updates | 18:47 |
| yofel | let me think a bit about how a good way to start helping with packaging would be... | 18:48 |
| yofel | santa_: hard to say, didn't you want to implement your workflow? I'm currently trying to meet that coming from the other direction by reworking and splitting the current scripts a bit so we can get work done. | 18:49 |
| yofel | santa_: I pretty much gave up using your workflow after making git-clone-all work for me took too long. I work in temporary directories, so I added -t to the script, but then it started doing nothing at all because you are relying on a fixed path in getFtpVersionMap(), which will fail depending on when the function is called | 18:52 |
| yofel | after I saw that I didn't feel like digging more into it | 18:52 |
| santa_ | yofel: what was that -t suposed to do? | 18:53 |
| yofel | same thing as -t does for staging-upload | 18:53 |
| yofel | make a tmpdir and do all the work in there | 18:53 |
| yofel | so it can later be easily deleted | 18:53 |
| santa_ | well the idea of automation-ng is using clones which you are suposed to keep | 18:54 |
| santa_ | if your intention is to clone all the repositories each time a new upstream release arrives that is going to take a long time | 18:55 |
| yofel | doesn't work for me. I know myself, and if I have repo clones lying around, I *will* use them for random stuff. And I need a clean environment for the automation stuff | 18:55 |
| yofel | sure, that's why -t is supposed to be optional | 18:56 |
| yofel | there's nothing wrong with the script supporting both persistent and volatile workflows | 18:56 |
| santa_ | well, I fail to see whats the problem | 18:57 |
| santa_ | if you don't trust your previous clones you can clone them again | 18:57 |
| santa_ | or if you screwed up and you want to start fresh or whatever | 18:58 |
| yofel | well, for starters, the git-clone-all I tried, cloned all repos inside of the kubuntu-automation checkout, without subfolder. Sorry, but that's a no-go for me | 18:58 |
| yofel | ok, I can go and make a symlink, and copy all the config stuff it needs to the temporary location, but the script should do that really. | 18:59 |
| yofel | And why would a *library* function rely on a specific execution location? If it needs the path, that's supposed to be a function parameter, not an in-function call | 19:00 |
| santa_ | wat | 19:00 |
| yofel | def getFtpVersionMap(releaseType): | 19:01 |
| yofel | [snip] | 19:01 |
| yofel | cwd = os.path.dirname(os.path.realpath(__file__)) | 19:01 |
| vertago1 | yofel: any tips on sorting out the packing issues from the kci / git repo issues? | 19:01 |
| yofel | well, guess what, cwd depends on where you run it... | 19:01 |
| yofel | or where you are when that function is called | 19:01 |
| santa_ | yofel: that cwd depends on where lib/utils.py is located | 19:06 |
| yofel | not if you os.chdir() before you run the function | 19:06 |
| yofel | which is what -t does | 19:06 |
| yofel | I mean, you can rework it to work even with that cwd call, but why would you rely on that in the first place? | 19:07 |
| santa_ | hmm before anything, could you clarify exactly what are you trying to do and the expected behaviour? | 19:08 |
| santa_ | I mean | 19:09 |
| santa_ | $ git-clone-all -r frameworks -t | 19:09 |
| santa_ | ↑ this? | 19:09 |
| santa_ | something else? | 19:09 |
| yofel | ./git-clone-all -r frameworks -t ./foobar/ | 19:10 |
| yofel | after that all repos should be inside ./foobar/ | 19:10 |
| yofel | which means that git has to be run in foobar/, which means that cwd is in foobar | 19:10 |
| santa_ | oh, ok, I don't think that is difficult to achieve | 19:11 |
| yofel | probably not, but I hit so many env assumtions in the script that it felt like I'm wasting me time | 19:12 |
| santa_ | give me some minutes to code and think about what you have said so we can reach an "agreement" here | 19:12 |
| santa_ | I don't think there isn't so many env assumptions but let me have a look | 19:13 |
| yofel | I'm actually happy that you're putting code into lib functions, but I want to eventually have those testable, and the current implementation is rather bad for that | 19:13 |
| yofel | vertago1: not where it's just a merge conflict. It's better if you look for build failures, missing files, lintian warnings etc. For those you can easily send me a patch that I can apply | 19:16 |
| yofel | git diff would be idea ofc. The repositories are on http://anonscm.debian.org/cgit/pkg-kde/ and the CI builds from the kubuntu_unstable / kubuntu_stable branches | 19:17 |
| santa_ | yofel: ok. I have seen you have implemented the -t option, it works as expected here | 19:17 |
| yofel | I think it actually worked for frameworks or so, but did nothing for applications. Something weird like that | 19:18 |
| yofel | been a while since I tried it | 19:18 |
| vertago1 | has anyone thought about having KCI guess what the problem is to apply tags to build failures? | 19:18 |
