[00:02] james_w, hello [00:02] * jml afk [00:02] I assume that it doesn't verify the key matches, essentially allowing spoofing of a consumer, but given that it allows any consumer anyway, that's not too much of an issue [00:02] hey jml [00:03] it just can't change the way it responds based on consumer, which given the consumer secret often won't be secret isn't something you would want to do anyway [00:04] james_w: Why isn't the consumer secret secret? [00:05] well, that depends on the consumer [00:05] for a desktop app you would have to distribute the "secret" to all users [00:05] LP and another hosted service could decide to agree on a specific consumer name an secret to check [00:06] if they wanted different behaviour [00:06] Why would you distribute the secret to all users? The users create the secret when they authorize the application to use their Launchpad account. [00:06] no, the consumer secret [00:06] wgrant: http://paste.ubuntu.com/223111/ - no apparent failure to see [00:06] Note that the Consumer Key and Consumer Secret are not the same thing. [00:07] the app could generate this on first use, but then you can't check the consumer key against the consumer secret anyway [00:07] The Key is the nameish thing which the example gives as 'just testing' [00:07] It isn't secret. [00:07] yeah [00:08] there's no change to the security properties of the protocol as far as I can see, you just can't identify particular consumers reliably [00:09] james_w: You mean you can't identify a particular piece of software? You can identify the user reliably. [00:09] yeah, consumer as distinct from user [00:09] OAuth has 3 parties, with the user and consumer distinct [00:10] It's a bit strange, because in the docs the consumer is generally a web service. [00:10] Whereas in LP's case it generally isn't. [00:10] if the consumer is able to keep it's secret secret then the service provide can reliably identify them and modify its behaviour [00:10] Ampelbein: Sure you didn't omit something there? There's no addMember request. [00:11] yeah, OAuth is designed for the web, and they state that it doesn't work very well here for the desktop case [00:11] james_w: Keep its *key* secret? [00:11] The key is not a secret. [00:11] no, secret [00:11] secret secret [00:11] Consumer Key: [00:11] A value used by the Consumer to identify itself to the Service Provider. [00:11] Consumer Secret: [00:11] A secret used by the Consumer to establish ownership of the Consumer Key. [00:11] The key is the name identifying the consumer. [00:11] Ah. [00:11] everyone can know "flickr" is called flickr [00:11] but we can't know their "secret" [00:11] Right. [00:11] wgrant: the addmember request comes directly before the "send: 'GET /beta/~ubuntu-my...." [00:12] it's conflicting with the usual definition of "key" as "secret part of the knowledge" [00:12] Ampelbein: Oh, you removed it? That's the one I needed to see - just drop the oauth_* parameters from the URL. [00:12] james_w: Right, it's rather confusing. [00:13] james_w: Why would LP want to change its behaviour based on the consumer key? [00:13] it might not, but the OAuth spec states that it is something that you can do if you trust the consumer === beuno_ is now known as beuno [00:14] I originally asked before thinking it all through, so I thought that having conflicting consumer secrets for a particular key would have implications beyond what I now think it does [00:15] wgrant: that is all: http://paste.ubuntu.com/223114/ [00:15] i just removed the oauth_ parts [00:15] interesting diversion though, thanks [00:16] Ampelbein: There was no line with ws.op=addMember? [00:17] no, none. for the record, myfenris has the issue, i've been msging with him for quite some time to get the job done. [00:18] thanks to Ampelbein [00:18] helping me up for hours.. [00:19] Ampelbein: Ah, now I see why I was confused. [00:20] Ampelbein: Now wadllib checks the parameters client-side. [00:20] As of a week ago. [00:20] But my wadllib checkout is from two weeks ago, so when I grepped for that error it only appeared in the server-side code. [00:21] http://paste.ubuntu.com/223119/ <-- all error [00:21] myfenris: i told you not to do that! there is private information there. [00:21] Ampelbein, my bad.. :( [00:22] It's not too dangerous - those signatures are only good for issuing those particular requests. [00:22] Which are all read-only. [00:22] But still not a good idea. [00:22] sorry for that .. [00:29] wgrant: the complete script I wanted myfenris to use is at http://paste.ubuntu.com/223124/ - it's a simple script and I don't really know why it's failing. [00:43] Ampelbein: Well, I can't get a recent version of launchpadlib to work at all. I'd suggest filing a bug against launchpadlib. [00:43] It's certainly not your fault. [00:44] And it might be some version incompatibility. [00:44] wgrant: ok, will do that. thanks for help. [01:15] beuno: Re. bug #394542: I see project group icons have returned, but not distribution ones. [01:15] Launchpad bug 394542 in launchpad-registry "Project group and distribution links lack custom icons" [High,Fix released] https://launchpad.net/bugs/394542 [01:15] Or is the revision missing from edge? [01:15] wgrant, hrm [01:15] * beuno looks [01:16] (https://edge.launchpad.net/ lacks the Ubuntu icon) [01:16] damn [01:16] you're right [01:18] wgrant, thanks [01:21] beuno: You can't slip this in with that other trivial sprite fix? [01:21] wgrant, I can try, yes [01:39] hey folks [01:39] I am trying to use LP OpenID stuff to login into one of my apps [01:39] and I get only my username from it [01:40] am I doing something wrong or can it indeed send only username? [01:40] pygi: Consumers have to be whitelisted to get other details. [01:40] wgrant, how does one do that? [01:40] is it a global LP stuff, or per-account? [01:40] pygi: Not sure. I'd try asking at https://answers.launchpad.net/launchpad/+addquestion [01:41] wgrant, will do. Thank your sir! :) [01:41] It's global. [01:42] ah [01:42] ok, so I don't have to ask now :P [01:43] Why not? [01:44] wgrant, well, you just said its global :) [01:44] and since the service is not ready yet... :) [01:44] Ah. [01:47] wgrant, thanks for the help :) [02:14] hi, any known issues with loggerhead at the moment? i'm getting "Internal Server Error" when trying to view files [02:14] for example: http://bazaar.launchpad.net/~washort/ecru/trunk/annotate/head%3A/pysrc/ecru/nodes.py [02:16] spm, ^^ [02:16] gah. looking. [02:17] * dash tries to not feel too smug about someone _else's_ server breaking [02:17] dash: cool. nothing wrong with the server. /me suspects something else. mwhudson? ^^ << [02:17] dash: heh [02:18] * jml tries a thing. [02:19] it doesn't appear to be a problem with the branch itself. [02:19] cool. so if you start here: http://bazaar.launchpad.net/~washort/ecru/trunk/files/head%3A/pysrc/ecru/ and select node.py - internal error. sweet. [02:19] nodes.py [02:19] yeah this was working earlier today [02:20] ~6hours ago [02:22] hrm. so codebrowse *was* restarted about 4 hours ago.... codebrowse update. hrm. mwhudson? [02:34] dash: it's a bzr bug [02:34] erm, maybe [02:34] Excitement [02:34] * mwhudson tries some tings [02:35] oh hm, maybe not [02:36] !! [02:37] dash: it's this bug http://bazaar.launchpad.net/~washort/ecru/trunk/annotate/head%3A/pysrc/ecru/nodes.py [02:37] no [02:37] obviously [02:37] https://bugs.edge.launchpad.net/loggerhead/+bug/382765 [02:37] Ubuntu bug 382765 in loggerhead "history.py uses deprecated (and deleted) ProgressBarStack" [Critical,Triaged] [02:37] ^ this one [02:38] but i've never been able to reproduce it before [02:38] well i get that when I run loggerhead _here_ [02:38] so I did vaguely wonder how you guys were getting loggerhead to work :) [02:38] so it's your fault! [02:39] sorry, didn't mean to collapse the waveform [02:39] once you've noticed the problem, quantum tunnelling means loggerhead breaks everywhere [02:39] :) [02:44] wow ./serve-branches http://bazaar.launchpad.net/~washort/ecru/trunk isn't very performant [02:44] (to no surprise at all( [02:45] i think i patched this once then made the mistake of installing a new loggerhead deb [02:49] well, now i can reproduce it, i'll fix the bug i guess [02:49] \o/ [02:49] i'm a little worried because the