| === work_alkisg is now known as alkisg | ||
| pitti | Good morning | 06:38 |
|---|---|---|
| alkisg | Good morning | 06:40 |
| Noskcaj | Should we sync dbus-glib from exp? It includes a few fixes and officially deprecates the package | 07:09 |
| alkisg | mvo, good morning, I did "verification-done" for 3 out of the 4 bugs for the update-manager precise SRU, but I'm having trouble with verifying LP #1341320 | 07:16 |
| ubottu | Launchpad bug 1341320 in update-manager (Ubuntu Precise) "empty apt-get install command suggested by hwe-support-status" [Medium,Fix committed] https://launchpad.net/bugs/1341320 | 07:16 |
| alkisg | (I did #1402706 #1415785 #1420217) | 07:16 |
| alkisg | hwe-support-status doesn't give me an apt line at all | 07:17 |
| mvo | alkisg: woah, you ROCK | 07:19 |
| alkisg | :) | 07:19 |
| mvo | alkisg: that happens if you have update-manager installed, then it won't tell you about apt-get :) | 07:19 |
| alkisg | Aaah | 07:19 |
| mvo | alkisg: is that maybe what you see? | 07:19 |
| alkisg | Yes I was trying with the desktop cd | 07:19 |
| alkisg | So I only need update-manager-core installed? | 07:19 |
| mvo | alkisg: yeah, if it detect the GUI being instlaled it will not bother with the low-level apt-get as update-manager in the UI can also do the transition | 07:20 |
| alkisg | Gotcha, trying... | 07:20 |
| mvo | alkisg: thanks so much for your help with! | 07:20 |
| alkisg | True, now it gives me an apt line. I'll do the verification-done step. | 07:21 |
| mvo | \o/ | 07:22 |
| alkisg | mvo: for 1341320, am I supposed to get an empty apt line with the previous update-manager-core? | 07:28 |
| alkisg | See https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/1420217 comment #3 for what I'm getting | 07:28 |
| ubottu | Launchpad bug 1420217 in update-manager (Ubuntu Precise) "HWE meta packages "linux-signed-generic"/"linux-signed-image-generic" not considered" [Medium,Fix committed] | 07:28 |
| mvo | alkisg: let me check | 07:28 |
| alkisg | I.e. I'm getting an apt line, but it just doesn't contain the signed variant | 07:28 |
| alkisg | And with .18, it does contain the signed variant | 07:28 |
| mvo | alkisg: so if there is no meta package anymore but a leftover package, then it will report a empty line, hm, maybe the test case is not good | 07:29 |
| mvo | alkisg: or just upgrade the meta packages first (i.e. run the commandline that is given there) | 07:29 |
| mvo | alkisg: and see if after that you get the empty output :) | 07:29 |
| alkisg | Gotcha | 07:29 |
| alkisg | mvo: that one was a little tricky, I had to update everything and remove all the previous meta packages and only keep a couple of old kernels around | 07:43 |
| alkisg | Could you check this output? http://sync.in/plinetio | 07:43 |
| mvo | alkisg: looks excellent | 07:43 |
| alkisg | OK, putting it to the bug report... | 07:43 |
| mvo | alkisg: yes, its a bit tricky my description should have been better, its a bit of a corner case | 07:44 |
| dholbach | good morning | 07:44 |
| alkisg | mvo: all done, thanks for the SRU! | 07:48 |
| mvo | alkisg: very nice, thanks for the verification | 07:50 |
| alkisg | tjaalton: hi, could I ping you to release an update-manager SRU? Thank you very much! http://people.canonical.com/~ubuntu-archive/pending-sru.html and https://wiki.ubuntu.com/StableReleaseUpdates#Publishing | 07:51 |
| tjaalton | alkisg: it's friday, so no packages released to -updates today :/ | 07:53 |
| alkisg | np, thank you | 07:54 |
| tjaalton | you'll get it on monday ;) | 07:54 |
| alkisg | :) | 07:54 |
| === alkisg is now known as work_alkisg | ||
| === darkxst_ is now known as darkxst | ||
| flexiondotorg_ | dholbach, Any chance you could help with a little sponsoring today? | 09:21 |
| dholbach | hi flexiondotorg_ | 09:23 |
| flexiondotorg_ | dholbach, Morning | 09:23 |
| dholbach | if you're talking about https://code.launchpad.net/~ubuntu-mate-dev/compiz/compiz-mate/+merge/249578 I'm not sure I'm a good person to make a decision | 09:24 |
| dholbach | bregma, Trevinho: ^ | 09:24 |
| dholbach | can you guys maybe help out with this? | 09:24 |
| dholbach | or maybe other folks in #ubuntu-unity or #ubuntu-desktop? | 09:24 |
| flexiondotorg_ | dholbach, Nope, not that 😃 | 09:25 |
| flexiondotorg_ | I have some merge proposal, that have been approved and merged. But not packages have been built since. | 09:25 |
| dholbach | aha? | 09:25 |
| flexiondotorg_ | https://code.launchpad.net/~ubuntu-mate-dev/ubiquity-slideshow-ubuntu/ubuntu-mate/+merge/244191 | 09:25 |
| flexiondotorg_ | https://code.launchpad.net/~profzoom/lightdm-gtk-greeter/add-mate-badge/+merge/243069 | 09:26 |
| flexiondotorg_ | https://code.launchpad.net/~ubuntu-mate-dev/indicators-gtk2/indicator-sound-gtk2/+merge/244120 | 09:26 |
| flexiondotorg_ | https://code.launchpad.net/~ubuntu-mate-dev/indicators-gtk2/indicator-application-gtk2/+merge/244116 | 09:27 |
| flexiondotorg_ | dholbach, There is also quite a serious bug in gtk2 affecting MATE but we have supplied a patch. | 09:28 |
| flexiondotorg_ | https://bugs.launchpad.net/ubuntu/+source/mate-control-center/+bug/1351890 | 09:28 |
| ubottu | Launchpad bug 1351890 in mate-control-center (Ubuntu) "Changing external screen resolution with dual monitors" [Medium,Confirmed] | 09:28 |
| flexiondotorg_ | Patch in #9 | 09:28 |
| caribou | When a package is synced from debian/experimental, does it get automatically synced again from experimental ? | 09:28 |
| flexiondotorg_ | dholbach, How do we progress that? | 09:28 |
| caribou | I mean, do I need to re-sync everytime ? | 09:29 |
| flexiondotorg_ | caribou, No. | 09:29 |
| caribou | flexiondotorg_: ok, thanks | 09:29 |
| flexiondotorg_ | caribou, At least not while the sync is still enabled. | 09:29 |
| dholbach | flexiondotorg_, for the last bug, add a patch including a d/changelog entry and subscribe ubuntu-sponsors | 09:29 |
| flexiondotorg_ | caribou, Debian sync end February 19th for Vivid. | 09:29 |
| dholbach | I'm fairly busy today, but I'll take a look at the bugs above - I would have expected somebody from the desktop team to land them :-( | 09:30 |
| flexiondotorg_ | dholbach, If you have contacts in the desktop team please introduce me. | 09:30 |
| flexiondotorg_ | So far I have been depending on you and cyphermox. | 09:31 |
| dholbach | seb128, didrocks, Laney, happyaron for example | 09:31 |
| flexiondotorg_ | seb128, didrocks, Laney, happyaron - I am the lead for Ubuntu MATE. Would any of you be able to help with some sponsoring today? | 09:32 |
| seb128 | dholbach, easier to say #ubuntu-desktop | 09:32 |
| didrocks | flexiondotorg_: note sure about today, I have some patch pilot round on Tuesday, I will take a look then if nobody beats me to it | 09:32 |
| didrocks | maybe I'll be able to do one or two beforehand, but with FF, quite busy… | 09:33 |
| flexiondotorg_ | didrocks, Understand. This is my concern. FF is upon us and I have quite a bit of outstanding stuff to be actioned 😞 | 09:33 |
| didrocks | dholbach: flexiondotorg_: I'm happy to get some today if I get some time (switching with Tuesday's patch pilot), but as usual, then I won't be marked has having done my shift when I did it, but anyway :p | 09:35 |
| flexiondotorg_ | didrocks, Don't mess with your schedule on my account. I understand you guys are super busy. | 09:36 |
| didrocks | flexiondotorg_: however, you should still see with bregma/Trevinho for the compiz one, I won't feel confortable on this one | 09:37 |
| flexiondotorg_ | didrocks, Yes. I will follow up with them regarding Compiz. | 09:37 |
| flexiondotorg_ | bregma, Trevinho - Please can you take another peek at https://code.launchpad.net/~ubuntu-mate-dev/compiz/compiz-mate/+merge/249578 | 09:38 |
| dholbach | looking at indicator-sound-gtk2 now | 09:40 |
| happyaron | flexiondotorg_: I can look at the slideshow and maybe merge it, but you need someone else to get it uploaded | 09:40 |
| happyaron | ah it's already merged | 09:41 |
| flexiondotorg_ | happyaron, Yeah, my issue is getting stuff uploaded. | 09:41 |
| flexiondotorg_ | dholbach, Thanks! | 09:41 |
| flexiondotorg_ | dholbach, You'll remember your review ubuntu-mate-settings and ubuntu-mate-artwork with me last year. | 09:42 |
| flexiondotorg_ | dholbach, I met with cyphermox last week and we did another review together. | 09:42 |
| flexiondotorg_ | dholbach, So they are ready for upload I believe. | 09:42 |
| dholbach | ok | 09:42 |
| didrocks | flexiondotorg_: bluesabre needs to do a release for lightdm-gtk-greeter it seems (he's doing upstream releases for it it seems) | 09:44 |
| didrocks | flexiondotorg_: seems that it's mostly the xubuntu guys, so maybe contact them? | 09:44 |
| didrocks | or ochosi maybe : | 09:44 |
| didrocks | :) | 09:44 |
| * ochosi hides | 09:44 | |
| flexiondotorg_ | didrocks, I have been in contact with Xubuntu team last night. | 09:44 |
| didrocks | spotted! | 09:44 |
| ochosi | gah :) | 09:44 |
| flexiondotorg_ | ochosi, Hello again. | 09:44 |
| ochosi | yeah, the greeter will get a 2.0 release and subsequent release soon | 09:45 |
| flexiondotorg_ | didrocks, I will chat with bluesabre later. | 09:45 |