| santa_ | yofel: also I'm inclined to think you are wrong about this: | 19:19 |
| santa_ | [20:06] <santa_> yofel: that cwd depends on where lib/utils.py is located | 19:19 |
| santa_ | [20:06] <yofel> not if you os.chdir() before you run the function | 19:19 |
| Riddell | vertago1: what's up with breeze-icons? I was packaging that today | 19:19 |
| yofel | possibly, I remember there being something wrong and thought that was it | 19:19 |
| santa_ | I think it works regardless of what you do with os.chdir | 19:19 |
| santa_ | yofel: ok, so would you give the thing another chance, so we can figure out the actual issues? | 19:21 |
| yofel | let me try | 19:21 |
| vertago1 | Riddell: I am fairly new to you all's packing workflow, but the merger script for breeze icons is complaining of a merge conflict | 19:21 |
| santa_ | we could try it for plasma 5.5.1 or so | 19:21 |
| Riddell | vertago1: url? | 19:21 |
| vertago1 | Riddell: http://kci.pangea.pub/job/merger_breeze-icons/ | 19:21 |
| vertago1 | is there any easy way to scrape the console outputs of the packages so I could look into writing a script to tag them? | 19:22 |
| yofel | you'll want to talk to sitter about that | 19:23 |
| vertago1 | actually I should wait on that. I am still trying to figure out which branch would be best to look for problems in. Senile? | 19:23 |
| Riddell | vertago1: thanks, fixed | 19:23 |
| vertago1 | Riddell: thanks | 19:24 |
| Riddell | vertago1: at some point a manual fix it needed, I'm not sure automating a tag on problems to label them would be worth any time it saved since manual look would still be needed | 19:24 |
| vertago1 | Riddel: yeah, I think it would be only useful if people were good at fixing particular kinds of problems and it was easier for them to find those problems by looking for a tag. | 19:25 |
| vertago1 | I am seeing this error a lot "Upload was likely rejected, we have been waiting for well over 30 minutes!" | 19:27 |
| vertago1 | It doesn't appear to be accurate because the timestamps don't reflect a 30minute wait | 19:28 |
| santa_ | yofel: is it possible that you got something like this: https://paste.kde.org/pnkresnlw ? | 19:29 |
| yofel | that too, I forgot the mkdir call | 19:29 |
| yofel | but the script does seem to work now | 19:30 |
| santa_ | yeah, was about to say that | 19:30 |
| clivejo | does anyone else find that when they click on the taskbar to switch to another prgramme, that the cursor picks up a shortcut instead? | 19:33 |
| vertago1 | yes | 19:33 |
| clivejo | its really getting on my nerves now | 19:33 |
| vertago1 | and sometimes it crashes plasma | 19:33 |
| vertago1 | or at least it used to | 19:33 |
| clivejo | I havent crashed plasma in months | 19:33 |
| vertago1 | yeah I think that part might be fixed | 19:33 |
| vertago1 | I don't understand why it is picking up an icon though | 19:33 |
| clivejo | but cant understand why is keeps picking up the shortcut | 19:34 |
| clivejo | and it keeps following me until I click on the desktop and create the icon | 19:34 |
| clivejo | I have to keep a section of desktop clear as an icon dump! | 19:34 |
| vertago1 | clivejo do you think an event might be getting lost so it thinks it is a click and drag instead of a single click? | 19:35 |
| clivejo | well I thought that it was my wireless mouse | 19:35 |
| clivejo | but I have a USB wired mouse and its doing the same | 19:35 |
| vertago1 | I have a logitech g9x and have the issue | 19:35 |
| clivejo | my wireless mouse is logitech | 19:35 |
| clivejo | but this wired one is a M$ one | 19:36 |
| clivejo | a Basic Optical Mouse version 1.0 | 19:36 |
| clivejo | 1.0A | 19:36 |
| yofel | I've seen that, but I can usually get rid by dropping the icon onto the app it came from | 19:37 |
| vertago1 | that is what I have done | 19:37 |
| shadeslayer | vertago1: re tags , you mean like http://kci.pangea.pub/job/wily_unstable_purpose/74/parsed_console/ | 19:37 |
| clivejo | yofel: its just a bloody pain having to dump the icon somewhere | 19:37 |
| yofel | agreed | 19:37 |
| clivejo | when you're busy trying to do something | 19:38 |
| clivejo | has anyone reported it? | 19:38 |
| vertago1 | shadeslayer: that is several steps toward what I would want to do. The next steps would involve filtering by errors or warnings. That would let you exclude all the build errors do to git | 19:39 |
| clivejo | I dont know if its my workflow, but the majority of the icons are firefox and dolphin | 19:39 |
| shadeslayer | Hm, might be possible | 19:39 |
| clivejo | firefox the vast majority | 19:39 |
| vertago1 | clivejo: I wonder if I could use wireshark's USB monitoring to troubleshoot it | 19:39 |
| shadeslayer | Iirc Jenkins allows Job tagging | 19:39 |
| clivejo | vertago1: if you know how, please do! Id love to see it resolved | 19:40 |
| vertago1 | clivejo I need to be able to trigger it though | 19:40 |