comment in the code about bzrlib's progress bar code not being thread safe looks to still be true [02:50] * mwhudson tries some things [02:53] grumf [03:13] do you have server issues on bazaar.launchpad.net? [03:14] compengi`: not that i know about, what are you seeing? [03:14] Internal Server Error [03:14] compengi`: at which url? [03:14] http://bazaar.launchpad.net/~compengi/%2Bjunk/Cirssi/annotate/head%3A/Cirssi.pl [03:15] compengi`: hmm [03:15] compengi`: yes, there is a problem here [03:15] compengi`: i'm fixing it now :) [03:15] okay :) === salgado is now known as salgado-afk [04:25] wgrant, still around? === fjlacoste is now known as flacoste === jml changed the topic of #launchpad to: https://launchpad.net/ | Launchpad is now open sourced: https://dev.launchpad.net/ | Help contact: - | Join https://launchpad.net/~launchpad-users | Channel logs: http://irclogs.ubuntu.com [05:20] \o/ [05:30] Oh hey, look at that. === flacoste is now known as flacoste_afk [06:16] hi [06:16] i was told that several parts of Launchpad will not be freed [06:16] i was told this is the automatic packager and some other part [06:16] meoblast001, it's all been opened [06:16] is this true? [06:16] yes [06:17] every last bit has been freed? [06:17] well, it's true that it's been opened [06:17] nothing has been left back [06:17] not even the history [06:17] history? [06:18] meoblast001: the source control repository (bzr, naturally) has been released, not just a tarball of the current code. [06:18] ok [06:18] what about that automatic packager [06:18] i heard it was "Canonical's secret sauce" and there was no plan to free it [06:18] If it's part of Launchpad, it's been released. [06:18] That includes soyuz. [06:19] Grab the code and see for yourself! :) [06:19] i want to make sure everything running on the official Canonical Launchpad server can be downloaded, modified, redistributed, etc [06:19] meoblast001, the decision to not open has been reverted [06:19] meoblast001: the only thing you can't download is the config files [06:19] so *everything* is open source now [06:19] :) [06:19] amazing [06:20] so i can remove the text at the bottom of this page then http://mysticgalaxies.com/amethyst/code.html [06:20] meoblast001: if not knowing the name of our database servers causes you a severe problem... [06:21] meoblast001, yes, you can remove the threatening text :) [06:21] it's not threatening :P [06:21] people were getting angry at me for using Launchpad [06:21] that specific program i write is written to make a free alternative to the flash format, ticked people off that they had to use nonfree software just to get to it [06:22] but i'll let them know [06:22] and put a note that Launchpad is being freed on that page (after i verify it myself online) [06:23] this is amazing news :D thank you everyone [06:23] feel free to blog/tweet/dent about it :) [06:23] :) i'm not angry at Canonical anymore [06:23] Shoot, I just yelled it. That's kinda like tweeting it, right? [06:24] rockstar, short-range tweet [06:25] who here works at Canonical? [06:25] * meoblast001 gets curious of these things [06:25] meoblast001, like half of us :) [06:25] beuno: you do? [06:25] meoblast001: me [06:26] i see [06:26] mwhudson, yes. And rockstar, and kfogel, and jml, and thumper, and sinzui, and.... [06:26] ok, i hope you guys don't hate me, i'm one of those guys who corrects people when they say "Linux" referring to a complete system [06:27] but thank you all on the decision [06:28] meoblast001: hey, we're glad too :-). You're welcome. [06:29] meoblast001: for the record: that was Mark Shuttleworth's call. [06:29] meoblast001: we're all pretty happy about it too :) [06:29] tell him i said thanks :) [06:30] took me a few minutes to remember what site i wanted to go to... brainstorm.ubuntu.com was it [07:02] yay on the LP release [07:02] but, I'm wondering why there's no deb package? [07:02] micahg, it's really not that type of application :) [07:04] * sharkbait imagines doing sudo apt-get install launchpad :) [07:04] why not? [07:04] it should be like any other? [07:04] If I want to install mantis or bugzilla I get pacakges [07:04] mediawiki [07:05] there's no pleasing some people! [07:05] * micahg is very happy :) [07:05] oh, there's a mediawiki package? [07:05] just asking why [07:05] I thought there was [07:05] that would have saved me lots of time :) [07:05] yep [07:06] micahg: anyway, i guess there's no real reason [07:06] lots easier than configuring apache :) [07:06] I'm backporting bzr so I can install for myself :) [07:06] but we don't deploy launchpad using a deb, so there's no packaging for it [07:06] should I file a bug? [07:06] ah [07:06] it would be a massive package [07:06] how big is it? [07:07] I just reread the blog post [07:07] I don't know, I'm not there yet [07:07] it seems like it's more for people to develop than to run... [07:07] but the dependencies are 281mb [07:07] micahg, ~100mb IIRC [07:07] micahg: yes, it is [07:07] that's not so bad [07:08] can I file a request for a deb package? [07:08] * micahg wants to install this at work :) [07:09] micahg: as mwhudson said, we don't deploy it as a deb, there's very little incentive for us to produce or maintain packages for it [07:09] micahg: OTOH, it's open source, so if you want to do that, you could always provide patches [07:10] well, wouldn't it be nice to offer it as a package on ubuntu-server [07:10] someone will eventually provide unofficial debs, just like with chromium :) [07:10] I guess I should file a needs packaging in ubuntu then :) [07:10] You really don't want to run 'your own Launchpad' except for development. You end up with an overly complicated system - you are better off using tools designed for single project or small groups of projects. [07:10] no, I do want my own launchpad [07:10] I need a private instance [07:11] stub: what tools compare to launchpad in functionality? [07:11] oh, what license is it under? [07:12] I personally want to install lp to contribute by fixing a bug I filed ~1 year ago :) [07:12] micahg: https://dev.launchpad.net/LaunchpadLicense [07:17] ok, bug 402073 [07:17] Launchpad bug 402073 in ubuntu "[needs-packaging] launchpad" [Wishlist,Triaged] https://launchpad.net/bugs/402073 [07:17] let ubuntu do it :) [07:18] no need for you guys to bother with it [07:18] micahg: you may not be aware, but launchpad can host private projects too - after all, it hosted launchpads development for a long time :) [07:18] yes, but for work, we can't stuff hosted [07:19] that's cool that it hosts its own development though === al-maisan_ is now known as al-maisan [07:46] jml, everyone, way to go! === jml changed the topic of #launchpad to: https://launchpad.net/ | Launchpad is now open sourced: https://dev.launchpad.net/ | Help contact: - | Join https://launchpad.net/~launchpad-users | Channel logs: http://irclogs.ubuntu.com | #launchpad-dev is the developer channel [07:47] \o/ [07:47] poolie, thanks :) [07:47] i turn my back for a bit and... :) [07:48] whohoo [07:53] noodles775: :-) === henninge is now known as henninge-brb [07:55] is LP postgres only? [07:56] you want to run it on oracle? :) [07:56] mysql :) [07:56] it's pretty much postgres only yes [07:57] is there a reason why postgres was chosen over mysql? [07:58] stub: ^^^ [07:59] (you might know some of the background there?) === henninge-brb is now known as henninge [08:02] Congratulations to all involved on the open sourcing of LaunchPad! I noticed that all 3 links under Tools on the front page of the wiki https://dev.launchpad.net are not readable by mere mortals (non-Canonical employees?), even when I am logged in to the LaunchPad wiki... is that intentional? Perhaps a comment that these are "for Canonical staff only" (or whatever the restriction is) would be helpful to newcomers? [08:02] * jml looks [08:02] micahg: PostgreSQL supported most of the features we thought we would need and MySQL did not. The other contender was Oracle, and for a while we made sure we would be able to switch to Oracle if necessary but PG has worked great. [08:02] This was about 2004 [08:02] jmarsden, I'll fix those up. I think at least one is now obsolete. [08:03] jml: Thanks :) [08:05] stub: how integrated is the DB in the code? [08:07] hmmm. [08:07] jmarsden, I don't have