| dholbach | responded on ~profzoom/lightdm-gtk-greeter/add-mate-badge | 09:45 |
| ochosi | we have all the features we want there, so it's just a matter of getting around to it | 09:45 |
| ochosi | dholbach: hmm, that MR is merged already though.. :) | 09:46 |
| dholbach | ok | 09:47 |
| dholbach | ah ok | 09:47 |
| flexiondotorg_ | dholbach, Replied to ~profzoom/lightdm-gtk-greeter/add-mate-badge | 09:48 |
| flexiondotorg_ | Regarding the gtk2 patch do you need a complete debdiff? e.g. new debian/changelog entry + patch itself + debian/patches/series diff (which specifies the patch file on the last line) | 09:48 |
| didrocks | flexiondotorg_: interesting, the slideshow has a release tag for 92 but didn't get uploaded to the archive | 09:48 |
| didrocks | (and more commits afterwards) | 09:49 |
| flexiondotorg_ | didrocks, Thanks for helping with this. Really appreciate it. | 09:49 |
| didrocks | flexiondotorg_: no worry, just having one remaining to do help :) | 09:50 |
| didrocks | flexiondotorg_: I'm trying to build this rev 92 (which contains your patch) and check | 09:50 |
| flexiondotorg_ | didrocks, Cheers. | 09:50 |
| flexiondotorg_ | dholbach, Regarding the gtk2 patch do you need a complete debdiff? e.g. new debian/changelog entry + patch itself + debian/patches/series diff (which specifies the patch file on the last line) | 09:51 |
| dholbach | flexiondotorg_, an entry in d/changelog, the patch in d/patches and the patch being mentioned in d/patches/series (http://packaging.ubuntu.com/html/patches-to-packages.html has more info) - in the indicator-sound-gtk2 package, the change was applied directly in the source | 09:53 |
| dholbach | ... or another option would have been to submit the patch directly to the "upstream" source | 09:53 |
| dholbach | and get a new release made | 09:53 |
| dholbach | commented on the badge proposal again | 09:54 |
| === vrruiz_ is now known as rvr | ||
| flexiondotorg_ | dholbach, I've added license information to ~profzoom/lightdm-gtk-greeter/add-mate-badge | 10:26 |
| ochosi | flexiondotorg_: humm, actually you can let dholbach off the hook since there will be a release for the greeter soon and your changes will be included ;) | 10:27 |
| ochosi | if you wanna add license information, please file a separate/new MR | 10:27 |
| dholbach | ok | 10:27 |
| dholbach | I'll leave you to it | 10:27 |
| flexiondotorg_ | ochosi, Thanks! | 10:28 |
| flexiondotorg_ | dholbach, Regarding the gtk2 patch do you need a complete debdiff? e.g. new debian/changelog entry + patch itself + debian/patches/series diff (which specifies the patch file on the last line) | 10:29 |
| didrocks | flexiondotorg_: uploaded slideshow rev 92, which should contain your fix | 10:31 |
| flexiondotorg_ | didrocks, Thanks. | 10:31 |
| dholbach | flexiondotorg_, I replied to that already, no? | 10:31 |
| dholbach | flexiondotorg_, it's already uploaded | 10:31 |
| flexiondotorg_ | dholbach, You've already uploaded the patched gtk2? 😃 | 10:32 |
| dholbach | sorry | 10:32 |
| dholbach | the indicator thing | 10:32 |
| dholbach | which gtk2 patch | 10:32 |
| dholbach | https://code.launchpad.net/~ubuntu-mate-dev/indicators-gtk2/indicator-application-gtk2/+merge/244116? | 10:33 |
| dholbach | or something different? | 10:33 |
| flexiondotorg_ | dholbach, Thanks for doing the Indicators. | 10:33 |
| dholbach | flexiondotorg_, I still don't understand | 10:34 |
| dholbach | flexiondotorg_, can you give me a link to the merge proposal your question was about? | 10:34 |
| flexiondotorg_ | dholbach, I have you at cross purposes. Sorry. | 10:35 |
| flexiondotorg_ | Can I jsut confirm have you uploaded the following? | 10:35 |
| flexiondotorg_ | https://code.launchpad.net/~ubuntu-mate-dev/indicators-gtk2/indicator-sound-gtk2/+merge/244120 | 10:35 |
| dholbach | yes | 10:35 |
| dholbach | you should have gotten a mail about it | 10:35 |
| flexiondotorg_ | and https://code.launchpad.net/~ubuntu-mate-dev/indicators-gtk2/indicator-application-gtk2/+merge/244116? | 10:36 |
| dholbach | https://lists.ubuntu.com/archives/vivid-changes/2015-February/005355.html | 10:36 |
| dholbach | the latter, not yet | 10:36 |
| dholbach | I'm no desktop expert, so I'm not quite sure what's re-added and how it all works | 10:36 |
| flexiondotorg_ | dholbach, I'm on really shitty wifi here at the momment. | 10:36 |
| dholbach | and I was additionally confused because bug 1319352 was marked as 'fix released' | 10:37 |
| ubottu | bug 1319352 in ubuntu-mate "indicator-application does not work in MATE 1.8" [Wishlist,Fix released] https://launchpad.net/bugs/1319352 | 10:37 |
| flexiondotorg_ | dholbach, That fixed release ralte to Trusty via a PPA because MATE isn't in the official 14.04 archive. | 10:38 |
| flexiondotorg_ | *relates | 10:38 |
| dholbach | that's the bug mentioned in the changelog entry | 10:38 |
| flexiondotorg_ | dholbach, The indicator-application-gtk2 change has been tested by Xubuntu team and myself on stock Ubuntu, Xubuntu and Ubuntu MATE. | 10:39 |
| dholbach | right, thanks a lot for that | 10:39 |
| flexiondotorg_ | So, from testing we are happy it works and doesn't conflict. | 10:39 |
| dholbach | it's still hard for me (not being an expert) to find out from the MP (and links in there) what's going on | 10:40 |
| flexiondotorg_ | dholbach, Basically the .service files have been re-instated. | 10:40 |
| dholbach | flexiondotorg_, ok... I'll play around with it | 10:46 |
| dholbach | flexiondotorg_, what I'm just confused about is that the bug that is referenced in the MP is marked as 'fix released' - to me that makes it look like the issue is already fixed | 10:46 |
| flexiondotorg_ | dholbach Would you like me to change that bug status? | 10:47 |
| flexiondotorg_ | MATE is not in the official archive. So, I maintain it in a PPA. | 10:47 |
| dholbach | right | 10:47 |
| flexiondotorg_ | I submitted a fix to the PPA for Trusty. And marked it fixed. | 10:48 |
| flexiondotorg_ | Therefore, the "fix" exist in my PPAs only right now. | 10:48 |
| dholbach | but that's the way we used those "(Ubuntu)" bug tasks - if it's "fix released", it means "this is no issue in Ubuntu any more" | 10:48 |
| dholbach | I really don't want to be pedantic - I just wanted to explain what I thought when I looked that the MP and bug, trying to understand what needs to be done | 10:49 |
| flexiondotorg_ | dholbach, I understand. | 10:49 |
| flexiondotorg_ | I'l avoid introducing any confusion like that in the future. | 10:50 |
| dholbach | the indicator-application-gtk2 branch fails to build for me | 10:50 |
| === _salem is now known as salem_ | ||
| dholbach | maybe somebody in #ubuntu-desktop can help fix it? | 10:51 |
| flexiondotorg_ | dholbach, It build in a PPA. Got a build log? | 11:05 |
| dholbach | flexiondotorg_, http://paste.ubuntu.com/10202955/ | 11:06 |
| flexiondotorg_ | dholbach, How strange. Same build in a PPA worked just fine?! - https://launchpadlibrarian.net/194192337/buildlog_ubuntu-vivid-amd64.indicator-application-gtk2_12.10.0.1-0ubuntu3~15.04~1_UPLOADING.txt.gz | 11:11 |
| dholbach | I don't know | 11:12 |
| dholbach | sorry, maybe somebody in #ubuntu-desktop or #ubuntu-quality can help to fix it? | 11:13 |
| flexiondotorg_ | dholbach, I'll chase that one. Thanks for looking into it. | 11:13 |
| dholbach | thanks | 11:13 |
| flexiondotorg_ | dholbach, Can you look at ubuntu-mate-settings and ubuntu-mate-artwork? | 11:14 |
| dholbach | do you have a link to those? | 11:14 |
| flexiondotorg_ | Just looking to see if the original package requests are still knocking about.... | 11:15 |
| dholbach | or is it https://code.launchpad.net/~ubuntu-mate-dev/ubuntu-mate/ubuntu-mate-settings and https://code.launchpad.net/~ubuntu-mate-dev/ubuntu-mate/ubuntu-mate-artwork to be uploaded to the archive? | 11:16 |
| flexiondotorg_ | Yes, precisely that. | 11:16 |
| dholbach | were you waiting for a long time to get them uploaded now? | 11:16 |
| flexiondotorg_ | dholbach, Well you review them around September last year. | 11:17 |
| flexiondotorg_ | But we missed the merge window. | 11:17 |
| dholbach | ah no... what I meant was: the need to get them uploaded just came up now? | 11:17 |
| flexiondotorg_ | They have changed since then, slightly. And cyphermox reviewed them with me in person last week. | 11:17 |
| dholbach | ok... I didn't see a sponsoring request for them, which is why I asked | 11:17 |
| flexiondotorg_ | dholbach, No, I did have bug to upload them last year. Trying to find them. | 11:18 |
| dholbach | nevermind, I'll take a look now | 11:18 |
| flexiondotorg_ | But, I didn't know about the adding ubuntu-sponsors back then. | 11:18 |
| dholbach | these packages are not in the archive yet, right? | 11:19 |
| bregma | flexiondotorg_, don't worry about the compiz MP, you proposed it after everyone was done for the day yesterday: it;s still in line for getting in by FF (assuming it passes all the tests) | 11:21 |
| flexiondotorg_ | bregma, Thanks. | 11:21 |
| flexiondotorg_ | dholbach, Correct, ubuntu-mate-settings and ubuntu-mate-artwork have never been in the archive. | 11:22 |