| clivejo | if 3 of us are experiencing it, must be an issue somewhere | 19:40 |
| vertago1 | clivejo: it seems to be sporadic | 19:40 |
| clivejo | <---- Xenial up-to-date as of yesterday morning | 19:41 |
| yofel | happens in wily for me | 19:41 |
| clivejo | the quicker, or more of a hurry Im in the worse it is | 19:41 |
| vertago1 | clivejo, it may be a click and drag threshold that needs to be adjusted | 19:42 |
| clivejo | wonder should I mention it in plasma | 19:43 |
| vertago1 | the drag start distance default seems to be 4pixels | 19:43 |
| clivejo | is it a kubuntu issue or all KDE distros | 19:43 |
| vertago1 | clivejo, I would try upping the drag start distance to 8 pixels or more and see if it makes it not as bad | 19:44 |
| clivejo | how do I do that? | 19:44 |
| vertago1 | if you search mouse on the application launcher it is under the advanced tab | 19:44 |
| vertago1 | called "drag start distance" | 19:44 |
| clivejo | ah yes | 19:45 |
| vertago1 | I use a high resolution: 1920x1200 so 4 pixels is fairly small | 19:45 |
| clivejo | upped to 8 | 19:45 |
| clivejo | will try that for a while | 19:45 |
| vertago1 | you may need it even higher | 19:45 |
| clivejo | what happened 5-6days ago to throw senial packages off? | 19:50 |
| santa_ | yofel, ok I found what I think it was wrong when you tried it | 19:50 |
| clivejo | back to over 300 broken | 19:50 |
| yofel | clivejo: branch switching for stable did that I guess | 19:51 |
| santa_ | yofel: maybe something like this https://paste.kde.org/pm7zbweh6 ? | 19:52 |
| yofel | don't think so, I think it literally did nothing (package list update from ftp failed?) | 19:52 |
| clivejo | yofel: any update on how to fix networkmanager-qt ? | 19:53 |
| yofel | update? that still is waiting for scarlett to update n-m | 19:53 |
| soee_ | Applications 15.12 released :) | 19:53 |
| === yofel changed the topic of #kubuntu-devel to: Kubuntu - Friendly computing | Plasma 5.5.1: TODO Apps 15.08.3: W/WIP X/TODO FW 5.17: W/TODO X/WIP | https://trello.com/kubuntu | http://qa.kubuntu.co.uk/ | No, plasma 5.5 isn't packaged yet | ||
| yofel | diff: - milestone link. That's better linked from qa | 19:54 |
| clivejo | soee_: when will you have the packaged and ready for testing :P | 19:54 |
| soee_ | :D | 19:55 |
| santa_ | yofel: in any case I think I have a couple of small changes for git-clone-all | 19:55 |
| soee_ | here i relay on the devs :_-) | 19:55 |
| === yofel changed the topic of #kubuntu-devel to: Kubuntu - Friendly computing | Plasma 5.5.1: TODO Apps 15.08.3: W/WIP 15.12.0: TODO FW 5.17: W/TODO X/WIP | https://trello.com/kubuntu | http://qa.kubuntu.co.uk/ | No, plasma 5.5 isn't packaged yet | ||
| yofel | 15.12 added | 19:56 |
| soee_ | topic says that: TODO Apps 15.08.3, should 15.12 be done not the old one ? | 19:56 |
| santa_ | btw this line | 19:56 |
| santa_ | workdir = args.workdir if args.workdir else cwd | 19:56 |
| yofel | now it does ;) | 19:56 |
| clivejo | soee_: you do know we are subscribed to the release mailing list and know this already :) | 19:56 |
| santa_ | disappeared misteriously after the merge | 19:57 |
| soee_ | clivejo: i;m like a smart bot that remainds it anyway :D | 19:57 |
| yofel | o.O | 19:57 |
| clivejo | like a facebook poke | 19:57 |
| yofel | now someone has to write a depression bot that constantly reminds us of all the things we have to do | 19:58 |
| vertago1 | you can put writing that bot on the list of things the bot should remind you to do | 19:59 |
| soee_ | :} | 19:59 |
| yofel | ahahaha | 19:59 |
| clivejo | oh kolab will be at FOSDEM | 19:59 |
| * yofel goes back to making ppa-build-status write out json | 20:02 | |
| * clivejo wishes he knew what was broken and how to fix it | 20:02 | |
| yofel | the script is currently both a) a data processor, b) a print-out-html-string-because-that's-the-cool-way thingy because that was easy to do and because nobody cares and because AAAAARGh | 20:03 |
| clivejo | like parley, why does the AMD64 build ok, but 386 wont install the deps?!? | 20:05 |
| yofel | where? | 20:05 |
| vertago1 | clivejo, that sounds like there may be a naming difference between amd64 and i386 | 20:05 |
| clivejo | vertago1: https://launchpadlibrarian.net/230041926/buildlog_ubuntu-xenial-i386.parley_4%3A15.08.2%2Bgit20151215.0457%2B16.04-0_BUILDING.txt.gz | 20:06 |
| yofel | probably more like other packages not building on i386 as well | 20:06 |
| vertago1 | hmm | 20:06 |
| clivejo | kio-dev but it is not going to be installed | 20:06 |
| * clivejo kicks KCI | 20:07 | |
| vertago1 | why can't it fetch teh gpg key? | 20:07 |
| yofel | yeah, you can't debug that without aptitude, apt with debug output, or dose output | 20:07 |
| yofel | good question, but this is launchpad internal, so it doesn't really need to | 20:07 |
| santa_ | yofel: pushed the simple fixes to https://github.com/jmsantamaria/kubuntu-automation-work | 20:08 |
| santa_ | this way it will work better | 20:08 |