permission to fix the front page, it seems! [08:08] Highly. We make use of PostgreSQL specific features, such as some of the SQL language extensions, PL/pgSQL and Python stored procedures, triggers, functional indexes, automatic load balancing over the replicas is Slony-I specific ... [08:08] ah, ok, so bascially to change DB engines is a total rewrite [08:08] jml: OK... thanks for trying :) [08:09] transactional DDL, tsearch2 full text search, database permissions... [08:09] * micahg is a web developer [08:09] kfogel, spm, can you give me permissions to edit the dev.lp.net front page? [08:09] jmarsden, I haven't finished trying yet :) [08:09] So I see... [08:10] jmarsden: you don't have them? [08:10] whups [08:10] jmarsden: sorry [08:10] jmarsden: bad completion of "jml" :-) [08:10] jml: you don't have perms? [08:10] let me see [08:10] No problem... [08:10] kfogel, apparently not! [08:12] jml: asking over in canonical irc, this is an admin question [08:15] jmarsden, fixed. [08:15] jml: Thanks [08:17] http://bazaar.launchpad.net/~launchpad-pqm/launchpad/db-devel/annotate/head%3A/lib/canonical/buildd/README gives me Internal Server Error. [08:19] pkern: me too [08:19] spm: if you've not gone to sleep yet... [08:20] Where's the Soyuz code in there? [08:20] pkern: soyuz is mostly in lib/lp/soyuz [08:21] pkern: and lib/lp/otherthings [08:21] grep is your friend ;) [08:21] find/grep even [08:21] I did just use loggerhead. [08:22] al-maisan: Oh hi (: [08:22] hello pkern :) [08:25] mwhudson, good morning again. it looks like it's a nasty issue with the server... ? [08:32] compengi`: it affects annotating certain files, it's not that bad, but it's a bit ugly [08:32] compengi`: the fix will get rolled out automatically in ~12 hours, it would be nice to find a sysadmin to blat it into place earlier [08:35] mwhudson, is it caused by my files? [09:05] hey [09:05] congratulations for the release! [09:10] hey carlos! [09:16] compengi`: no [09:18] The bzr branching takes ages... [09:21] pkern: just released, huge branch.... [09:23] woo. Thanks all, and congratulations. [09:24] did you split out lazr from launchpad so it can be used to develop other web applications or there is still something left to reach that point? (I know some lazr libraries where already released) [09:25] but I'm more interested on the whole framework [09:25] carlos: yeah it was split out a while ago [09:26] hi [09:26] bigjools: and is there any documentation about how to put all that together? or is just a matter of getting Zope and start adding such libraries on top of it? [09:26] we need review https://translations.edge.launchpad.net/wordpress/2.8/+imports anyone can her accept that? [09:27] imported file is downloaded from launchpad self, also don't send a strings error to me :) [09:27] carlos: I've no idea, sorry. [09:28] bigjools: ok, thanks [09:31] How big is the repo approx.? [09:31] 150M approx [09:32] The Getting page probably should mention that. I'll edit it now. [09:32] Well, it looks like bzr slowness also contributes to the download time a lot. [09:33] Or the servers can't do more than 40kB/s but I doubt that. [09:35] bzr itself is much faster than 40kB/s [09:35] Congratulations for the release Launchpad people! Happy open-sourcing day ;) ! [09:36] Hah, MemoryError and a crash. [09:37] LarstiQ: we'rehaving folk pull from http, which will add some overhead [09:37] bzr on my server did this, OpenVZ for the blame. But it ate 900M on my laptop already. [09:37] LarstiQ: and there is some chance we're getting dos:P [09:37] lifeless: yes, and everyone at it at the same time [09:37] pkern: which version of bzr? [09:37] It's HTTP of course that's being a reason for slowness. But seriously, fetching shouldn't load "everything" into memory. [09:38] Is it possible to continue pulls? [09:38] pkern: it doesn't load everything into memory [09:38] LarstiQ: 1.17 on my laptop, 1.16 on the server. [09:38] lifeless: That's why I quoted it. [09:38] pkern 18650 31.0 55.0 933572 837628 pts/4 S+ 10:10 8:43 /usr/bin/python /usr/bin/bzr branch lp:launchpad [09:38] 55% of my memory is currently in bzr... [09:38] [#########| ] 112129KB 