| dholbach | ok | 11:22 |
| dholbach | flexiondotorg_, is ./usr/share/glib-2.0/schemas/zubuntu-mate-live.gschema.override supposed to be "zubuntu"? | 11:23 |
| flexiondotorg_ | dholbach, Yes. | 11:25 |
| flexiondotorg_ | I work with the upstream Debian packaging team and correct naming of gschema overrides didn't arrive before the Jessie freeze. | 11:26 |
| flexiondotorg_ | So, alpha-numerically this is how I have to inforce the live session overrides 😞 | 11:26 |
| dholbach | flexiondotorg_, I'm a bit unsure about the stuff going into etc/skel and if it couldn't conflict with other files at some stage | 11:28 |
| dholbach | and etc/xdg/autostart/ | 11:29 |
| dholbach | and etc/X11/Xsession.d | 11:29 |
| flexiondotorg_ | Nothing in etc/skel has conflicted during the 14.04 and 14.10 releases. | 11:30 |
| flexiondotorg_ | I discussed etc/xdg/autostart/ with cyphermox last week. | 11:31 |
| flexiondotorg_ | The nm-applet-mate.desktop is required to overcome an issue with patches in nm-applet for GNOME3 and Unity. | 11:31 |
| flexiondotorg_ | cyphermox, Has confirmed my work around is correct. | 11:32 |
| flexiondotorg_ | The issue is MATE is the only desktop, other than GNOME3, that actually parses AutoStartConditations. | 11:32 |
| tjaalton | pitti, Mirv: dropped libopenvg1-mesa-dev from qtbase-opensource-src{,-gles} BD so they build again | 11:32 |
| pitti | tjaalton: yay, thanks! | 11:32 |
| dholbach | regarding etc/skel: there you ship a default config file for tilda | 11:32 |
| flexiondotorg_ | But because MATE is not GNOME3, nm-applet.desktop as shipped will not start under MATE. | 11:33 |
| tjaalton | almost forgot again :P | 11:33 |
| pitti | tjaalton: nice, that'll clean up the last bit on http://people.canonical.com/~ubuntu-archive/nbs.html | 11:33 |
| dholbach | I know that it (right now) probably doesn't conflict with anything - I'm just wondering if the mate package is best place for this - normally something like this would be shipped with tilda itself | 11:33 |
| flexiondotorg_ | So, I provide a MATE specific nm-applet that will only run under MATE. | 11:33 |
| pitti | I cleaned up some days ago, but I didn't quite know what to do with that openvg bit | 11:33 |
| dholbach | for the stuff in autostart I was just wondering if the files should probably have something mate specific in their name? | 11:34 |
| flexiondotorg_ | dholbach, Tilda doesn't have a system level configuration. So this way the only way to integrate it. | 11:34 |
| Mirv | tjaalton: argh, that means I need to bump and rebuild my landing, but good | 11:34 |
| dholbach | I was talking about shipping a file in /etc/skel | 11:34 |
| tjaalton | pitti: ah, I was trying to find where they were showing up | 11:34 |
| flexiondotorg_ | It is only used at system install or when adding a new account, so won't clobber existing user confiurations. | 11:34 |
| dholbach | I understand | 11:34 |
| tjaalton | Mirv: right, it's a simple change at least | 11:34 |
| tjaalton | well, the uploads were | 11:35 |
| dholbach | what I'm saying is: let's say somebody else wants to ship some tilda configuration in etc/skel - in that case it'd make sense to agree on it in one canonical place - I just found it a bit confusing to see mate settings ship a default config file for tilda | 11:35 |
| flexiondotorg_ | dholbach, I see your point. | 11:36 |
| flexiondotorg_ | dholbach, ShouldI removed Tilda? | 11:36 |
| dholbach | maybe ask the folks in #ubuntu-desktop if there's another good way to do it? | 11:37 |
| dholbach | flexiondotorg_, do the yuyo files come from somewhere? | 11:37 |
| flexiondotorg_ | dholbach, I did check what other flavour are doing. I saw that Xubuntu and Lubuntu ship configs in their settings packages so assumed it was OK to do the same. | 11:38 |
| flexiondotorg_ | yuyo are mirroed from github. | 11:38 |
| flexiondotorg_ | https://github.com/snwh/yuyo-gtk-theme | 11:38 |
| dholbach | could you mention that in d/copyright? | 11:39 |
| flexiondotorg_ | Sure. | 11:39 |
| dholbach | cool | 11:40 |
| dholbach | ok, I found two files https://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/xubuntu-default-settings/vivid/files/head:/etc/skel/ | 11:40 |
| dholbach | but mayb still bring it up with the desktop team to get some advice | 11:40 |
| dholbach | I'm running out of time on a few other things now | 11:40 |
| flexiondotorg_ | dholbach, Thanks for helping. | 11:42 |
| dholbach | anytime | 11:43 |
| flexiondotorg_ | dholbach, Yuyo Source corrected. | 11:45 |
| flexiondotorg_ | dholbach, In GitHub at least. Will need to sync with LP. | 11:45 |
| dholbach | ok | 11:46 |
| flexiondotorg_ | Is ubuntu-mate-artwork in good shape? | 11:46 |
| flexiondotorg_ | Because if so, I can seek upload assistance elsewhere later. | 11:47 |
| flexiondotorg_ | With regard to ubuntu-mate-settings, bit concerned that will torpedo getting Ubuntu MATE official this cycle 😞 | 11:47 |
| dholbach | from what I've seen it looked good, it just needed an update of d/copyright to explain which code/assets are borrowed from elsewhere, under which license and who the copyright holders are | 11:48 |
| dholbach | send a mail to the desktop team list, ask the people there, and ask folks from other flavours too - in a few places I just wasn't sure if that was "the way to do it" | 11:48 |
| dholbach | lunch time, bbiab | 11:52 |
| === marcusto_ is now known as marcustomlinson | ||
| flexiondotorg_ | dholbach, ubuntu-mate-artwork copyright is updated accordingly. | 12:16 |
| tjaalton | Mirv: uh, looks like dpkg-gensymbols wasn't too happy about dropping libopenvg1-mesa-dev build-dep | 12:30 |
| tjaalton | or maybe unrelated, dunno | 12:31 |
| Mirv | tjaalton: I think there's possibly a recent toolchain change causing that. | 12:32 |
| Mirv | while preparing 5.4, I noted at one point I needed another symbols update | 12:32 |
| tjaalton | ok.. you'll sort it out? | 12:33 |
| Mirv | tjaalton: but it's only gles, right, the main package seems ok? | 12:33 |
| tjaalton | apparently so | 12:33 |
| tjaalton | still building the main package | 12:33 |
| Mirv | tjaalton: the amd64 is past the symbol checking stage | 12:33 |
| tjaalton | yeah just checked that myself :) | 12:34 |
| Mirv | tjaalton: I'm landing Qt 5.4.0 (all of https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-005/+packages) on Monday anyway, but I guess there's only one symbol that'd need fixing on the -gles package | 12:34 |
| tjaalton | ah ok, nice | 12:35 |
| Mirv | those optional ones are on purpose there and part of the normal/-gles packaging, so it's the first symbol in the log that's actually wrong | 12:36 |
| Mirv | I can upload a fix 5.3.2 -gles for that | 12:36 |
| tjaalton | great | 12:40 |
| tjaalton | actually the main package failed too | 12:41 |
| tjaalton | on amd64 | 12:41 |
| Mirv | hmh, and the gles did not on i386 | 12:43 |
| Mirv | flaky symbols, my favorites | 12:44 |
| Mirv | so all build logs would be needed to make sure which archs have that one symbol and which do not | 12:44 |
| Mirv | tjaalton: so if you have time, wait the needed 2-3h still and then manipulate that single symbol only adding !amd64 etc as needed, and reupload. otherwise I'll probably simply land 5.4.0 on Monday. | 12:47 |
| tjaalton | Mirv: probably just wait for monday then, I'll be gone in 1h :/ | 12:54 |
| Kano | hi, why does systemd-analyze not work in a current vivid live iso | 12:55 |
| Mirv | tjaalton: yeah, I've the same problem | 12:55 |
| xnox | Kano: beacause live-cd boots with upstart by default. | 13:04 |
| Kano | why? | 13:05 |
| Kano | i would remove upstart | 13:05 |
| xnox | Kano: because ubuntu hasn't switched to systemd yet. | 13:05 |
| xnox | Kano: it's not like there is multi-year effort to achieve that, pleora of things to fix and integrate still, and have working.... | 13:05 |
| Kano | lost time with upstart it seems | 13:06 |
| Kano | next is mir ;) | 13:06 |
| rbasak | hallyn: are you aware of cgmanager being stuck in vivid-proposed? lxc dep8 failure. | 13:10 |
| rbasak | stgraber: ping, for when you get in, for kimchi NEW | 13:19 |
| rbasak | stgraber: I have some questions - the developer is asking whether some issues are blockers for acceptance, and I'd like an answer from the person who will actually review it :) | 13:20 |
| rbasak | stgraber: who gaughen says is you? | 13:20 |
| rbasak | pitti: may I have some help with adt-run --changes please? I have built already using sbuild. Using just --changes seems to not do anything. | 13:23 |
| rbasak | But the changes file doesn't include the source. | 13:23 |
| rbasak | So do I need to specify the .dsc as well? In which case, which order? | 13:24 |
| rbasak | I ask because it's taken dozens of minutes each try. | 13:24 |
| rbasak | If you're around :) | 13:24 |
| * rbasak tries --changes -B .dsc | 13:25 | |
| === MacSlow is now known as MacSlow|lunch | ||
| stgraber | rbasak: hi | 13:36 |
| stgraber | rbasak: I just read your e-mail (gaughen forwaded it to me) | 13:37 |
| pitti | rbasak: re from lunch; what did you run exactly? and does your .changes include the source? | 13:42 |