| santa_ | btw I shouldn't have called that variable cwd, since it's confusing | 20:08 |
| vertago1 | santa_ scriptdir would probably make sense | 20:09 |
| santa_ | or installdir or something like that | 20:09 |
| yofel | merged. Into master btw. as we should really start using them (And I kind of messed up while juggling branches) | 20:11 |
| santa_ | ok, I will try to be here as much as possible to assist with issues | 20:12 |
| yofel | #packages we know use a different version than the rest of the release | 20:12 |
| yofel | differentVersion = ["baloo-kf5", "kfilemetadata-kf5", "kdepimlibs", "kdepim", "kdepim-runtime", "kde4libs"] | 20:12 |
| yofel | wow, even ppa-build-status has hardcoded configs /o\ | 20:13 |
| vertago1 | clifejo: so that is the kubuntu_xenail_archive branch here http://anonscm.debian.org/cgit/pkg-kde/applications/parley.git/refs/heads? | 20:24 |
| clivejo | the KCI is using the unstable and stable branches | 20:25 |
| vertago1 | ok | 20:25 |
| vertago1 | is that parley the stable branch? | 20:25 |
| clivejo | both stable and unstable are failing | 20:27 |
| clivejo | due to the 386 build | 20:27 |
| clivejo | I attempted to fix it and got the AMD64 to build locally and then on KCI | 20:27 |
| clivejo | but the 386 keeps FTBFS | 20:28 |
| clivejo | I have no idea why | 20:28 |
| vertago1 | clivejo are all the dependencies available? | 20:28 |
| vertago1 | I am looking at the list | 20:28 |
| vertago1 | cmake, | 20:28 |
| vertago1 | debhelper (>= 9), | 20:28 |
| vertago1 | extra-cmake-modules, | 20:28 |
| vertago1 | imagemagick, | 20:28 |
| vertago1 | kdoctools-dev, | 20:28 |
| vertago1 | kio-dev, | 20:28 |
| clivejo | as far as I can see | 20:28 |
| vertago1 | kross-dev, | 20:28 |
| vertago1 | libkeduvocdocument-dev, | 20:28 |
| vertago1 | libkf5config-dev, | 20:28 |
| vertago1 | libkf5coreaddons-dev, | 20:28 |
| vertago1 | libkf5crash-dev, | 20:28 |
| vertago1 | libkf5i18n-dev, | 20:29 |
| vertago1 | libkf5kcmutils-dev, | 20:29 |
| vertago1 | libkf5khtml-dev, | 20:29 |
| vertago1 | libkf5newstuff-dev, | 20:29 |
| vertago1 | libkf5sonnet-dev, | 20:29 |
| vertago1 | libqt5svg5-dev, | 20:29 |
| vertago1 | libxml2-dev, | 20:29 |
| vertago1 | libxslt-dev, | 20:29 |
| vertago1 | pkg-config, | 20:29 |
| vertago1 | pkg-kde-tools (>= 0.12), | 20:29 |
| vertago1 | qtmultimedia5-dev, | 20:29 |
| vertago1 | qtscript5-dev | 20:29 |
| clivejo | vertago1: we prefer to use a paste bin :) | 20:29 |
| vertago1 | yeah I didn't realize it was going to do that | 20:29 |
| clivejo | https://paste.kde.org/ | 20:29 |
| clivejo | or the likes | 20:29 |
| vertago1 | quassel keeps sending me to file : ///home..../kioexec/krun/ when I click URLs | 20:30 |
| clivejo | I did attempt to setup a 386 pbuild but I ran into problems | 20:30 |
| vertago1 | how do you sent one up? | 20:30 |
| vertago1 | I can setup a i386 container | 20:31 |
| clivejo | https://wiki.ubuntu.com/PbuilderHowto | 20:31 |
| * clivejo doesnt like 386 | 20:31 | |
| vertago1 | have you use lxc at all? | 20:32 |
| vertago1 | they arn't that bad. I set one up because of a regression in samba | 20:32 |
| yofel | clivejo: what problem did you run into regarding i386 pbuilder? | 20:32 |
| * clivejo has a slow internet connection and had to upload and downland calligra source on numurious occassions | 20:33 | |
| clivejo | yofel: getting the packages | 20:33 |
| yofel | o.O that should work the same as always when in an i386 chroot | 20:33 |
| santa_ | The following packages couldn't be cloned: | 20:35 |
| santa_ | spectacle | 20:35 |
| santa_ | ↑ what's that? | 20:35 |
| vertago1 | pbuilder create --architecture i386? | 20:37 |
| yofel | seems to be some new app in 15.12 | 20:37 |
| vertago1 | looks like it wants sudo | 20:37 |
| vertago1 | I am going to see if I can run it in userland | 20:37 |
| * yofel has a very custom pbuilder setup (not userland) | 20:38 | |
| yofel | so can't really help with the defaults | 20:38 |
| * clivejo wonders how yofel has pbuilder set up | 20:38 | |
| vertago1 | it doesn't work in userland because of debbootstrap | 20:38 |
| vertago1 | I got it to start fetching stuff | 20:39 |
| clivejo | mine fetched lots of stuff, but then stopped | 20:39 |
| clivejo | that ppa seems to corrupt things | 20:39 |
| yofel | enjoy my pbuilderrc paste.ubuntu.com/14056289/ - lots of stuff you don't need, but maybe you'll find something interesting | 20:39 |
| vertago1 | the problem is I am on wiley so I don't know if it will be a good test | 20:39 |
| yofel | hm, I need to update the debian codenames ^^ | 20:40 |
| clivejo | vertago1: you can create pbuilder instances for xenial too | 20:40 |
| clivejo | BTW senial is sitters codename for xenial | 20:41 |
| clivejo | sorry senile | 20:41 |
| vertago1 | is that how it is pronounced anyway? | 20:42 |
| clivejo | he has me typing senial by mistake | 20:42 |
| BluesKaj | zenial | 20:42 |
| clivejo | only if you have a certain accent ;) | 20:42 |
| vertago1 | hahaha | 20:42 |
