[00:06] Aha! Excellent I did not know about that [00:06] And many other people didn't [00:08] "A lot of people don't understand Blueprints or consider it a misfeature." [00:08] I think that *everyone* considers it a misfeature, especially us. [00:09] wgrant: Yeah, I didn't say that on purpose. [00:09] wgrant: It sounds so overly negative. [00:09] The only good thing about is it that I used it once and it gave me a virtually unbeatable contribution score almost instantly [00:09] lifeless: what worries me is that thing is from 2006 [00:10] lvh: It's actually on the roadmap now, though. [00:10] wgrant: I think that's great [00:11] But I heard "nobody actually develops lp" a few times [00:11] I know that's wrong but perception is hard to erase or ignore [00:11] Development has been glacial for a few years now. [00:11] But that is hopefully changing with the restructure and various other improvements. [00:12] I also ranted about your pricing model a while ago which gave me the impression Launchpad was being slated as an Ubuntu-package hosting sideshow [00:12] Which is unfortunate. [00:12] Because it'll be a cold day in hell before I mistake Github Issues for a bug tracker. [00:13] I believe the commercial offerings are going to be rethought this year, after our complete privacy rework is done in the next couple of months. [00:14] Yay. [00:14] It's not that $250/yr is expensive for what Launchpad offers. [00:14] It's just a pretty big gap. [00:14] Yes. [00:15] I believe the $250/year thing was introduced before GitHub and BitBucket had their current models... and we sort of never evolved past that. [00:15] right [00:15] Plus, it's cheaper to just dump everything for an entire company into one project even when it is conceptually many projects [00:15] we are going to redo our commercial offering [00:15] But the next feature on the list is making privacy not suck. After that we can vastly improve our commercial offerings. [00:15] its currently just treading water [00:15] Like I've ranted on before: number of users is a metric for how big something is which is also a far better herustic for you in terms of cost. [00:16] lifeless: I was going to buy lp [00:16] lifeless: Because if I don't, I'd use github or bitbucket and one has a terrible bugtracker and the other has a terrible pull-request-code-review story [00:16] So I'm going to have to buy or set up more services [00:16] And I'm really, really really lazy [00:17] (Also at my hourly wages it's unlikely I can get that done for $250/yr anyway) [00:17] I noticed that GitHub's new pull requests are suspiciously similar to LP merge proposals. [00:17] They are virtually a 1:1 clone! The problem is bug tracker integration [00:17] 23:47 < Jerub> lvh: you know what my primary annoyance with lp has always been? [00:17] 23:48 < Jerub> lvh: it's always so long between projects that i forget that the correct way to throw my branch under the project is to push to the non existant branch under the project name [00:17] 23:48 < Jerub> it just doesn't make any sense to me. github and bitbucket have it right with the 'clone' button. [00:18] 00:16 < Jerub> lvh: i want my fucking 'fork' button and then the 'bzr' invokation for pushing to my fork [00:18] It's funny because I consider that a feature [00:18] But apparently this is less than obvious! [00:18] 00:18 < arkanes> lvh: I like being able to auto-vivify branches by pushing, but the fork button is good UI [00:18] I think everyone except people used to Git consider it a future. [00:18] Because Git *has* to do it the other way, because they don't have stacked branches. [00:18] I realise this is a ridiculously small sample size [00:19] Right. [00:19] But lots of Git users have this same complaint. [00:19] Perhaps I should make an illustrated introduction to contributing to my thing. [00:19] "no you don't even need to fork yes really" [00:19] That takes me a long time to explain to git uses [00:19] users* [00:19] They just go like "yeah that can't actually work" for a while [00:20] And then I give up and just give them the sequence of bzr commands [00:20] I don't understand why people want to click around in the UI, then run a command to push... when they