| pitti | rbasak: if you built without source, you have to give the .dsc explicitly | 13:42 |
| pitti | rbasak: otherwise, adt-run foo.changes foo.dsc --- ... | 13:43 |
| rbasak | pitti: I built with sbuild, so I have an amd64 changes. I suppose I do have a separate source changes. | 13:45 |
| rbasak | stgraber: arch-dep-package-has-big-usr-share has come up, because he switched to arch: any. He was arch: all, but a commenter on the ITP requested more specific per-arch qemu dependencies, and apparently you can't have that with arch: all (I didn't know that before this). | 13:48 |
| pitti | rbasak: that's sbuild's -s option | 13:48 |
| pitti | rbasak: but no matter, giving the .changes (or just *.deb which is pretty much equivalent) and then the .dsc should work fine | 13:48 |
| rbasak | pitti: ah, thanks. I didn't know about that, not being a Debian uploader. I never use binary changes files for anything else ;) | 13:49 |
| pitti | rbasak: right, it's the standard thing for a DD | 13:49 |
| stgraber | rbasak: right. I think that's fine to override, another way around it would be to split the /usr/share stuff into a -data package which is arch:all and depend on it | 13:49 |
| pitti | rbasak: for ubuntu I mostly run adt-run *.deb *.dsc in my /tmp/build-area/ | 13:49 |
| rbasak | stgraber: so you'd accept this without a -data split? | 13:50 |
| rbasak | stgraber: finally, it's a web application, depends on nginx, which sets up a "system" default nginx on port 80, but he doesn't use that - he starts his own nginx daemon instance on a high port. | 13:51 |
| stgraber | rbasak: I'd be fine with it yeah. Splitting to -data would basically make the main package empty which to me just makes things more confusing. Since you're not likely to want to co-install multiple arches of the main one (which would be a valid reason for the split), I think it's fine to override. | 13:51 |
| rbasak | stgraber: OK, thanks. Do you want specific overrides for these tags explaining the reasoning? | 13:51 |
| rbasak | stgraber: I don't like that at all. I suggested integrating properly with nginx, using conf-available et al, and providing the web app on http(s?):localhost/kimchi | 13:52 |
| rbasak | He wanted to know if that would be a blocker, or if he can fix it later. | 13:52 |
| stgraber | rbasak: yeah, I usually like to see all packages I review be "lintian -iI --pedantic" clean by providing an lintian-overrides file which overrides any irrelevant warning with a comment as to the why | 13:53 |
| rbasak | OK | 13:53 |
| rbasak | pitti: thanks. I understand this better now. My run of --changes -B .dsc seemed to work, but I guess the -B wasn't necessary there. | 13:54 |
| stgraber | hmm, and I don't suppose there's a binary-only package for nginx (similar to dnsmasq-base vs dnsmasq)? | 13:54 |
| rbasak | No. That was my first thought, but there isn't. | 13:54 |
| rbasak | I don't want to introduce an Ubuntu delta just for that. We could request it in Debian though. | 13:54 |
| rbasak | http://webapps-common.alioth.debian.org/draft/html/index.html suggests using http://localhost/<webapp> though. | 13:55 |
| stgraber | hmm, so my concern then is that you'll have a change in behavior when you fix the issue as it'll be on a high port until then and will then switch to /<webapp> on upgrade | 13:55 |
| rbasak | Unless there's a specific need to have another web server daemon instance running, it seems friendlier to the user to me to do it that way. | 13:55 |
| stgraber | so I'd rather we do the right thing from the start | 13:55 |
| rbasak | Yeah - I did wonder about that. | 13:55 |
| rbasak | Probably makes it unsuitable for changing after feature freeze. | 13:55 |
| stgraber | well, for a new package, getting an FFe isn't very difficult typically, but still, if we intend on making noise about this when it hits the archive, changing behavior a week later is suboptimal | 13:56 |
| rbasak | So he wanted to know if this is a blocker for upload. As you're also concerned, I guess the answer is yes? | 13:56 |
| stgraber | yeah, you'd need to provide me with a very good reason why the high port thing is needed for me to be fine with this and so far, there doesn't appear to be one | 13:57 |
| rbasak | OK, thanks! | 13:57 |
| rbasak | There were other minor issues but they shouldn't be any issue to fix up quickly. | 13:57 |
| zul | can someone promote python-tempest-lib (#1420006) please its blocking the nova build | 14:18 |
| flexiondotorg_ | cyphermox, Are you available? | 14:24 |
| === elmo_ is now known as elmo | ||
| stokachu | @pilot in | 14:46 |
| === udevbot changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support or app devel) | build failures -> http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of lucid -> utopic | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: stokachu | ||
| cyphermox | flexiondotorg_: sure, what's up? | 14:52 |
| flexiondotorg_ | cyphermox, How are you fixed today? | 14:52 |
| flexiondotorg_ | cyphermox, I got some stuff done with dholbach earlier. | 14:52 |
| cyphermox | yeah, I read backlog | 14:53 |
| flexiondotorg_ | cyphermox, I think dholbach is now on to other tasks. | 14:53 |
| * dholbach hugs stokachu | 14:53 | |
| dholbach | yes I am | 14:53 |
| flexiondotorg_ | cyphermox, Can you pick up the baton? | 14:53 |
| dholbach | stokachu can maybe also help out - he just went on review duty - but I'll leave you guys to it | 14:54 |
| stokachu | o/ | 14:54 |
| flexiondotorg_ | stokachu, Hi | 14:55 |
| flexiondotorg_ | stokachu, If you have some time for package uploading please let me know. I'm the Ubuntu MATE dev and trying to get everything I need in the official archive. | 14:55 |
| stokachu | flexiondotorg_: hey there | 14:55 |
| pitti | apw: I noticed you have a "merge initramfs-tools: INPROGRESS" on https://blueprints.launchpad.net/ubuntu/+spec/core-1411-systemd-migration | 14:56 |
| stokachu | flexiondotorg_: sweet! what can i help with | 14:56 |
| pitti | apw: question 1: will that still actually happen for vivid, as it's quite intrusive? | 14:56 |
| flexiondotorg_ | I have a package that hasn't been reviewed yet. | 14:56 |
| pitti | apw: question 2: how does that affect the systemd transition? | 14:56 |
| flexiondotorg_ | And a couple that have been reviewed a couple of time are most likely ready for upload. | 14:57 |
| pitti | apw: i. e. was that just a convenient place to record "merging i-t is overdue", or is there some actual effect on booting there? | 14:57 |
| flexiondotorg_ | StevenK, Which do you fancy? | 14:57 |
| flexiondotorg_ | StevenK, sorry. | 14:57 |
| flexiondotorg_ | stokachu, Which do you fancy. | 14:57 |
| stokachu | flexiondotorg_: lets do the uploads first | 14:57 |
| pitti | apw: the one thing that comes to my mind there is mounting /usr from the initramfs (as in 0.118 in Debian) | 14:57 |
| stokachu | flexiondotorg_: then ill review the other package | 14:57 |
| pitti | apw: but systemd has several patches to deal with mounting /usr from the real system, and 0.118 still didn't promote in Debian because apparently there was some fallout? | 14:58 |
| flexiondotorg_ | stokachu, https://code.launchpad.net/~ubuntu-mate-dev/ubuntu-mate/ubuntu-mate-artwork | 14:58 |
| flexiondotorg_ | That has been review by cyphermox and dholbach. All feedback implemented. | 14:58 |
| stokachu | flexiondotorg_: is there a MP? | 14:59 |
| cyphermox | stokachu: it's a new package ;) | 14:59 |
| stokachu | is there a new package bug somewhere? | 15:01 |
| pitti | tkamppeter: hallo, wie gehts? | 15:03 |
| pitti | tkamppeter: did you notice the cups test regression in https://jenkins.qa.ubuntu.com/job/vivid-adt-cups/75/ARCH=amd64,label=adt/console? that holds the last cups upload in -proposed | 15:03 |
| pitti | tjaalton: hm, the qt uploads are FTBFS.. | 15:04 |
| === MacSlow|lunch is now known as MacSlow | ||
| pitti | yay C++ symbol mangling | 15:05 |
| flexiondotorg_ | stokachu, There was at some point or though dholbach and I could track it down earlier. | 15:06 |
| flexiondotorg_ | stokachu, If you need a bug against what should I file it? | 15:06 |
| stokachu | flexiondotorg_: https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages#Requesting_a_new_package_for_Ubuntu | 15:07 |
| stokachu | just need a new package bug filed | 15:07 |
| stokachu | you dont need to worry about debian as this is ubuntu specific | 15:07 |
| stokachu | flexiondotorg_: heres an eg: https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages/ExamplePackageRequest | 15:08 |
| flexiondotorg_ | stokachu, Thanks. | 15:08 |
| stokachu | flexiondotorg_: no thank you :) and please let me know when those are done and ill get started on it | 15:08 |
| hallyn | rbasak: sigh | 15:08 |
| stokachu | flexiondotorg_: you'll want to create one for each new package you need uploaded | 15:09 |
| stokachu | flexiondotorg_: assuming they don't exist yet in ubuntu | 15:09 |
| flexiondotorg_ | stokachu, Hah. Here is the original bug 😃 | 15:09 |
| flexiondotorg_ | https://bugs.launchpad.net/ubuntu/+bug/1368218 | 15:09 |
| ubottu | Launchpad bug 1368218 in Ubuntu "[needs-packaging] ubuntu-mate-artwork and ubuntu-mate-settings" [Wishlist,New] | 15:09 |