| vertago1 | BlueKaj that is what I thought | 20:42 |
| clivejo | grrrr stupid symbols | 20:44 |
| clivejo | what is purpose anyway? | 20:46 |
| clivejo | Ive packaged it and have no idea what its for! | 20:46 |
| yofel | Description: library for abstractions to get the developer's purposes fulfilled | 20:47 |
| yofel | uhuuuuh... | 20:47 |
| yofel | sounds very meta | 20:47 |
| * clivejo feels enlightened | 20:47 | |
| clivejo | yofel: its part of framesworks now? | 20:48 |
| yofel | looks like it | 20:48 |
| yofel | maybe it has a purpose there | 20:48 |
| clivejo | why doesnt it have a frameworks version number? | 20:48 |
| vertago1 | hmm | 20:49 |
| vertago1 | it looks like my system config is incompatible with pbuilder as is | 20:49 |
| yofel | oh, it's not in 5.17 | 20:50 |
| yofel | #libnamingconfusion | 20:50 |
| vertago1 | I am going to have to get it to download the packages to the same physical device I am making the base configs on | 20:50 |
| yofel | if you're getting ln errors, you can disable the aptcache and use bindmounts or so | 20:51 |
| vertago1 | it is trying to hardlink across two hard drives | 20:51 |
| clivejo | strange, its under frameworks in debian git - http://anonscm.debian.org/cgit/pkg-kde/frameworks/purpose.git/ | 20:52 |
| yofel | right, I get that too when building in tmpfs, so I'm using this when I do: | 20:52 |
| yofel | BINDMOUNTS="/var/cache/apt/archives/ /run/" | 20:52 |
| yofel | APTCACHE="" | 20:52 |
| vertago1 | I have a ssd for / and a platter drive for /home | 20:52 |
| vertago1 | I am just going to let it use /var for the base | 20:53 |
| vertago1 | I may need to move it later though | 20:53 |
| santa_ | https://github.com/jmsantamaria/kubuntu-automation-work/commits/master | 20:59 |
| santa_ | ↑ yofel stuff needed to bump the build depends for the last versions | 20:59 |
| yofel | I really don't see the point of that list you know | 21:02 |
| yofel | but thanks for the update | 21:02 |
| santa_ | the json with all the -dev packages? | 21:02 |
| yofel | yep | 21:03 |
| yofel | or well, the package list has a potential use, but the versions? | 21:03 |
| santa_ | they have the epoch or not | 21:03 |
| santa_ | so no need to hardcode which packages are epoch'ed | 21:04 |
| vertago1 | clivejo so after getting a i386 pbuilder setup where do I get the .dsc to do the build? | 21:04 |
| santa_ | note that the list is generated automatically with dev-package-names-list | 21:05 |
| santa_ | also note that this script could be glued with the bump-* script in the future | 21:05 |
| yofel | aaaaah, that script actually does look up the real epoch | 21:05 |
| yofel | ok | 21:05 |
| santa_ | yeah, parses the changelogs | 21:06 |
| clivejo | vertago1: debuild -S | 21:08 |
| vertago1 | clivejo, in what context? | 21:08 |
| clivejo | to get the .dsc to build in pbuilder | 21:09 |
| vertago1 | I need the tar files | 21:10 |
| vertago1 | should I enter the pbuild environment and then use apt-source? | 21:12 |
| vertago1 | I mean apt-get source | 21:12 |
| clivejo | well I grab the debian git branch and then use uscan to grab the source | 21:14 |
| vertago1 | it is complaining that I don't have a sources.list | 21:14 |
| yofel | that's an option. Other one is to look up the URL on the web ui, or to use pull-ppa-source from lp:kubuntu-dev-tools, or to just rebuild from git as clive said | 21:14 |
| clivejo | but KCI is grabbing the source directly from git | 21:15 |
| vertago1 | I already have the source from git and I checkout out the xenail-unstable branch | 21:15 |
| clivejo | so in those cases I download the source from KCI PPA's | 21:15 |
| clivejo | then you need to debuild -S to build the source | 21:16 |
| clivejo | and instead of uploading to your PPA, you build it in pbuilder | 21:16 |
| clivejo | for example use pbuilder-dist xenial build ../*dsc | 21:17 |
| clivejo | that builds it in a nice clean xenial install | 21:17 |
| clivejo | on my machine | 21:17 |
| clivejo | yofel: do you know what top level package creates libkf5kcmutils-dev ? | 21:19 |
| vertago1 | it is complaing about no .orig.tar.xz even though it is there | 21:19 |
| yofel | clivejo: apt-cache showsrc libkf5kcmutils-dev | 21:19 |
| vertago1 | it looks like the version number is wrong | 21:19 |
| santa_ | kscreenlocker (exit status = 1) | 21:20 |
| santa_ | discover (exit status = 1) | 21:20 |
| santa_ | breeze-gtk (exit status = 1) | 21:20 |
| clivejo | check debian/changelog and make sure the versions are the same | 21:21 |
| santa_ | ↑ this packages from plasma doesn't have a kubuntu_xenial_archive branch | 21:21 |
| vertago1 | how do you keep uscan from downloading newer versions? | 21:21 |
| yofel | they're new | 21:21 |
| santa_ | could you create the branches then so I can continue to test the thing? | 21:22 |
| santa_ | (no hurry) | 21:22 |
| vertago1 | I must have the wrong download url for uscan it says no matching hrefs for 15.08.2 | 21:25 |