103KB/s | Fetching revisions:Inserting stream [09:38] pkern: resuming is possible, might need some shuffling around depending on how you started [09:39] we're definitely seeing interest :) [09:40] pkern: thats certainly higher than I'd have expected [09:40] Well it's late and huge though. It's probably a pain to set up. [09:41] pkern: I'd let it finish [09:41] pkern: you've pulled ~2/3rds of it now [09:41] I life non-linear progress bars... not. [09:42] pkern: are you at Debconf? [09:42] lifeless: last time I pulled a branch that size, bzr got to 900M resident [09:42] with 1.17rc-something === proppy1 is now known as proppy [09:43] LarstiQ: I am. [09:43] k [09:46] Are there any launchpad .debs? I can't seem to see them on dev.launchpad.net [09:47] pkern: its a bug [09:47] it's not packaged [09:47] both the slowness, and probably the speed [09:47] I'm filing it now [09:47] lifeless: 900M resident is a bug for sure :) [09:49] thrashing... [09:50] X stopped responding for a while. [09:53] bigjools: a) always file bus when that happens, please. [09:53] bigjools: can't fix what we don't know about [09:54] bigjools: b) was that over http or bzr [09:55] lifeless: a) yes, I told kfogel and thought he was handling it, but my bad. b) http [09:55] it was part of the pull testing we did before going open [09:55] I mean you must have noticed such issues earlier? ;-) [09:56] pkern: first I heard of it; and am a bzr dev [09:56] bigjools: well future ref - make sure a bzr dev knows, please:) [09:56] is the only way to have LP is to use bzr? there isn't a packaged LP? (.deb or .tar.gz or else) [09:56] MaWaLe: bzr only [09:57] thx lifeless : the purpose is to hold a classroom for our LoCo on it. And this way, we have to give them a little intro to the use of bzr :) [09:58] lifeless : is it planned to add it like an available package with Ubuntu? [09:59] MaWaLe: LP is complex to get up and running, even if someone did they'd need to rebrand it (see the announcement and wiki pages for details) [09:59] there are no plans to package it [09:59] is there a FAQ this question could be added to? It's the third time in a couple of hours it gets asked now. [10:00] not even a PPA? [10:00] yes, I'll sort it [10:00] MaWaLe: certainly we don't have any plans to package it - we don't deploy it via packages [10:00] thx all for your answers [10:01] bigjools: kfogel's text was pretty good, I thought [10:02] bigjools: in one of the announcement thingies [10:02] lifeless: I missed that then, I'll look again [10:02] it wasn't precisely on packaging [10:03] bigjools: elmo gave I think a reasonable concise answer in here [10:03] but rather about the icons etc aspect, which is connected [10:04] ok, I'll see if I can copy & paste that bit [10:04] we should probably add stub's bit about postgres dependency too [10:05] I suspect that'll become a FAQ too [10:05] elmo: oh right [10:08] stub: can you give me a paste of what you said, I can slot it into the wiki. [10:10] bigjools: http://paste.ubuntu.com/223374/ [10:10] ta [10:11] Already pulling for an hour, oh well. === jtv is now known as jtv-afk [10:20] pkern: lots of folk are downloading it it seems ;) [10:20] I've filed bug 402114 about the performance [10:20] Launchpad bug 402114 in bzr "too many http requests with 2a fetches" [Critical,Triaged] https://launchpad.net/bugs/402114 [10:22] lifeless: does that cover the huge memory usage or should I file another? [10:23] bigjools: please file another; I suspect they are linked though, so perhaps mention 402114 [10:23] ok [10:24] branching tip *should* be 'read 14 packs, churn CPU while checking they are compressed well. Done.' === noodles775 is now known as noodles775-afk [10:24] but its doing lots of little reads; this could be data specific [10:24] or a bug [10:25] and if its data specific we may be creating that ourselves (bug in the writer), or working badly with whats on disk (bug in the reader) [10:30] Probably you should tar up a current checkout and offer that for download and later pulling. [10:31] Memory usage doesn't grow anymore but... [10:31] [#########- ] 198662KB 0KB/s | Fetching revisions:Inserting stream [10:31]