| stokachu | flexiondotorg_: ah ok perfect | 15:09 |
| hallyn | rbasak: i'm at the publishing history - how do i get to the lxc dep8 failure results page? | 15:11 |
| rbasak | hallyn: go from http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html | 15:11 |
| === roadmr is now known as roadmr_afk | ||
| === kickinz1 is now known as kickinz1|afk | ||
| hallyn | stgraber: I don't supose https://jenkins.qa.ubuntu.com/job/vivid-adt-lxc/lastBuild/ARCH=amd64,label=adt/console is a known transient failure? | 15:14 |
| === kickinz1|afk is now known as kickinz1 | ||
| stokachu | flexiondotorg_: ok great, give me a few minutes and ill work on these 2 | 15:15 |
| flexiondotorg_ | stokachu, Thanks! | 15:15 |
| stgraber | hallyn: oh right, I meant to look into that. Looks like it's stuck on test-ubuntu which is known to take quite a while. | 15:16 |
| cyphermox | flexiondotorg_: you'll also need one for the metapackage | 15:16 |
| stgraber | hallyn: it may be that a firewall change or similar is now preventing the jenkins builder from reaching cloud-images.u.c | 15:16 |
| flexiondotorg_ | cyphermox, Yep and yuyo. | 15:17 |
| cyphermox | indeed. | 15:17 |
| flexiondotorg_ | I'm going to file them shortly 😃 | 15:17 |
| hallyn | stgraber: ok, so you know that bc it just runs them alphabetically? | 15:17 |
| stgraber | hallyn: yep | 15:18 |
| stgraber | hmm, so it hanged quite a few times, certainly not transient... | 15:20 |
| stgraber | lets see if I can reproduce the issue in kvm here (since we clearly don't hit that problem in LXC CI) | 15:20 |
| hallyn | trying to reproduce locally with the new cgmanager | 15:20 |
| stgraber | it's been failing since the first LXC 1.1 upload, so it's not caused by cgmanager, my guess is that it may be lxcfs | 15:21 |
| stgraber | or a firewall/proxy change in Canonical CI | 15:21 |
| hallyn | oh yeah, if i can clear other things i need to track downthat lxcfs issue | 15:22 |
| valavanisalex | Hi there, can someone please advise on why/when packages in main should use dh_translations? I implemented in in the Inkscape package ages ago as a replacement for some much older translations handling code http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/inkscape/vivid/revision/6 although I've never fully understood why we needed it | 15:23 |
| tkamppeter | pitti, I got the notification but did not set up VPN access yet here. I never had a triggering of this kind of test. What happened exactly. | 15:31 |
| pitti | tkamppeter: you don't need VPN access, the jenkins.qa.ubuntu.com URL is public | 15:32 |
| pitti | tkamppeter: latest cups upload fails its autopkgtest, while the previous versions succeeded; so it's a regression and britney holds it back (http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#cups) | 15:33 |
| hallyn | stgraber: tests passed here... | 15:37 |
| stokachu | flexiondotorg_: building now, ill ping you once ive tested these 2 | 15:37 |
| flexiondotorg_ | stokachu, Thanks. | 15:37 |
| === pgraner is now known as pgraner-afk | ||
| stgraber | hallyn: same here | 15:42 |
| stgraber | hallyn: so I'm going to blame the Canonical CI environment and push an override | 15:43 |
| stgraber | pitti: FYI ^ | 15:43 |
| hallyn | stgraber: thx | 15:43 |
| hallyn | rbasak: say, in uvtool, have youconsidered tuning the backing store files? | 15:43 |
| rbasak | hallyn: tuning how? | 15:44 |
| hallyn | i.e. preallocation=metadata,compat=1.1,lazy_refcounts=on is supposed to be a nice speed increase | 15:44 |
| === roadmr_afk is now known as roadmr | ||
| rbasak | Oh. I didn't know about that. | 15:44 |
| hallyn | also i've tried forcing uvtool to give me 'cache=none' by default, but it doesn't seem to work. (even wehn i hard-code 'unsafe=true' in the code) | 15:44 |
| hallyn | rbasak: i'm not sure how you're creating the files, i.e. if you're using a library that doesn't expose those... | 15:45 |
| hallyn | just thought i'd bring it up | 15:45 |
| rbasak | I'm downloading the files from the simplestreams source. | 15:45 |
| hallyn | since i saw it | 15:45 |
| hallyn | since i say my disk is compat-0.10 | 15:45 |
| rbasak | Is there any downside to this tuning? I wonder if we should be publishing from cloud-images like this already if not. | 15:45 |
| rbasak | uvt-kvm has an --unsafe-caching option, which messes with the template XML provided. Maybe that's affecting it? | 15:46 |
| hallyn | well, precise couldn't do compat=1.1 i guess, only 1.0 | 15:46 |
| rbasak | I didn't notice much of a speedup with that though. | 15:46 |
| rbasak | I see. | 15:46 |
| hallyn | did you look at the xml afte rusing that option? | 15:46 |
| * hallyn tries again | 15:46 | |
| hallyn | you don't actually support precise right? | 15:46 |
| pitti | stgraber: for lxc? well, the tests passed before.. | 15:47 |
| rbasak | So on releases that support it, I could rewrite the image to make it faster? That sounds like it's worth doing - at least as an option, or by default with an option to disable or something. | 15:47 |
| Daviey | I've done some qemu testing, and unsafe io has a signficant perf' improvement (obv. with the bad karma attached to it) | 15:47 |
| hallyn | rbasak: looks like i messed up before, bc unsafe-caching is working for me, thx | 15:47 |
| pitti | hallyn, stgraber: I overrode the lxc failure for cgmanager and dnsmasq already, but for lxc itself it's a real test regression | 15:47 |
| rbasak | I do support precise, sort of. Juju uses uvtool on precise for KVM support. But the feature set needed is only a subset of developer CLI usage. | 15:47 |
| stgraber | pitti: well, both hallyn and I managed to get LXC to pass fine under adt on our own network | 15:47 |
| stgraber | pitti: so in theory, exact same kvm environment as Canonical CI | 15:48 |
| stgraber | pitti: since it's hanging on a test which pulls packages and cloud images over the network, I'm tempted to blame a firewall/proxy change | 15:48 |
| hallyn | well no mine was just a handbuilt vm | 15:48 |
| pitti | stgraber: the host on CI is trusty, i. e. it's trusty's QEMU; and then there's the network proxy; maybe it's related to either of that? | 15:48 |
| stgraber | ah, ok, well, mine was proper adt-run using a vivid kvm | 15:48 |
| pitti | stgraber: ah, so proxy then? | 15:48 |
| pitti | so what changed in the test or in LXC wrt. proxy handling? | 15:49 |
| stgraber | yeah, that's my guess. Since I can't reproduce it here it's hard to tell, but my best guess is that it's hanging on the wget of the cloud image, passes the timeout and gets killed | 15:49 |
| stgraber | nothing, the tests haven't changed since rc1 | 15:49 |
| hallyn | Daviey: yeah especially on apt-get dist-upgrades (especially with kernel updates) cache=none or cache=unsafe is a huge difference | 15:49 |
| stgraber | and I also need to use a proxy in my home environment so that part sure is working | 15:49 |
| stgraber | my guess is that what changed is Canonical's network/proxy | 15:50 |
| rbasak | hallyn: I filed https://bugs.launchpad.net/uvtool/+bug/1421690 for your optimisation suggestion - thanks. | 15:50 |
| hallyn | rbasak: might be worth finding time to do some perf measurements with some of those tunables | 15:50 |
| ubottu | Launchpad bug 1421690 in uvtool "Images could be more optimised for better performance" [Wishlist,Confirmed] | 15:50 |
| hallyn | rbasak: cool, thanks | 15:50 |
| hallyn | btw http://sevennet.org/2015/01/08/how-to-incredibly-slow-kvm-disk-performance-qcow2-disk-files-virtio-programming-dev-development/ was one example reference | 15:50 |
| pitti | stgraber: hm, doesn't work very well on my host either (vivid, no proxy): http://paste.ubuntu.com/10206330/ | 15:51 |
| stgraber | odd | 15:52 |
| stgraber | oh, I see one problem with my manual test at least, let me rerun that | 15:52 |
| pitti | that's different than the timeout on CI, though; so it could certainly be that the proxy in the DC changed in some way | 15:53 |
| Daviey | hallyn: I also experimented with using eatmydata within the guest and saw *some* improvement (so lie about fsync's within the guest and outside), but it was inconsistent.. probably just skew that would averaged out if watched long enough | 15:53 |
| pitti | stgraber: the above error happens in vivid too, though (~rc4) | 15:53 |
| stgraber | pitti: I wrongly assumed that adt-run would use -proposed... how do I tell it to run using vivid-proposed? | 15:53 |
| rbasak | stgraber: --apt-pocket=proposed -U | 15:54 |
| pitti | stgraber: --apt-pocket=proposed -U | 15:54 |
| pitti | stgraber: well, if you don't specify anything it will use whatever is configured in the VM | 15:54 |
| pitti | stgraber: but if you just used adt-buildvm-ubuntu-cloud, that doesn't have -proposed enabled by default | 15:54 |
| stgraber | right, that's what I used | 15:54 |
| stgraber | ok, so rc4 was clean here, retrying with 1.1.0 now | 15:55 |
| stgraber | hmm, the VM fails to upgrade | 15:56 |
| hallyn | Daviey: oh, on btrfs eatmydata is the difference between 1 min and 20 mins | 15:56 |