| yofel | no, that's uscan not properly fetching all versions it could find. | 21:26 |
| yofel | I believe there's an uscan bug report for that | 21:26 |
| vertago1 | looks like I will have to do it manually | 21:27 |
| vertago1 | ls | 21:30 |
| vertago1 | hmm my build failed but for a different reason | 21:33 |
| vertago1 | actually when I looked at it more closely it is the same error, but my base probably has less installed by default | 21:34 |
| vertago1 | http://paste.ubuntu.com/14056655/ | 21:36 |
| clivejo | yofel: https://launchpadlibrarian.net/230149652/buildlog_ubuntu-xenial-i386.kdeclarative_5.17.0%2Bgit20151216.0354%2B16.04-0_BUILDING.txt.gz says libkf5declarative-dev_5.17.0+git20151216.0354+16.04-0_i386.deb is being built ok | 21:36 |
| yofel | okaaaay..? | 21:36 |
| clivejo | yet sbuild-build-depends-kcmutils-dummy : Depends: libkf5declarative-dev (>= 5.16.0~) but it is not going to be installed | 21:37 |
| clivejo | why? | 21:37 |
| yofel | apt saying that doesn't mean that package is the problem | 21:37 |
| yofel | it is the package that apt fails on, due to dep issues with other packages | 21:37 |
| yofel | as I said, you need the output from aptitude, or apt with -o debug::pkgproblemresolver=true, or some kind of dose output to debug this | 21:38 |
| yofel | the apt output in the buildlog is next to useless | 21:38 |
| clivejo | it only seems to be an issue in 386 | 21:39 |
| vertago1 | yofel is there a way to get apt to include that option inside pbuilder? | 21:39 |
| yofel | you can put it into the chroot's apt.conf I gues | 21:39 |
| yofel | s | 21:39 |
| vertago1 | how do you change the base and save updates? | 21:40 |
| vertago1 | nvm | 21:40 |
| vertago1 | I will just manually edit the .tgz | 21:40 |
| yofel | pbuilder login --save-after-login | 21:40 |
| vertago1 | that is easier | 21:40 |
| yofel | just make sure your last command exist with 0 | 21:41 |
| yofel | otherwise it won't save | 21:41 |
| yofel | python 2's datetime class has no easy way to generate a unix timestamp, seriously? | 21:42 |
| yofel | first time that I see something useful in 3 | 21:42 |
| vertago1 | yofel are you talking about an epoch timestamp? I am pretty sure you can get that in python 2 | 21:44 |
| yofel | I'm properly bllind, but I see nothing on https://docs.python.org/2/library/datetime.html | 21:45 |
| yofel | I'll just port the script to py3. Needs to be done anyway | 21:45 |
| vertago1 | yofel there is no obvious function, but it looks like most people just use .strftime("%s") | 21:47 |
| vertago1 | which is ugly because you have to parse a format string | 21:47 |
| clivejo | yofel: I used sudo DIST=xenial ARCH=i386 pbuilder create does this overwrite my AMD64 environment? | 21:47 |
| vertago1 | clivejo I used pbuilder create --basetgz xenial_i386_base.tgz --architecture i386 --distribution xenial | 21:48 |
| yofel | not with my config if that's what you're trying... | 21:48 |
| vertago1 | my config is empty | 21:48 |
| vertago1 | minus the parallel build lines I copied from yofel's config | 21:49 |
| clivejo | Im reviewing my 386 setup, trying to find out why its not working | 21:49 |
| clivejo | but I only have one xenial-base.tgz | 21:51 |
| clivejo | I think Ive overwritten my amd64 environment | 21:51 |
| vertago1 | clivejo I explicity name my base.tgz to keep from doing that | 21:51 |
| yofel | my setup will e.g. create a /var/cache/pbuilder/xenial-amd64.tgz | 21:52 |
| clivejo | vertago1: thanks, I must do that in future | 21:52 |
| clivejo | mine seems to put it in ~/pbuilder | 21:52 |
| yofel | right, depends on your config | 21:52 |
| clivejo | which is what I want my /home is on a HDD and keeps writes off my SSD | 21:53 |
| vertago1 | this is the new error I get after adding pkgproblemresolver to the config: http://paste.ubuntu.com/14056915/ | 21:54 |
| vertago1 | it looks like my source urls are wrong | 21:54 |
| vertago1 | kdoctools-dev which is a virtual package and is not provided by any available package. | 21:54 |
| clivejo | did you add the KCI PPA? | 21:55 |
| vertago1 | no | 21:55 |
| vertago1 | I will do that now | 21:56 |
| * clivejo waits on new pbuilder base to download | 21:57 | |
| clivejo | sucks being on a slow connection | 21:57 |
| clivejo | vertago1: what are you trying to build? | 21:59 |
| vertago1 | parley on i386 | 21:59 |
| vertago1 | since it was broken | 21:59 |
| vertago1 | I am having trouble finding the exact ppa I should add | 22:00 |
| vertago1 | this one: https://launchpad.net/~kubuntu-ci/+archive/ubuntu/unstable? | 22:01 |
| clivejo | looks like it | 22:01 |
| clivejo | I think its higher up in the dep tree | 22:03 |
| clivejo | a lot of those failing on KCI are due to the 386 build | 22:04 |
| vertago1 | I just got the repo added | 22:04 |
| clivejo | vertago1: are you on LP? | 22:05 |
| yofel | wait, you're keeping writes off the SSD? why? They're built for stuff like that | 22:05 |
| vertago1 | yofel my home dir is much much larger than my ssd | 22:05 |