could just run a command to push :( [00:20] me neither but hey [00:20] I said this in my first post about lp [00:20] to be fair I've only had three people tell me this [00:20] "Your UI sucks -- I can't find the fork button" [00:20] I've had lots of people tell me that. [00:21] https://code.launchpad.net/launchpad should really have push instructions. [00:21] I was thinking an instructional in Chunky Bacon style. [00:21] And a Fork button that tells people to go away. [00:21] Because you can't explain things to people without cartoons. [00:22] lvh: TBH, I don't like the idea of fork that github pushes [00:24] A lot of feedback doesn't actually help [00:24] https://twitter.com/#!/olix0r/status/44542610735513600 [00:24] "It should be simpler" [00:24] Okay, I know, but how [00:25] wgrant: Having it on the code page is probably a big win [00:25] wgrant: Perhaps a sort of tutorial mode that you can turn off if you're already familiar with launchpad? [00:26] wgrant: As someone who already knows how to use launchpad You can browse the source code for the development focus branch or get a copy of the branch using the command: [00:26] bzr branch lp:bzr [00:26] * wgrant points at huwshimi. [00:26] That just means I have to scroll more. [00:26] wgrant: file a bug about some of these ;) [00:27] Perhaps the active reviews and the stats could be moved to the RHS: the important part are the barnches [00:27] lifeless: They could be summarised as "Our UI sucks, go through and actually look at every page" [00:27] I'm sure there already is a bug about a "fork" button [00:27] wgrant: thats not a discretely actionable thing though [00:27] hey, i've got a really simple LP question. [00:27] where are the docs? [00:27] lifeless: It is clear that no developer has ever looked at most of our pages. [00:27] I'm here: https://launchpad.net/ [00:27] Jerub: help.launchpad.net [00:27] Jerub: help.launchpad.net [00:27] i can't find them. [00:27] https://help.launchpad.net/ [00:28] What about people who think that a fork button is a flawed concept and shouldn't be added? [00:28] have a 'how to fork' popup with instructions [00:28] StevenK: We have to cater to people damaged by Git, unfortunately. [00:28] Jerub: On the front page, it says Get started [00:28] Which says: Learn more about Launchpad in the user guide or try it for yourself in our sandbox environment. [00:28] I followed the user guide link [00:28] "Learn more about Launchpad in the _user guide_ or try it for yourself in our sandbox environment." [00:28] It should probably have a main menu and say "Help" or something. [00:28] That link there. [00:28] Not entirely obvious. [00:28] Right. [00:29] ah right. the user guide [00:29] It should just be "Help" [00:29] I'm of the opinion that the fork button is there since git is braindead and can't do the equivalent of 'bzr push' [00:29] i was actually seaching for 'documentation' [00:29] i went to 'Find answers' first tbh [00:29] StevenK: That's right. [00:29] but quickly found that wasn't what i was after. [00:29] StevenK: That's what everyone agrees on I think [00:29] StevenK: But Git people want it to work on LP too, because the brokenness is more familiar :) [00:30] StevenK: the reason i push the 'fork' button is because it gives me a neat way of cutting+pasting the commands for creating my working copy locally that i can push from. [00:30] People don't like changing workflows, even if they are smoother. [00:30] wgrant: Did you see my suggestion about having code duplicated to github [00:30] I seriously doubt people will agree 'git is braindead', but okay [00:30] Jerub: But that's what it does right now [00:30] Jerub: I think we should give you those commands [00:30] You can browse the source code for the development focus branch or get a copy of the branch using the command: [00:30] bzr branch lp:bzr [00:31] lvh: And monitoring pull requests there? [00:31] Jerub: So if you could cut-n-paste the command directly it would be fine too? [00:31] Jerub: we do already but only once you're deep into the site [00:31] yep, just seeing that in the lp ui, i'm having a browse around. [00:31] wgrant: Yup: when pull requests occur, I create a branch from trunk and apply