| pitti | stgraber: are "/usr/sbin/deluser: The user `lxcunpriv' does not exist." and "ERROR: Unable to fetch GPG key from keyserver." expected, or causes for the failure that I see? | 15:57 |
| hallyn | stgraber: fwiw i did use proposed in my test | 15:57 |
| Daviey | hallyn: in tandem with unsafe cache or alone? | 15:57 |
| pitti | but for sure "container creation template for tmp.fc9zKfXCVv failed" is a hard fail | 15:57 |
| stgraber | pitti: the GPG error is fatal | 15:57 |
| stgraber | pitti: http://paste.ubuntu.com/10206408/ | 15:57 |
| cjwatson | bdmurray: The Contents publisher race should be fixed now, so I expect it to be reliably updated as of tomorrow morning's run. Please let me know if you see it falling behind again. | 15:57 |
| pitti | stgraber: uh, how old is that VM? that bug was fixed ages ago | 15:58 |
| flexiondotorg_ | stokachu, This would be next I guess - https://bugs.launchpad.net/ubuntu/+bug/1421694 | 15:58 |
| stgraber | pitti: about 30min old | 15:58 |
| ubottu | Launchpad bug 1421694 in Ubuntu "[needs-packaging] grub2-themes-ubuntu-mate" [Undecided,New] | 15:58 |
| pitti | stgraber: ok, which autopkgtest version do you have? | 15:58 |
| stgraber | pitti: whatever's in trusty | 15:58 |
| pitti | this was fixed in 2.17 | 15:58 |
| stgraber | got 2.14 | 15:59 |
| pitti | during utopic when we merged the new sysvinit, which now requires LSB headers | 15:59 |
| pitti | ok, so that explains that bit | 15:59 |
| stgraber | is there a more recent version of autopkgtest I can use on trusty? (also, please SRU) | 15:59 |
| pitti | stgraber: http://archive.ubuntu.com/ubuntu/pool/main/a/autopkgtest/autopkgtest_3.9.5git1_all.deb will work fine on trusty | 15:59 |
| stgraber | pitti: I guess I've got to rebuild the VM image after installing that right? | 16:00 |
| pitti | stgraber: happy to upload current vivid to trusty, but it's a helluva change | 16:00 |
| stokachu | flexiondotorg_: cool will build those next | 16:00 |
| pitti | stokachu: right | 16:00 |
| stokachu | pitti: \o/ | 16:00 |
| pitti | err, stgraber: right | 16:00 |
| pitti | stokachu: sorry :) sometimes weechat's tab completion is too clever :) | 16:00 |
| stokachu | pitti: haha no worries | 16:01 |
| seb128 | shell question | 16:01 |
| seb128 | if used [ ... -o ... ] | 16:01 |
| seb128 | is the second condition evaluated even if the first one is true? | 16:02 |
| pitti | I don't think shell can even have a concept of short-circuiting, as it replaces variables and substitutions first | 16:02 |
| seb128 | pitti, well, if you do [ ... ] || [ ... ] it seems to be different? | 16:03 |
| pitti | seb128: it depends on what you put there really | 16:03 |
| seb128 | pitti, context is https://code.launchpad.net/~mathieu-tl/mtp/lp1421664/+merge/249669 | 16:03 |
| seb128 | so it's | 16:04 |
| seb128 | if [ -z "$disconnected" ] || [ "$disconnected" -ge 1 ]; | 16:04 |
| stgraber | I'd expect test to return immediately after the first, but as pitti said, if you're using command output or similar, those are parsed before anything else so will run anyway | 16:04 |
| seb128 | vs | 16:04 |
| seb128 | if [ -z "$disconnected" -o "$disconnected" -ge 1 ]; | 16:04 |
| seb128 | the second version returns an illegal argument in case the variable is emptyu | 16:04 |
| seb128 | illegal argument +error | 16:04 |
| seb128 | the first one works | 16:05 |
| ogra_ | because you call -o "" -ge 1 | 16:05 |
| stgraber | right, because: "test -z "" -o "" -ge 1" isn't valid | 16:05 |
| seb128 | ogra_, right, and in the other case as well? | 16:05 |
| seb128 | is [ "" -ge 1 ] valid? | 16:05 |
| ogra_ | yeah | 16:05 |
| seb128 | but "" is not a number? | 16:05 |
| pitti | no, it's not numeric | 16:05 |
| ogra_ | -o expects some content | 16:05 |
| stgraber | in the other case you call 'test -z "" || test "" -ge 1' so you never call the second's test argument parser | 16:06 |
| ogra_ | tthat too | 16:06 |
| stokachu | flexiondotorg_: hate to ask this but do you mind creating another needs-packaging but for just ubuntu-mate-settings | 16:06 |
| pitti | seb128: how about [ "${disconnected:-0}" -ge 1 ] ? | 16:06 |
| stokachu | flexiondotorg_: im finishing up ubuntu-mate-artwork now | 16:06 |
| flexiondotorg_ | stokachu, Sure thing. | 16:06 |
| hallyn | Daviey: oh, sorry - this was in containers actually | 16:06 |
| stokachu | bug* | 16:06 |
| hallyn | (eatmydata on btrfs) | 16:06 |
| seb128 | pitti, well, the [ ] || [ ] version works, I was just trying to understand the difference | 16:06 |
| stokachu | flexiondotorg_: awesome thank you | 16:06 |
| pitti | seb128: yeah, [ -o ] is one test command which is syntactially invalid | 16:07 |
| pitti | seb128: while the || are two commands, and the second isn't evaluated at all if it's empty | 16:07 |
| pitti | seb128: so that will also do what you want indeed | 16:07 |
| seb128 | pitti, ok, makes sense | 16:07 |
| seb128 | stgraber, ogra_, pitti, thanks | 16:07 |
| pitti | seb128: although using a default value might be more concise | 16:07 |
| ogra_ | but even "" -ge 1 should work without error | 16:07 |
| pitti | no, it shouldn't | 16:07 |
| stgraber | seb128: [ blah ] is a shell convenience syntax for "test blah" so the difference is in the first case, the test built-in command does the parsing of the whole condition, in the second, you do two calls to shell with the second conditional on the first passing | 16:07 |
| pitti | an empty string isn't a number | 16:08 |
| pitti | $ test "" -ge 1 | 16:08 |
| pitti | bash: test: : Ganzzahliger Ausdruck erwartet. | 16:08 |
| pitti | (bah, German FTW!) | 16:08 |
| ogra_ | oh, k | 16:08 |
| * ogra_ fires up google translate | 16:08 | |
| ogra_ | :P | 16:08 |
| seb128 | lol | 16:09 |
| flexiondotorg_ | stokachu, https://bugs.launchpad.net/ubuntu/+bug/1421703 | 16:10 |
| ubottu | Launchpad bug 1421703 in Ubuntu " [needs-packaging] ubuntu-mate-settings" [Undecided,New] | 16:10 |
| stokachu | flexiondotorg_: thank you! | 16:10 |
| bdmurray | cjwatson: will do, thanks for fixing that! | 16:12 |
| flexiondotorg_ | stokachu, I need to run some errands. I'll be back later. | 16:13 |
| stokachu | flexiondotorg_: np ill ping you when ive done the 2 and reviewed that 3rd one | 16:13 |
| stokachu | if the 3rd package are easy fix ill go ahead and make the changes | 16:14 |
| flexiondotorg_ | stokachu, Thanks! | 16:14 |
| stokachu | flexiondotorg_: thank you :) | 16:14 |
| flexiondotorg_ | stokachu, I've got a couple more packages, one is the meta package and one is a theme. | 16:15 |
| flexiondotorg_ | The theme is a bit busted. Should be fixed by Monday. | 16:15 |
| flexiondotorg_ | As for the meta package, it is currently hooked up to my PPAs. Which I assume need unpicking? | 16:15 |
| rbasak | pitti: how much RAM does the adt-virt-qemu get in our infrastructure, please? I'm hitting a failure that only affects me locally, and we suspect RAM size. But I just tried 2G, and that didn't seem to fix it. | 16:16 |
| stokachu | flexiondotorg_: is the meta package ready for some review? | 16:18 |
| flexiondotorg_ | stokachu, Sure - https://code.launchpad.net/~ubuntu-mate-dev/ubuntu-mate/ubuntu-mate-meta | 16:19 |
| stokachu | flexiondotorg_: cool, please file a needs packaging bug on that one and ill review it as well today | 16:19 |
| flexiondotorg_ | stokachu, Nice one! | 16:19 |
| stokachu | i can at least get that one started if there are changes to be made | 16:20 |
| stgraber | pitti: adt really doesn't like me today... http://paste.ubuntu.com/10206702/ | 16:21 |
| tkamppeter | piti, how can I test this before upload and where do I find the autopkgtest so that I can perhaps adapt it to newer features of CUPS? | 16:22 |
| flexiondotorg_ | stokachu, https://bugs.launchpad.net/ubuntu/+bug/1421711 | 16:23 |
| ubottu | Launchpad bug 1421711 in Ubuntu "[needs-packaging] ubuntu-mate-meta" [Undecided,New] | 16:23 |
| stokachu | flexiondotorg_: perfect, thanks! | 16:24 |
| stokachu | bdmurray: mind flipping the bit on https://bugs.launchpad.net/ubuntu/+source/python-netaddr/+bug/1421375? | 16:27 |
| ubottu | Launchpad bug 1421375 in python-netaddr (Ubuntu Trusty) "AttributeError when attempting to access _sys.maxint" [High,Fix committed] | 16:27 |
| flexiondotorg_ | cyphermox, stokachu Is on a roll. Anything else I need to do for the build system? | 16:28 |
| stokachu | flexiondotorg_: nah man you good, ill ping you if i need anything else | 16:28 |
| === pgraner-afk is now known as pgraner | ||
| flexiondotorg_ | Cool. | 16:28 |
| stokachu | thanks again | 16:28 |
| cyphermox | flexiondotorg_: the artwork, setiings, meta and yuyo packages are the main things i believe | 16:30 |
| flexiondotorg_ | cyphermox, Well MATE Tweak is really important. But we might get that in via Debian. | 16:31 |
| flexiondotorg_ | cyphermox, If not then hopefully we can get an exception in Ubuntu. | 16:31 |
| flexiondotorg_ | I'll check in irc later and over the weekend. | 16:32 |
| tkamppeter | pitti, how can I test this before upload and where do I find the autopkgtest so that I can perhaps adapt it to newer features of CUPS? | 16:32 |
| flexiondotorg_ | I'll assess where we are on Monday to see how if using Debian is an option. | 16:32 |
