[03:36] teward: It is common for my armhf builds to take a couple of hours when the usual build is 30 minutes. Emulation has its costs I guess. [06:20] im not getting updates from launchpad [06:20] i tried to dput something up there yesterday and today [06:20] but i cant tell whats going wrong [06:31] noting [06:31] nothing [06:46] karstensrage: Which PPA were you uploading to? [06:48] v [06:48] ppa:ufpidentity/identity4c [06:48] is that what you mean wgrant ? [06:50] karstensrage: Two problems: you uploaded a binary package rather than a source package (probably ran dpkg-buildpackage or debuild without the "-S" flag), and you didn't sign the .changes file. https://help.launchpad.net/Packaging/PPA/BuildingASourcePackage#Options_when_building may be helpful. [06:50] Since you didn't sign the upload, Launchpad has no idea who to email about the problem. [06:50] i did sign it [06:50] but probably not the same fingerprint as is up there [06:51] Hm, the error I see usually indicates that there was no signature, but possibly. [06:51] im using bzr builddeb [06:51] It must match a key that is on your Launchpad account, anyway. [06:51] ok [06:51] i guess i do want the source though, [06:51] You have no choice; Launchpad does not accept binary uploads. [06:52] Successfully uploaded identity4c_1.0.orig.tar.gz to ppa.launchpad.net for ppa [06:52] thats the source [06:52] wgrant, was the non-signed one from yesterday? [06:53] karstensrage: But there was a binary in that upload too. [06:53] You must build with -S. [06:53] Ah yes, the most recent upload was signed, but not with a registered key. [06:54] And earlier upload was entirely unsigned. [06:55] ok [06:59] wgrant, i have two .changes, do upload them both (assuming they are signed correctly) or just the source one? [07:00] karstensrage: Just the source one. Launchpad only accepts source uploads; it always builds the binaries itself. [07:00] ok [07:01] ill try again tomorrow [07:01] thanks for your help === coreycb` is now known as coreycb [08:38] git push lp:~ubuntu-server-dev/ubuntu/+source/ntp 1_4.2.6.p5+dfsg-3ubuntu9:refs/heads/ubuntu/devel [08:38] ! [remote rejected] 1_4.2.6.p5+dfsg-3ubuntu9 -> ubuntu/devel (failed to write) [08:38] rbasak: I am only a bot, please don't think I'm intelligent :) [08:38] Is that because I used "ubuntu/devel" with a '/' in it? [08:42] Though I managed to create "debian/sid" just fine. [08:44] rbasak: slashes are fine, and you created ubuntu/sid OK. is it reproducible? [08:44] wgrant: I did a little experimenting. [08:45] wgrant: 1_4.2.6.p5+dfsg-3ubuntu9 is a (local) tag so I had to specify refs/heads/ubuntu/devel instead of just ubuntu/devel for git. That fails. [08:45] wgrant: OTOH if I make a local branch and push that without disambiguating, then it works. [08:45] So workaround was: [08:45] git branch ubuntu/devel 1_4.2.6.p5+dfsg-3ubuntu9 [08:45] git push ... ubuntu/devel [08:46] (I made a mistake with ubuntu/sid instead of ubuntu/devel but I think that didn't matter in this case) [08:46] can you reproduce the original failure reliably? [08:46] Let me try. [08:48] wgrant: yes, reproducing. Now on git push lp:~racb/ubuntu/+source/ntp 1_4.2.6.p5+dfsg-3ubuntu9:refs/heads/ubuntu/devel [08:48] (I mistakenly created that as a tag on that repo first, I hope that didn't affect anything) [08:50] I think the failure case might be attempting to push :refs/heads// [09:06] rbasak: indeed. does that work on other git servers? [09:07] i don't see what we do that is special here, but it's possible [09:07] I would expect it to. I can check locally. [09:09] Interesting. It does not. [09:09] error: Trying to write non-commit object 8bd94ee19a36c6e1fae755a8a572150b442b0e54 to branch refs/heads/ubuntu/devel [09:09] remote: error: cannot update the ref 'refs/heads/ubuntu/devel'. [09:09] To /tmp/foo ! [remote rejected] 2.11-1ubuntu1 -> ubuntu/devel (failed to update ref) [09:10] ah, signed tag, i guess [09:11] Annotated tag, yes. [09:11] OK, that makes sense I guess. I'd like git to dtrt, but it certainly isn't a Launchpad issue. SSorry! [09:11] np, easy fix :) [09:12] Although could Launchpad give me a better error in this case? Or do you not want to for security reasons? [09:12] I've seen git itself give that sort of non-descriptive thing. [09:12] Have you tried pushing over SSH locally? [09:13] It's really bad at reporting errors over the remote protocol in some cases, and I think this might be one of them :( [09:13] We definitely don't deliberately suppress it. [09:13] But it's also possible that we detect the backend exiting before we forward any error, though I thought I fixed that. [09:15] Pushing to localhost:/tmp/foo does still give me the descriptive error [09:15] http://paste.ubuntu.com/14864577/ [09:24] rbasak: Hum, could you file a bug? I'll investigate. [09:24] rbasak: https://launchpad.net/turnip/+filebug [09:56] wgrant: https://bugs.launchpad.net/turnip/+bug/1541284 [09:56] Launchpad bug 1541284 in turnip "Push failures drop some of git's error explanations" [Undecided,New] [10:14] rbasak: Thanks. === axino` is now known as axino [10:45] Hey again! :) [10:46] Anyone have an idea how often LP verify every mirror? → Last probe [10:46] This mirror was last verified 22 hours ago. → Is it checked every 24h? every 28h? [10:46] Or it's random? [10:47] I'm not closing this channel, please you see this and you have the answer just send it :) If I lost connection I'll check the logs too :) [10:48] elacheche: Every 23 hours, usually, but it sometimes gets a bit behind. [10:48] Let me