the diff and create a merge proposal [00:31] lifeless: 'code' tab is hardly deep. i like it there. it's also on every branch page. [00:32] lvh: we should look to make the bazaar locations easier to update, and document the append_path bits clearly [00:32] lvh: for all my LP work, I never specify a target, just "bzr push" [00:32] lvh: and my locations does the rest [00:32] perhaps a command in the launchpad bazaar plugin would be an idea [00:33] wgrant: merge proposals and pull requests are sufficiently similar that nobody serious could possibly object to being told to use merge proposals instead [00:33] the other thing that's confusing to me is what the deal with first pushing to my own user account's subdirectory before doing something to make it lp:projectname, i never really got that [00:33] bzr lp-set-project-locations ... [00:33] thumper: You definitely should because I would consider myself a bzr fan and I have no idea what you just said [00:33] thumper: bzr push just works for me because ... [00:33] Jerub: You don't actually have to do that any more. [00:33] Oh [00:33] well [00:33] Okay, so it just works for me but I have no idea why [00:33] Jerub: bzr push lp:someprojectwithoutabranch will now create it if it doesn't exist. [00:33] wgrant: that's interesting. i don't understand the docs or how to do it. [00:33] I suppose that's actually a compliment to bzr :-) [00:33] lvh: ok, here is some lines from my .bazaar/locations.conf file: [00:34] the docs https://help.launchpad.net/Code/QuickStart here say bzr push lp:~your-id/project-id/branch-name [00:34] on second thoughts, I'll use a pastebin [00:34] Jerub: This changed pretty recently. [00:34] oh goodie [00:34] Jerub: We should probably update the docs. [00:35] lvh: http://pastebin.ubuntu.com/576720/ [00:35] lvh: that way, inside my src directory, i create a new branch, and go "bzr push" to get it on Launchpad [00:36] lvh: I've just shown the wikkid section, my locations.conf file is large [00:36] larger than it needs to be, but I've not gone through and trimmed it recently [00:39] Cool [00:39] thanks :) === Pici` is now known as Pici [03:40] hello, i'm experiencing a strange FTBFS in a buildd (in a P3A) that i can't reproduce locally in my pbuilder chroot; on the buildd, i386 builds fine, but amd64 breaks; my local pbuilder is am64 and builds fine [03:40] here's a partial log from the buildd failure -- http://pastebin.ubuntu.com/576787/ [03:41] achiang: It's probably because amd64 builders don't build arch-indep packages. [03:41] achiang: So the indep targets of debian/rules aren't run. [03:42] wgrant: it FTBFS on armel too; same issue? [03:42] Try 'dpkg-buildpackage -B'; it should give the same failure. [03:42] achiang: Yes. [03:42] I'm not sure if pbuilder lets you run binary-arch instead of binary-indep, but sbuild does. [03:43] wgrant: hm, i don't have enough packaging-fu to take your statement and figure out how to fix the FTBFS [03:44] achiang: Do you understand what's going on? [03:44] We only build architecture-independent packages on i386. [03:44] So we run debian/rules binary-indep on i386, and debian/rules binary-arch on the others. [03:45] Your package builds some indep files unconditionally, but only copies them into the package in binary-indep. [03:45] This fails if binary-indep isn't run. [03:46] i see [03:48] wgrant: so, that means i should explicity copy the files in binary-arch? [03:49] achiang: Ideally you should not build them in binary-arch. [03:49] wgrant: hm. this is debian/rules -- http://pastebin.ubuntu.com/576792/ [03:52] somehow, this built for upstream maverick; i'm just trying to backport it for a project of mine, and it's failing [03:53] achiang: You'll have to look for a debhelper change between Lucid and Maverick, I guess. [04:06] wgrant: Who was it that was working on LP dkim integration (sorry, I've blanked)? I think I just found a significant bug in pydkim and I'd like a second opinon. [04:07] ScottK: poolie, but he's away this week. [04:07] OK. [04:07] Anyone else familiar with the topic? [04:08] handwavingly so [04:08] Vaguely. [04:12] I've been unable to get my messages to verify. [04:12] It looks like upstream assumes that a DKIM