| cyphermox | flexiondotorg_: tweak is just a utility no? | 16:32 |
| flexiondotorg_ | cyphermox, It is but it is actually the big ticket item for Ubuntu MATE 15.04 😃 | 16:33 |
| flexiondotorg_ | It make Compiz and the Interface switcher come to life. | 16:33 |
| flexiondotorg_ | For mere mortals. | 16:33 |
| cyphermox | tbh as long as the metapackages are installable, you should be good to build images | 16:33 |
| cyphermox | ah | 16:33 |
| cyphermox | well, we'll nsee | 16:33 |
| flexiondotorg_ | cyphermox, It would be an option to drop it and any other packages that don't make the cut and host them in a PPA with user post-install instructions. | 16:35 |
| flexiondotorg_ | But, I'd really like to land MATE Tweak. | 16:35 |
| cyphermox | no worries | 16:35 |
| cyphermox | if you really need to, it can land in ubuntu first, but it would be best if not | 16:35 |
| flexiondotorg_ | Must dash. Thanks cyphermox stokachu dholbach and others. | 16:35 |
| cyphermox | ttyl | 16:36 |
| dholbach | bye flexiondotorg_ | 16:36 |
| bdmurray | stokachu: if you mean releasing the package it needs to wait 7 days and its Friday | 16:38 |
| stokachu | bdmurray: ok | 16:40 |
| pitti | tkamppeter: it's in the source package in debian/tests; see http://packaging.ubuntu.com/html/auto-pkg-test.html how to run it | 16:44 |
| pitti | tkamppeter: since that seems to be the first time that you see it, I take it OdyX wrote that? :-) | 16:45 |
| pitti | rbasak: 1 GiB by default; we have some overrides for packages which need more | 16:46 |
| pitti | stgraber: oh, wow -- I've been trying for a year or so to reproduce that locally | 16:46 |
| rbasak | pitti: do you happen to have an override for mysql-5.6? :) | 16:46 |
| stgraber | pitti: well, I had it happen 3 times in a row :) | 16:47 |
| pitti | we occasionally hit that on our infrastructure (I suppose some qemu 9p weirdness), but I never got it here | 16:47 |
| pitti | stgraber: heh; I guess I want ssh access to your machine then :) | 16:47 |
| rbasak | pitti, stgraber: oooh, I hit that about five minutes ago. | 16:47 |
| pitti | rbasak: no, neither for 5.5 | 16:47 |
| rbasak | I just tried again. | 16:47 |
| rbasak | pitti: thanks | 16:47 |
| pitti | rbasak: are you on trusty as well? | 16:49 |
| pitti | perhaps trusty's qemu makes that more likely for some reason | 16:49 |
| rbasak | pitti: yes, with a custom backported autopkgtest | 16:49 |
| rbasak | Of some random version that fixed an issue I had months ago. | 16:50 |
| rbasak | I'd never actually seen this one before. | 16:50 |
| rbasak | Just coincidence that you mentioned here just now I guess. | 16:50 |
| * stgraber runs adt-run in a while loop... | 16:54 | |
| stgraber | pitti: well, here it just won't ever succeed... 10 times in a row now | 16:55 |
| stgraber | pitti: "adt-run -s --apt-pocket=proposed -U lxc --- qemu adt-vivid-amd64-cloud.img" <- that's valid right? | 16:55 |
| pitti | stgraber: eww; I guess I do want ssh on your machine next week.. | 16:55 |
| pitti | stgraber: correct | 16:56 |
| pitti | stgraber: CI machines are pretty clogged right now, but I confirm that at least one machine now can't access the proxy any more | 16:56 |
| flexiondotorg_ | stokachu, OK, Yuyo theme is fixed 😃 | 16:57 |
| tkamppeter | pitti, yes OdyX must have done it. | 16:59 |
| smoser | stgraber, since you're acdtive... could you archive-admin move python3-httpretty to main for me ? depencency for cloud-init (python-httpretty is already in main). | 17:00 |
| smoser | https://launchpad.net/ubuntu/+source/cloud-init/0.7.7~bzr1055-0ubuntu1/+build/6969483 | 17:00 |
| pitti | stgraber: RT sent to fix albali's proxy access (but the recent runs happened on other machines, so that still doesn't explain the whole story) | 17:07 |
| === ChickenCutlass is now known as ChickenCutlass_a | ||
| stgraber | smoser: done | 17:07 |
| === ChickenCutlass_a is now known as ChickenCutlass | ||
| === ChickenCutlass is now known as ChickenCutlass_a | ||
| stgraber | pitti: can you try connectivity to https://cloud-images.ubuntu.com from the CI lab? | 17:08 |
| pitti | stgraber: running | 17:08 |
| stgraber | pitti: since it's https, it may be bypassing the proxy or confusing it somehow | 17:08 |
| pitti | I'm running https_proxy=http://squid.internal:3128 wget -O- https://cloud-images.ubuntu.com | 17:08 |
| pitti | works on alderamin and aldebaran (where the last run happened) | 17:08 |
| pitti | albali is known broken, and wazn doesn't respond at all (it's overloaded) | 17:09 |
| * pitti restarts the LXC test, just to be sure | 17:09 | |
| stgraber | ok | 17:09 |
| flexiondotorg_ | stokachu, cyphermox https://bugs.launchpad.net/ubuntu/+bug/1421734 | 17:10 |
| ubottu | Launchpad bug 1421734 in Ubuntu " [needs-packaging] yuyo-gtk-theme" [Undecided,New] | 17:10 |
| pitti | wazn works, too | 17:10 |
| stgraber | pitti: so both lxc-test-apparmor-mount and lxc-autostart use the proxy to access GPG and images.linuxcontainers.org, lxc-test-ubuntu does a regular debootstrap using archive.u.c and creates a container using the cloud image, one of those two must be failing | 17:12 |
| stgraber | ok, 20 times and still stuck on the same adt issue, time to boot a vivid box... | 17:12 |
| stgraber | pitti: if it's reasonably easy for you, can you start a VM on one of those hosts, then install lxc-tests from vivid-proposed, do the two exports (http_proxy and https_proxy) and then "sudo sh -x /usr/bin/lxc-test-ubuntu" and see where it gets stuck | 17:15 |
| stokachu | flexiondotorg_: thanks, i commented on the ubuntu-mate-meta one | 17:17 |
| pitti | stokachu: it is; but I need to run now, I'm afraid; I'll do that Monday morning | 17:19 |
| pitti | bah! | 17:19 |
| stgraber | :) | 17:19 |
| pitti | stgraber: it is; but I need to run now, I'm afraid; I'll do that Monday morning | 17:19 |
| stgraber | pitti: ok. I'm getting ready to run adt on a vivid machine here so I can make sure that they pass fine | 17:20 |
| stgraber | (lxc-test-ubuntu is run for every commit upstream so I see no reason why it wouldn't work...) | 17:20 |
| pitti | cheers | 17:21 |
| flexiondotorg_ | stokachu, Regarding germinate I copied what the Ubuntu and the other flavours do. | 17:25 |
| stokachu | cyphermox: ^ im not familiar on the use of germinate in debian packaging | 17:26 |
| cyphermox | for ubuntu-mate-meta? | 17:27 |
| stokachu | yea | 17:27 |
| cyphermox | the update script writes to the relevant files for you based on what is on the seeds | 17:27 |
| stokachu | ah ok | 17:28 |
| cyphermox | these files get used to generate the list of Depends for the packages | 17:28 |
| cyphermox | (and Recommends) | 17:28 |
| stokachu | gotcha | 17:28 |
| stokachu | flexiondotorg_: with that settled do you want me to update the other recommendations i made? | 17:28 |
| stokachu | i can make the changes myself if you want | 17:29 |
| flexiondotorg_ | stokachu, If you could that would be great. | 17:29 |
| cyphermox | stokachu: as I recall you'll see there is one of the files which maps packages to the "seed" files, that gets read in to replace $germinate:Depends | 17:29 |
| stokachu | flexiondotorg_: yep will do | 17:34 |
| flexiondotorg_ | stokachu, Cheers. | 17:35 |
| smoser | stgraber, thank you | 17:37 |
| stokachu | flexiondotorg_: done | 17:39 |
| flexiondotorg_ | stokachu, Oh oh, what is? 😃 | 17:40 |
| flexiondotorg_ | stokachu, meta package. Nice! | 17:41 |
| flexiondotorg_ | cyphermox, Here comes the breakage right? 😉 | 17:42 |
| stokachu | flexiondotorg_: all of them | 17:43 |
| flexiondotorg_ | stokachu, Super cool. | 17:43 |
| stokachu | flexiondotorg_: thanks man, lemme know if you need anything else | 17:44 |
| flexiondotorg_ | stokachu, That is absolutely brilliant. | 17:45 |
| flexiondotorg_ | Thank you so much! | 17:45 |
| stokachu | np :) | 17:45 |
| flexiondotorg_ | stokachu, You've totally made my weekend. | 17:45 |
| stokachu | flexiondotorg_: haha, /fingers crossed i didn't break anything | 17:45 |
| stokachu | flexiondotorg_: glad to help, thanks for your contributions to Ubuntu :) | 17:46 |
| flexiondotorg_ | stokachu, My pleasure. | 17:47 |
| popey | flexiondotorg_: killed any kittens with your changes yet? :) | 17:48 |
| flexiondotorg_ | popey, Not yet. cyphermox Will builds "happen" now? popey wants to see some carnage. | 17:49 |
| popey | hey dude it's friday evening, friday is best day for carnage | 17:50 |
| flexiondotorg_ | Right, I'm off then. Everyone have fun with that. I look forward to error logs or something 😉 | 17:50 |
| flexiondotorg_ | I'll go back to the Debian guys and try to get the last packages uploaded. | 17:51 |
| cyphermox | popey: not quite yet | 17:51 |
| flexiondotorg_ | cyphermox, What is next? | 17:51 |
| cyphermox | but I'll get that shepherded to completion | 17:51 |
| flexiondotorg_ | cyphermox, When it explodes please point me to error logs so I can help clean up the mess. | 17:51 |
| flexiondotorg_ | The meta package is requesting caja-actions, mate-tweak and mate-menu that I am trying to get upload via Debian. | 17:52 |
| flexiondotorg_ | So I imagine it will fail. | 17:52 |