check. [10:49] OK.. Because a week ago on Monday it checked a specific mirror by 09am.. if it's 23h so it's logical to change the check time (that I can see in the html source of the page).. [10:50] I'm trying to generate some reports from those pages, so it'll be better to know if there is a fixed range of time :) [10:52] elacheche: There's no real way to predict. A mirror is requeued when its last check is 23 hours old, but the mirror at the top of the queue now was last checked 39 hours ago. So it's lagging more than usual. [10:53] :D OK :) So I should not use that as an actual status, but I should mention the value in my report [10:53] thx wgrant [10:54] If my project went alive I'll come here and share with you what I was working on.. Thank you wgrant and all the others who helped me here :) :D [10:54] elacheche: I'll be interested to see it! [10:55] BTW, do you know if there is a way to check the status of a mirror without using LP. [10:55] ? [10:56] elacheche: I don't quite understand what you mean. You could of course check for existence of the various indexes and packages yourself, but beyond that... [10:57] Emm.. I'll investigate that :) thank you wgrant [11:09] hey all, any idea why this MP requires login https://code.launchpad.net/~aacid/unity8/new_dash_navigation/+merge/271617 ? [11:13] Saviq: Fun. Looks like it supersedes https://code.launchpad.net/~aacid/unity8/new_dash_navigation/+merge/270634, which is proposed for merging into some private branch [11:13] Saviq: Please file a bug - not much to be done about the latter case directly, but we should probably redact invisible superseded MPs [11:16] cjwatson, will do, thanks [11:20] cjwatson, bug #1541331 [11:20] bug 1541331 in Launchpad itself "Public MPs superseding private ones require logging in" [Undecided,New] https://launchpad.net/bugs/1541331 [11:20] ta [13:39] https://launchpad.net/ubuntu/+source/llvm-toolchain-3.8/1:3.8~+rc1-1~exp1/+build/8933083 has run out of disk space it seems [13:41] xnox: 12:58 maybe remove the build dir after installation as a workaround; I'm not sure at what point increasing disk means we have to decrease number of builders [13:41] (#launchpad-ops internal) [13:41] doko asked the same question there [13:42] ah ok. [13:42] maybe i should be on that channel. [22:04] hi [22:04] with github, I can push stuff by typing my username + password [22:04] what's the trick with launchpad git repos? [22:05] trying to use the https option [22:05] then try username "sergio-br2" [22:05] and my launchpad password === [HeOS] is now known as HeOS [22:17] sergio-br2: We haven't yet implemented HTTPS access tokens for git, although I'm working on prerequisites for it at the moment. You can only push using SSH right now. [22:17] :/ [22:17] thanks [22:18] sergio-br2: But when it is done, it certainly won't involve using your Launchpad password directly; there'll be a separate token for it that you'll be able to generate. [22:18] nice === zequence_ is now known as zequence [23:22] git is getting my linux user as the "user" [23:22] :/ [23:22] this is why I like https [23:22] sergio-br2: username@ [23:23] ? [23:25] i give up [23:26] ... [23:26] sergio-br2: What is the problem? [23:26] it's using "sergio" as user [23:26] Giving up because you didn't put YOUR_LP_USERNAME@ in the URL seems silly. [23:26] which is my desktop user [23:27] What is the command you're using? [23:27] git config --global user.name [23:27] it changes the ~/.gitconfig right? [23:27] That's your full name for committer identification. [23:28] It's got nothing to do with authentication. [23:28] humm [23:29] [url "git+ssh://USER@git.launchpad.net/"] [23:29] insteadof = lp: [23:29] so just add sergio-br2@git.launchpad.net/ [23:29] ? [23:29] Yes. [23:30] not working yet [23:30] We can help more easily if you describe what isn't working and how it isn't working. [23:30] Launchpad user 'sergio' doesn't have a registered SSH key [23:31] it's pretty confuse, sorry [23:31] That's clearly not using the lp: alias you just created. [23:31] so are there other file that I have to change? [23:32] I don't know what you've done, nor what the command that's currently failing is. [23:32] So far you've pasted an irrelevant command, plus a config snippet that you're apparently not using. [23:32] Really don't have enough information. [23:32] I'm just trying to clone a repo... with git+ssh [23:32] . [23:32] git clone git+ssh://git.launchpad.net/~random-stuff/mgba/+git/mgba-debian mgba-debian-git [23:33] If you have that "insteadof = lp:" bit in your git config, you can just "git clone lp:~random-stuff/mgba/+git/mgba-debian mgba-debian-git" [23:33] Otherwise, use "git clone git+ssh://sergio-br2@git.launchpad.net/~random-stuff/mgba/+git/mgba-debian mgba-debian-git" [23:34] it's working [23:34] let's see to push it now [23:46] thanks wgrant, it's working [23:46] it's just a little trick [23:48] Great. [23:50] wgrant, does "daily build" option work? [23:50] sergio-br2: It should. [23:51] there's no button like in the bzr recipes [23:51] below the daily build [23:51] the dolphin emu recipe I tested at least, it was not working [23:51] I had to click in the Request build [23:52] sergio-br2: The button only appears if the recipe is stale (that is, the underlying code has changed so a new build makes sense) [23:52] yeah, I know [23:52] but I just commit, and it did not appear [23:52] Does LP show the new commit? [23:53] yup [23:54] Which recipe? [23:54] https://code.launchpad.net/~endrift/+recipe/mgba-daily [23:54] the main lp:mgba is broken, ugh