| clivejo | yofel: I read keeping writes off a SSD helps its life | 22:06 |
| yofel | the split itself I understand, I had that myself for a while | 22:06 |
| vertago1 | I just have to make sure I don't fill up the ssd, so far it isn't a problem but I have had to put /var/lib/mysql off the ssd before | 22:07 |
| yofel | clivejo: you should measure the wear and see if it matters. From my experience it doesn't unless you have really heavy workloads. And for package building the SSD is a nice thing | 22:07 |
| vertago1 | I have enough ram I could do the package building in ram | 22:07 |
| vertago1 | unless it is a big package like gcc | 22:07 |
| vertago1 | or plasma | 22:08 |
| clivejo | yofel: its small so I just use it for system, all tmp and cache are on HDD | 22:08 |
| yofel | ah, then it might matter | 22:08 |
| yofel | the smart data usually has some vendor dependent wear indicator | 22:08 |
| clivejo | then /home is backed up to 2Tb external HDD | 22:08 |
| clivejo | I dont want to kill my SSD, I love it too much! | 22:09 |
| yofel | that's how I felt about my first 60G SSD, in the end I stopped worrying, but still only managed to wear it out ~50% after ~5 years | 22:10 |
| vertago1 | I have 1.02TB of writes to mine | 22:10 |
| clivejo | oh, this time pbuilder create has made xenial-i386-base.tgz | 22:11 |
| clivejo | in /var/cache/pbuilder/xenial-i386-base.tgz | 22:12 |
| clivejo | strange | 22:12 |
| clivejo | didnt do that last time | 22:12 |
| yofel | my server SSD (300GB intel 320) says it wrote 86TiB, and the wear indicator sits at 94% life left (The warranty was 35TiB). I guess they measure the absolute worst case workload | 22:13 |
| vertago1 | here is the most recent failure | 22:14 |
| vertago1 | http://paste.ubuntu.com/14057389/ | 22:14 |
| yofel | is universe enabled in the PPA? | 22:15 |
| clivejo | I cant get i386 to work | 22:21 |
| vertago1 | i am enabling universe now and will retry it | 22:27 |
| vertago1 | it appears to be working | 22:29 |
| * clivejo gives up again, too tired | 22:29 | |
| vertago1 | yofel it looks like for some reason the repos are not right for the i386 builds | 22:30 |
| vertago1 | the error matches the error I got before I added the KCI ppa | 22:30 |
| clivejo | vertago1: the dep's are failing too | 22:31 |
| yofel | could be, still leaves the question of what the issue is | 22:31 |
| vertago1 | mine is building though | 22:31 |
| vertago1 | no dependency problems after enabling the ppa and universe | 22:32 |
| vertago1 | it succeeded but I don't know where it put the .deb file | 22:34 |
| vertago1 | where is the kci build code for setting up pbuilder for i386? | 22:34 |
| vertago1 | I have enough information I could probably write a wiki page tutorial on how to setup a local test | 22:37 |
| yofel | kci doesn't use pbuilder but uploads the builds to launchpad (which uses some customized sbuild AFAIR) | 22:42 |
| vertago1 | yofel any idea on how to configure its source repos? | 22:43 |
| yofel | grrr, whoever said that string handling in py3 improved was crazy | 22:43 |
| yofel | vertago1: pbuilder has an OTHERMIRROR setting that you might be able to use to inject a ppa into --create | 22:44 |
| vertago1 | I meant for launchpad so kci can be fixed | 22:44 |
| clivejo | vertago1: do you have buildlog for your parley build in pbuilder? | 22:45 |
| clivejo | see what version of the packages its installing | 22:45 |
| vertago1 | it looks like i need to run it again and pipe the output to a file | 22:46 |
| vertago1 | or use --logfile | 22:47 |
| clivejo | could be installing older stable versions | 22:47 |
| vertago1 | http://paste.ubuntu.com/14058223/ | 22:52 |
| clivejo | its installing older versions | 22:52 |
| clivejo | libkf5kcmutils-data [5.15.0-0ubuntu2 (xenial)] | 22:53 |
| clivejo | libkf5kcmutils5 : Depends: libkf5kcmutils-data (= 5.17.0+git20151214.0845+16.04-0) but it is not going to be installed. | 22:53 |
| clivejo | installing from the archive rather than the PPA | 22:54 |
| clivejo | can you add a hook to drop into the terminal | 22:55 |
| clivejo | see why it cant/wont install 5.17.0 | 22:55 |
| vertago1 | how do I do that? | 22:56 |
| clivejo | https://wiki.ubuntu.com/PbuilderHowto#Using_the_Kubuntu_pbuilder_hooks | 22:57 |
| clivejo | theres a hook to drop into the pbuilder environment after the build | 22:58 |
| clivejo | cant remember now to do it exactly | 22:59 |
| clivejo | how | 22:59 |
| vertago1 | I see list missing | 23:00 |
| yofel | C10shell is what you need | 23:01 |
| clivejo | yofel to the rescue! | 23:02 |
| yofel | helping you guys is more fun than fixing python str != byte issues | 23:02 |
| clivejo | vertago1: http://bazaar.launchpad.net/~kubuntu-packagers/pbuilder/pbuilder-hooks/view/head:/C10shell | 23:02 |
| clivejo | yofel: does it only drop to shell if it fails? | 23:03 |
| yofel | C10shell yes, B10shell would drop on success | 23:03 |
| clivejo | cause it is building successfully | 23:03 |