| cyphermox | once all the packages have been uploaded and reviewed by the archive team, a ubuntu-cdimage merge to be reviewed by the gods among us, and then kicking off an image. | 17:52 |
| cyphermox | yeah, things will fail to install and/or build | 17:52 |
| cyphermox | before you get a live cd these packages will need to exist or be removed from the seed. | 17:53 |
| cyphermox | I'm hungry | 17:55 |
| flexiondotorg_ | cyphermox, Enjoy your soup 😉 | 17:55 |
| flexiondotorg_ | Bye for now. And thanks to everyone who helped me today. You've been great. | 17:56 |
| stgraber | pitti: ok, I think I've found the problem with lxc-test-ubuntu | 18:13 |
| stgraber | pitti: it's got to do with the ubuntu-cloud template failing to recognize vivid | 18:13 |
| stgraber | I'll upload an updated cloud-image-utils fixing that | 18:17 |
| stgraber | hallyn: FYI ^ | 18:18 |
| hallyn | stgraber: d'oh | 18:19 |
| hallyn | then why wouldn't it fail when i ran it locally? | 18:19 |
| stgraber | did you run it on vivid? | 18:19 |
| hallyn | yup | 18:20 |
| stgraber | hmm, then I don't know. Here "lxc-create -t ubuntu-cloud -n test -- -r vivid" was failing immediately with an error saying that vivid is invalid | 18:23 |
| hallyn | and lxc-test-ubuntu does that? | 18:23 |
| stgraber | it doesn't pass -r but that defaults to current series | 18:24 |
| stgraber | root@autopkgtest:~# lxc-create -t ubuntu-cloud -n vivid | 18:24 |
| stgraber | ubuntu-cloudimg-query is /usr/bin/ubuntu-cloudimg-query | 18:24 |
| stgraber | wget is /usr/bin/wget | 18:24 |
| stgraber | confused by argument: vivid | 18:24 |
| stgraber | There is no download available for release=vivid, stream=daily, arch=amd64 | 18:24 |
| stgraber | root@autopkgtest:~# | 18:25 |
| stgraber | that's what lxc-test-ubuntu does ^ | 18:25 |
| hallyn | maybe mine was automatically setup for a differnet stream than daily? | 18:26 |
| === wedgwood is now known as Guest68861 | ||
| === JanC_ is now known as JanC | ||
| hallyn | slangasek: stgraber: infinity: so cgmanager 0.36-1 regresses some systemd-shim users bc it can't reach /var/run/nscd/. trying to think of the cleanest solution | 19:01 |
| hallyn | no wait, that cnsa't be it | 19:01 |
| hallyn | but it does want /usr available to read libnss | 19:04 |
| === roadmr is now known as roadmr_afk | ||
| hallyn | ok think i've got it | 19:26 |
| hallyn | i assume pitti is eow so i'll need to bug someone else to sponsor the fix :) | 19:27 |
| hallyn | xnox: hm, having a libnih issue - cgmanager --daemon is not causing /run/cgmanager.pid to be created | 19:37 |
| === roadmr_afk is now known as roadmr | ||
| hallyn | nih/main.c leads me to believe that should happen | 19:38 |
| teward | probably a stupid question, but is there a way to make a build environment for Debian packages against the Debian repository as well as Ubuntu? For when I set up my own repository, and I build nginx packages, I would like to have packages available for my Debain server on my own net as well in my own repository. | 19:48 |
| xnox | hallyn: let me check. | 19:48 |
| xnox | hallyn: where is the source code of cgmanager? | 19:48 |
| xnox | https://github.com/lxc/cgmanager? | 19:49 |
| hallyn | xnox: yeah | 19:49 |
| sarnold | teward: ought to be doable via sbuild / schroot .. | 19:51 |
| teward | sarnold: so, what, just follow the sbuild guide, but choose a schroot of `unstable`, `testing`, etc.? | 19:51 |
| sarnold | teward: I think so, I've got the schroots set up but never tried building in them :) hehe | 19:51 |
| teward | heheheh | 19:51 |
| teward | sarnold: no time like the present to test XD | 19:52 |
| xnox | hallyn: it writes out /var/run/argv[0].pid | 19:53 |
| xnox | hallyn: unless e.g. nih_main_set_pidfile() is called to set something else. | 19:53 |
| hallyn | xnox: ooooh. | 19:54 |
| hallyn | xnox: it's upset bc /var/run/cgmanager is already a directory, maybe? | 19:55 |
| hallyn | nah, that would be stupid | 19:55 |
| xnox | hallyn: .pid is appended as well. | 19:55 |
| hallyn | but anyway /var/run/cgmanager.pid is not being created, and that makes sysvinit stop of cgmanager very slow | 19:55 |
| hallyn | xnox: i'll get to the bottom of that, but in th emeantime, do you mind sponsoring http://mentors.debian.net/debian/pool/main/c/cgmanager/cgmanager_0.36-2.dsc ? | 19:56 |
| hallyn | i'd ask pitti but i thin khe's out fo rthe week | 19:56 |
| xnox | hallyn: i'm on trusty, and cgmanager is spawned with --sigstop, rather than --daemon. | 19:57 |
| xnox | hallyn: oh, debian. | 19:57 |
| hallyn | yeah, sysvinit | 19:58 |
| hallyn | it's fine in upstart | 19:59 |
| hallyn | well it's only a pain on restarting in sysvinit. like i say i'll dig into it, was just wondering whehter you know offhand since for some reason i have this feeling you deal with libnih a bit | 20:00 |
| xnox | hallyn: so, on trusty it seems to be fine. | 20:00 |
| xnox | hallyn: as in .pid file is created. However debian's libnih might be different here. | 20:01 |
| hallyn | xnox: oh. yeah. didn't even think of that | 20:01 |
| hallyn | you run trusty with sysvinit? | 20:01 |
| xnox | hallyn: no, i run trusty with upstart. I've stopped cgmanager and did $ sudo cgmanager --daemon and then check /var/run/cgmanager.pid to be the one that is running. | 20:02 |
| hallyn | xnox: d'oh | 20:05 |
| hallyn | xnox: ti's not libnih's fault. the pidfile is being created, but in the new mntns. (i thought that happened earlier) | 20:06 |
| hallyn | i'll have to think about that one. | 20:06 |
| hallyn | so /run on host is a tmpfs, and /run/cgmanager.pid is being created on the rootfs /run | 20:06 |
| xnox | hallyn: huh? is /run mounted as tmpf _after_ cgmanager started?! | 20:07 |
| hallyn | no, | 20:07 |
| hallyn | cgmanager bind-mounts / (non-recursively) and pivots into that | 20:07 |
| hallyn | it needs to somehow shed all mounts it doesn' tneed. BC otherwise if /foo is mounted on the host, then cgmanager starts, then hos tumount s/foo, then /foo stays mounted in cgmanager's namespace | 20:08 |
| hallyn | (starting to feel dirty yet?) | 20:11 |
| hallyn | xnox: so in case it got lost in the noise, woul dyou mind sponsoring http://mentors.debian.net/debian/pool/main/c/cgmanager/cgmanager_0.36-2.dsc to sid? | 20:11 |
| xnox | hallyn: i'm not quite sure where you shedding the mountpoints, but clearly you want to keep /var/run mountpoint (or anyhow that points to, e.g. usually symlink to -> /run) | 20:13 |
| hallyn | why clearly? | 20:14 |
| hallyn | i'm shedding the mountpoints by doing a non-recursive bind-mount of / onto NEWROOT, then pivot_root'ing to NEWROOT | 20:15 |
| hallyn | I have a /run/cgmanager for my own private (cgroup_ mounts, but I don't need the /run tmpfs, really. Now that one would be safe, so yeah I could bind-mount that too, and that would probably fix this | 20:15 |
| hallyn | really this *should* be a common idiom for any daemons that want to run in MS_PRIVATE | 20:16 |
| sarnold | one almost suspects there should be a kernel-provided mechanism to bring only selected mounts into the new namespace.. | 20:18 |
| hallyn | lemme see if bind-mounting /run fixes the issue | 20:18 |
| hallyn | sarnold: agreed - with the caveat that i can't think of a good way to choose 'selected' mounts | 20:18 |
| sarnold | hallyn: no, me neither :) | 20:18 |
| hallyn | heck, | 20:18 |
| === pgraner is now known as pgraner-afk | ||
| hallyn | maybe this is something worth discussing at a conference. use cgmanager as an example, | 20:18 |
| hallyn | ask whether others feel this needs generalization | 20:19 |
| hallyn | xnox: oh, there's a bug in the .dsc i was pointing you at anyway | 20:25 |
| hallyn | (minor one; extra arg sent to printf) | 20:26 |
| hallyn | ok ,so bind-mounting just /run (no sub-mounts) fixes the sysvinit issue. i'll push that fix to git and then push a new .dsc | 20:27 |
| cjwatson | teward: sbuild is *from* Debian, it works fine *for* Debian :-) | 20:29 |
| hallyn | sigh, ok, that's pushed to mentors. now to take a walk and clear my head | 20:33 |
| Unit193 | So can I ignore the warning at the top to remove the duplicate 'contents' title on https://wiki.ubuntu.com/UbuntuDevelopment/DeveloperApplicationTemplate ? | 20:48 |
| cjwatson | I would think so | 20:50 |
| cjwatson | the point is to discourage people from editing their own application text into it | 20:51 |
| Unit193 | Yes, figured that. But since it's somewhat of an important page figured I'd ask. Thanks. | 20:52 |
| === kickinz1 is now known as kickinz1|afk | ||
| === salem_ is now known as _salem | ||
| hallyn | slangasek: hey, would you mind sponsoring http://mentors.debian.net/debian/pool/main/c/cgmanager/cgmanager_0.36-2.dsc into sid? | 21:44 |
| stokachu | @pilot out | 22:11 |
| === udevbot changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support or app devel) | build failures -> http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of lucid -> utopic | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: | ||
| slangasek | hallyn: so if we're bind mounting the rootfs, what's the purpose of having a separate namespace here? | 23:28 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!