key record won't end with a ';' [04:13] If it does, it ends up with an empty string at the end after it's split the key record and fails the key [04:13] That sounds plausible. [04:14] Upstreams dkim key is at dig txt greg._domainkey.hewgill.com [04:14] Mine is at dig txt 2007-00_domainkey.controlledmail.com [04:14] It's not just me either: dig txt ip-dk-1k._domainkey.ironport.com [04:14] Missing . there, I presume. [04:15] Hmm. [04:15] I've got the ; on the end, he doesn't. [04:15] What does the spec say? [04:15] Mine don't have a trailing semicolon. [04:15] The spec isn't terribly clear on this. [04:15] steven@liquified:~% host -t txt 2007-00_domainkey.controlledmail.com [04:15] Host 2007-00_domainkey.controlledmail.com not found: 3(NXDOMAIN) [04:15] http://tools.ietf.org/html/rfc4871#section-3.6.1 [04:15] Uh? [04:15] Missing a . sorry. [04:16] dig txt 2007-00._domainkey.controlledmail.com [04:16] ScottK: I was wondering what was going on :-) [04:16] tag-list = tag-spec 0*( ";" tag-spec ) [ ";" ] [04:16] That seems pretty clear. [04:16] It's allowed. [04:17] (that's from 3.2) [04:17] Thanks. [04:18] I hadn't gotten that far. [04:18] dig txt s1024._domainkey.yahoo.com says to me this is not just a theoretical concern for LP. [04:18] Indeed. [04:18] Sounds like somebody needs to fix pydkim. [04:18] :( [04:19] Yeah. [04:19] My not probably not very pythonic solution (but works) is: [04:19] if a[len(a) - 1] == '': [04:19] a.pop(len(a) - 1) [04:20] Where a is the list of public key hunks split at the ; [04:20] Is this in verify()? [04:20] Yes [04:20] * wgrant dies quietly. [04:20] Line ~561 [04:21] If you could suggest an elegant fix, I can take care of getting an SRU into Lucid. [04:22] (I assume that's the relevant Ubuntu release for you) [04:22] That would be reat. [04:22] +g [04:22] Y'all are way more pythonic than I am, so I thought I ask for suggestions. [04:22] if a[-1] == '': [04:22] a.pop() [04:23] I'd do a.pop(-1) just for clarity if an unfamiliar reader is reading [04:23] Maybe. [04:24] given upstream seem to not know how string.split works [04:24] That works. [04:24] I would assume they count as an unfamiliar reader :) [04:24] lifeless: Yeah, fair point. [04:25] It's not clear that this was much more than a quick proof of concept that they published and forgot about. poolie and I both have written them at differnt times and AFAIK never heard back. [04:25] I'm --><-- that close to just forking it. [04:25] I think we probably should. [04:25] I'll write them with this bug and if I don't hear back, I'll do it. [04:26] Sounds like a good plan. [04:26] Thanks for handling this. [04:29] wgrant: thank you for the help earlier. i figured out how to fix it -- there was a utouch-geis-doc binary package that was specified as Architecture: all (as you predicted); i changed it to 'any' and now it is building correctly in the lucid buildd [04:30] achiang: That's not really the right fix, but OK.. [04:30] wgrant: better than my first attempt, which was to remove the --fail-missing arg from override_dh_install. :-/ [04:31] No, that would have been better. [04:31] Still not good, but better :) [04:31] really? [04:31] why better? [04:31] Now you're duplicating the -doc package across every arch. [04:32] true [04:40] wgrant and lifeless: Looking at it, you all probably are using your own copy of pydkim since we didn't SRU poolie's canonicalization patch. Please let me know if you actually need an SRU? [04:41] Fixed in Natty now. [04:41] ScottK: Ah, good point. [04:41] We are using an egg. [04:41] I'll grab the patch from natty and get that landed. [04:41] Thanks. [04:41] OK. [04:42] I doubt anyone else is actually using it, so unless someone complains, I'll not bother with the SRU. [04:42] Heh. [04:54] OK. Patch sent upstream. We'll see what happens. === jtv is now known as jtv-eat === jtv-eat is now known as jtv [06:44] hy it is posible to macke diffrent packages with diffrent configure with one soure? === Ursinha is now known as Ursinha-afk === vila_ is now known as vila === ferrousw1eel is now known as ferrouswheel === tumbleweed_ is now known as tumbleweed === dev-0 is now known as dev-zero [09:05] Morning === soren_ is now known as soren === gmb` is now known as gmb === kirkland` is now known as kirkland [11:39] hi guys, im trying to uplaod to my ppa, but lately idk why the orig.tar.gz gets stuck at 93622k out of 93623k, irrrespective of file size [11:40] X3lectric: There is a known bug in the Launchpad FTP server, which has recently been completely rewritten to avoid it, but not deployed yet. [11:40] Alternatively, consider uploading over sftp rather than ftp [11:40] Which avoids the issue by going through a different server [11:41] no difference [11:42] do I have to logout or reboot to make chages to .dput.cf? [11:44] X3lectric: no [11:45] maxb: also copying packages form one ppa to other if packages are same but different distros the second distro errors out [11:45] X3lectric: but its ~/.dput.cf I think, not ./.dput.cf [11:46] i know where .dput.cf should be it worked just fine, its just lately nothing seems to work ok i launchpad [11:47] copying packages is by far the worst [11:49] X3lectric: What's the error you get when copying? [11:49] anyone know why copying same packages of two dristros the second distro packages errors out saying source conflicts with existing or similar error [11:51] X3lectric: Copying two packages at once? [11:51] Or in separate requests? [11:51] at once or separate requests same error [11:51] Which packages? Which series? [11:53] wgrant: https://launchpad.net/~team-xbmc/+archive/unstable [11:53] all the karmic and lucid packages [11:53] from that ppa to another ppa [11:54] if request is at oonce errors out and nothing gets copied [11:54] if separate I only get a choice of either one karmic or lucid but not both [11:55] X3lectric: What is the exact text of the error message? [11:56] and after error ocurss you get to copy absolutily nothing because sources the are already in destination ppa despite nothing showing up === henninge_ is now known as henninge [11:58] The following sources cannot be copied: [11:58] * crystalhd 1:0.0~git20101029.6df10a0-1~ in lucid (binaries conflicting with the existing ones) [11:58] * rtmpdump 2.3-2~ in lucid (binaries conflicting with the existing ones) [11:58] option is just to copy binaries not rebuild copied sources [11:59] Which PPA are you copying into? [12:00] wgrant: destination https://launchpad.net/~x3lectric/+archive/xbmc-git-head [12:01] in fact doenat matter what ppa source or destination this is a general error when you try to cipy same packages from differnt distros [12:02] only can copy 1 distros packages once error occurs you just cannot copy second distro packages EVER [12:05] X3lectric: I can copy them separately after the error. [12:11] X3lectric: Try copying it from karmic to lucid. [12:11] They are the same package. [12:11] mmm [12:12] it wont copy [12:13] X3lectric: It will. [12:15] the karmic to lucid worked [12:15] but shouldnt have errors like this [12:15] It is due to the lpia binaries, and occurs when you have the same sources and binaries across several series. [12:15] also have another problem [12:16] oh? [12:16] ah [12:16] my dput I changed it form ftp to sftp and when I start uploading it says its using ftp [12:17] is this cached somewhere [12:17] No. [12:17] What is the command you are using to upload? [12:17] dput ppa:x3lectric/xbmc-git-head *source.changes [12:18] And what is in ~/.dput.cf? [12:18] * X3lectric is a networking Jedi but this ppa stuff is doing my force in [12:19] [x3lectric ppa] [12:19] fqdn = upload.launchpad.net [12:19] method = sftp [12:19] incoming = ~%(ppa)s/ubuntu [12:19] X3lectric: There's your problem. [12:19] What is that 'x3lectric ' doing there? [12:20] idk its what how to says [12:21] It's not. Try removing it. [12:21] ok [12:21] now I get No module named testtools [12:22] Do you have bzr installed? dput needs it for sftp. [12:22] im sure I do lemme double check [12:23] * X3lectric is waving fist at ppas and lauchpad in great fury [12:24] Package: bzr [12:24] State: installed [12:24] Where is the testtools import that is failing? [12:24] er idk what you mean [12:25] There's no traceback for the "No module named testtools" error? [12:26] i just run the dput cmd as above [12:26] then I get all sorts of signature checks ok [12:26] then I get this [12:26] Uploading to ppa (via sftp to upload.launchpad.net): [12:26] (your package copying error is even more of a corner case than I expected: it only occurs when you are copying one package to two series, where one of them has fewer architectures than the other, *and* the packages you're copying were copied from another PPA into the PPA you are copying from) [12:26] No module named testtools [12:26] E: Error connecting to remote host. [12:26] I don't see why, but try installing python-testtools. [12:27] kk [12:27] wgrant: some corner case :) [12:29] sounds like a big boobed up job on ppas corner case [12:29] ok now its getting worst [12:29] need at least testtools 0.9.2: /usr/lib/python2.6/dist-packages/testtools/__init__.pyc is () [12:30] That doesn't make much sense. Did you install a non-packaged version of bzr at some point? [12:30] no [12:30] its staright from apt-get [12:34] lemme downgrade [12:37] ok [12:37] says uloading but get no progress [12:38] Right, sftp has no progress info at the moment. [12:38] But it is reliable. [12:38] mmm idk [12:39] gotta love the way people moan about a free service [12:39] its not even showing what files its at [12:39] eh Im not moaning [12:39] just trying to get it to work past all the bugs [12:40] im not expereinced at ppa stuff [12:41] wgrant: right no dice [12:41] wgrant: ssh: connect to host upload.launchpad.net port 22: Connection timed out [12:41] X3lectric: upload.launchpad.net doesn't exist... did you mean ppa.launchpad.net? [12:42] mmm that the dput stuff [12:42] lemme change dput [12:44] omg [12:44] just gets better [12:45] Are you sure you want to continue connecting (yes/no)? yes [12:45] Warning: Permanently added 'ppa.launchpad.net,91.189.90.217' (RSA) to the list of known hosts. [12:45] Permission denied (publickey). [12:45] Unable to connect to SSH host ppa.launchpad.net; EOF during negotiation [12:46] * X3lectric doesnt knwo what to do anymore [12:52] wgrant: any clew now [12:52] googling it suggest deleting know_hosts [12:55] no dice [13:01] help?? :( === Ursinha-afk is now known as Ursinha [13:21] wgrant: thx for the help anyway Im gonna give up for now, juts getting way frustrated === henninge is now known as henninge-lunch === doko_ is now known as doko === jcsackett changed the topic of #launchpad to: https://launchpad.net/ | Help contacts: jcsackett | Launchpad is an open source project: https://dev.launchpad.net/ | This channel is logged: http://irclogs.ubuntu.com/ === henninge-lunch is now known as henninge [14:48] X3lectric: exactly what are you trying to do? [14:52] hyperair: upload packages to be built on ppa via sftp since ftp is buggy [14:52] X3lectric: i've found that dput's sftp support is buggy though [14:53] well packages get stuck in ftp [14:53] at 1 byte [14:53] http://blog.launchpad.net/cool-new-stuff/sftp-uploads [14:53] my packages get stuck in sftp after completion. [14:53] oh well [14:54] e.g out of 12310 12309 get uploaded and then its stuck [14:54] heh i've seen that issue before too [14:54] you could upload them manually via ftp. =p [14:54] no fix [14:54] we're switching to a new ftp server this week [14:54] my daily builds have been going up without the one-byte issue though [14:54] i wonder what's up with dput's ftp client [14:54] X3lectric: presumably you added your ssh key to your launchpad account? [14:54] and sftp requires a ^C to finish [14:55] yes ssh has been added for ages [14:55] then you're not connecting with the right credentials [14:55] X3lectric: could you do: sftp ppa.launchpad.net [14:55] see what that says [14:55] X3lectric: you need a username in the dput config [14:55] there is [14:55] you could also use -vvvv [14:56] X3lectric: the one you pasted earlier did not [14:57] [ppa] [14:57] fqdn = ppa.launchpad.net [14:57] method = sftp [14:57] incoming = ~%(ppa)s/ubuntu [14:57] login = x3lectric [14:57] X3lectric: could you please try connecting manually via sftp [14:58] or i'll just leave and you can deal with your problem on your own. [14:58] sftp -vvvv ppa.launchpad.net [14:58] actually sftp -vvvv x3lectric@ppa.launchapd.net [14:58] er launchpad [14:59] actually i need to be catching the train now [14:59] goodl uck [14:59]