| yofel | :96 | 23:04 |
| yofel | oops | 23:04 |
| clivejo | its dropping back to 5.15.0-0ubuntu2 instead of using 5.17.0 | 23:04 |
| yofel | can you install the 5.17 version by hand? | 23:06 |
| clivejo | I cant even get a i386 environment setup :( | 23:07 |
| vertago1 | ok I got a shell after the build | 23:07 |
| vertago1 | The following packages have unmet dependencies: | 23:08 |
| vertago1 | libkf5khtml5 : Depends: libkf5khtml-data (= 5.17.0+git20151214.0838+16.04-0) but 5.17.0+git20151215.0437+16.04-0 is to be installed | 23:08 |
| vertago1 | This one probably needs to be fixed first: http://kci.pangea.pub/job/xenial_unstable_khtml/ | 23:12 |
| vertago1 | https://launchpadlibrarian.net/230040872/buildlog_ubuntu-xenial-i386.khtml_5.17.0%2Bgit20151215.0437%2B16.04-0_BUILDING.txt.gz | 23:13 |
| clivejo | thats failing on kio-dev, libkf5parts-dev and libkf5wallet-dev | 23:14 |
| vertago1 | those are the same as before | 23:14 |
| clivejo | thats why I think its up the dep tree somewhere | 23:14 |
| clivejo | ah - https://launchpadlibrarian.net/230149363/buildlog_ubuntu-xenial-i386.kio_5.17.0%2Bgit20151216.0344%2B16.04-0_BUILDING.txt.gz | 23:15 |
| clivejo | symbols | 23:15 |
| vertago1 | it looks like a function name was changed | 23:17 |
| vertago1 | - _ZNK13KUrlNavigator7Private17retrievePlacePathEv@Base 4.96.0 + _ZNK13KUrlNavigator7Private16retrievePlaceUrlEv@Base 5.17.0+git20151216.0344+16.04-0 | 23:18 |
| clivejo | yofel: ping | 23:27 |
| yofel | hm? | 23:27 |
| clivejo | Ive updated the symbols for kio | 23:27 |
| yofel | good? | 23:28 |
| clivejo | but not sure on what to put in the changelog | 23:28 |
| yofel | for CI? nothing | 23:28 |
| yofel | or what branch are you on? | 23:29 |
| clivejo | kubuntu_unstable | 23:29 |
| yofel | nothing then | 23:29 |
| yofel | just write a proper commit message | 23:29 |
| vertago1 | hopefully that is at least part of a fix if not a full fix | 23:30 |
| clivejo | soon see :) | 23:31 |
| clivejo | KCI has picked it up and building now | 23:31 |
| clivejo | vertago1: how long did it take to build on your machine? | 23:33 |
| vertago1 | not very long but I was using 8 cores | 23:33 |
| clivejo | last build on KCI was 24mins | 23:34 |
| clivejo | so I guess LP is slower | 23:34 |
| yofel | that includes the publishing | 23:35 |
| clivejo | i386 is 25% | 23:36 |
| clivejo | https://launchpad.net/~kubuntu-ci/+archive/ubuntu/unstable/+build/8449523 | 23:36 |
| clivejo | yofel: BTW can I start KCI rebuilds? | 23:38 |
| clivejo | I logged in with my LP ID but says I dont have permission | 23:39 |
| yofel | that requires being in ~kubuntu-members | 23:39 |
| clivejo | oh | 23:39 |
| yofel | wtf? in my script here, log.index() always throws a "ValueError: substring not found", but does return the correctly found index | 23:41 |
| stacksUpHigh | is gnome good? new to linux.. | 23:41 |
| stacksUpHigh | python programming is so much better using ubuntu though | 23:41 |
| yofel | #ubuntu-gnome will likely answer with yes, we will with no | 23:41 |
| yofel | so I don't think we can help much there ^^ | 23:41 |
| clivejo | stacksUpHigh: We use the KDE desktop, hence the K in Kubuntu | 23:42 |
| vertago1 | I haven't used gnome much since gnome 2 | 23:42 |
| * clivejo is not a fan of Gnome or Unity | 23:43 | |
| clivejo | well the buildlogs look good :) | 23:44 |
| clivejo | yofel: KCI retries all builds at midnight UTC? | 23:44 |
| yofel | during the night, no idea what exact time | 23:45 |
| yofel | sounds about right though looking at the logs | 23:45 |
| clivejo | I seem to get pinged every night around that time! | 23:45 |
| vertago1 | that is 15minutes from now | 23:46 |
| vertago1 | hopefully it finishes before that | 23:46 |
| clivejo | its finished | 23:47 |
| clivejo | just needs to publish the files to the PPA | 23:47 |
| clivejo | always forget about that part | 23:48 |
| clivejo | takes ages! | 23:48 |
| clivejo | the bot should post a message in here | 23:49 |
| yofel | *facepalm* | 23:51 |
| yofel | I spent an hour debugging a corrupt cache. RAGE | 23:51 |
| vertago1 | ouch | 23:51 |
| vertago1 | ccache? | 23:51 |
| * clivejo offers Phillip a stiff drink | 23:52 | |
| yofel | no, the script has it's own caching | 23:52 |
| yofel | heh thanks | 23:52 |
| yofel | and kio is happy | 23:52 |
| clivejo | yippee we saved a Lemming | 23:52 |
| vertago1 | hopefully all its dependents become happy once kci gives them the kick to rebuild | 23:52 |
| clivejo | wheres wily | 23:53 |
| * clivejo kicks kubuntu-ci bot | 23:53 | |
| vertago1 | does the merge work? | 23:53 |
| yofel | brrr, there's even more html inlining *-.- | 23:53 |
| clivejo | hopefully purpose can build tonight and I can fix the symbols tomorrow | 23:54 |
| yofel | I triggered a build for it | 23:56 |
| clivejo | there are other deps need to build first | 23:56 |
| clivejo | kaccounts, kcmutils and kdeconnect | 23:58 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!