[02:31] <trbs> how can i remove a branch from my launchpad account ?
[02:32] <lifeless> delete it in the web ui
[02:32] <trbs> i cannot find the link for it :(
[02:33] <trbs> aaah ic ... it's the little trash can icon
[02:33] <trbs> thanks lifeless ! :)
[05:24] <RenatoSilva> I want to put my useful scripts in the web. They're kind of mini-projects: not so big, often a single file, but useful...
[05:26] <RenatoSilva> I think it would be annoying to create one project for each one, so I was think about creating a single project called "<mylogin>"  or "<mylogin>-utils" or so.
[05:26] <RenatoSilva> How about it?
[05:26] <thumper> RenatoSilva: you can just use +junk for that
[05:26] <RenatoSilva> what's +junk?
[05:27] <thumper> stuff not related to a project :)
[05:27] <RenatoSilva> I don't get you yet, I'm still reading some docs
[05:27] <wgrant> RenatoSilva: Just use '+junk' in the branch URL where you'd normally put the project name.
[05:27] <thumper> eg https://code.edge.launchpad.net/~thumper/+junk/icerobots
[05:27] <wgrant> It'll create a personal branch unrelated to a project.
[05:28] <RenatoSilva> btw, what's the difference about a project and a branch?
[05:29] <RenatoSilva> thumper: so +junk is kind of meta-project to store code snippets and utilities?
[05:29] <spiv> RenatoSilva: projects can have many branches, and bugs, and translations, download files, etc.
[05:29] <thumper> RenatoSilva: kinda
[05:29] <thumper> RenatoSilva: there are no bugs or answers for +junk
[05:30] <spiv> RenatoSilva: more like "no project" than a meta-project.  It's named after http://www.samba.org/junkcode/ IIRC.
[05:30] <RenatoSilva> spiv: can't branches themselves have bugs associated, etc?
[05:30] <thumper> RenatoSilva: junk branches are just branches you want to put on LP but are not part of something larger
[05:30] <wgrant> RenatoSilva: Not in the same way as projects have bugs, no.
[05:30] <spiv> RenatoSilva: branches can be linked to bugs, but they can't *have* bugs, if you see what I mean.
[05:31] <RenatoSilva> junk looks a weird name
[05:31] <wgrant> (when a branch is linked to a bug, it normally means that the bug is fixed - not present - in the branch.
[05:31] <thumper> junk is a weird name
[05:32] <thumper> chosen on purpose :)
[05:32] <wgrant> thumper: To dissuade people from using it?
[05:32] <thumper> wgrant: to make people want to have real projects for stuff they really care about
[05:32] <thumper> if a branch is just a simple script, then sure, use +junk
[05:32] <wgrant> That's the reason 'junk' was chosen for a similar purpose in my project, well before anybody who knew of LP was on the scene.
[05:33] <RenatoSilva> spiv: link/have bugs? I need to read more in the wiki to understand this...
[05:33] <thumper> RenatoSilva: a bug can be linked to a branch, which can be seen from the bug and the branch
[05:33] <thumper> bugs have tasks, some of which can be on projects
[05:34] <thumper> a bug cannot have a task for a branch
[05:34] <thumper> other tasks are on distros, or source packages
[05:35] <wgrant> RenatoSilva: Have you been through the LP tour? That has a bit of discussion on how bugs work, IIRC.
[05:35] <wgrant> It is fairly confusing.
[05:36] <RenatoSilva> So let me see, branches are just actual code, and projects is a more general idea around this code.
[05:37] <wgrant> Right.
[05:37] <wgrant> A branch is the actual substance of the project.
[05:37] <wgrant> But a project can have hundreds of branches.
[05:37] <wgrant> And bugs.
[05:37] <wgrant> And blueprints.
[05:37] <wgrant> And downloads.
[05:37] <RenatoSilva> wgrant: I tried to use translations a long time ago to help Ubuntu. I found that part confusing.
[05:46] <RenatoSilva> wgrant: here in Brazil we have a saying: tamanho não é documento. I think in english it's like "Good things come in small packages.", i.e., the value of a software can't be measured by how many files it consists on.
[05:46] <RenatoSilva> some of my useful stuff are just single files thanks to guys like Matz which has invented stuff like Ruby
[05:48] <thumper> RenatoSilva: LP doesn't care how big your code is, if it is useful and open source, we're happy for you to have a project for free
[05:49] <thumper> in fact I'm not sure we even discriminate on useful...
[05:49] <RenatoSilva> +junk seems that the code is , hmm, junk.
[05:49] <thumper> some code is junk
[05:50] <wgrant> Junk has a couple of different meanings, unfortunately.
[05:50] <RenatoSilva> I think +junk is not a good place, it's not even a project
[05:50] <RenatoSilva> it's kind of sand box, it seems to me
[05:50] <RenatoSilva> Would it be a better idea to create a <-my-login-name> project?
[05:50] <spiv> RenatoSilva: See the description on http://www.samba.org/junkcode/; "The reason I call this 'junkcode' is that I have no plans to properly document, package or support this code. If you find it useful then that's great ..."
[05:51] <spiv> RenatoSilva: that's the sort of thing that Launchpad's +junk is good for.  Code you want to share, but not turn into fully-fledged projects for whatever reason.
[05:52] <wgrant> And if you do eventually want to turn it into a project, you can easily do so later.
[05:53] <lifeless> RenatoSilva: it is a per-user sandbox; calling it by the usersname would lead to urls like ~USERNAME/USERNAME, and make usernames have to be distinct from projects.
[05:54] <RenatoSilva> spiv: that's not the idea of little projects consisting in one file. Just imagine you created 5 nice and useful softwares, but "unfortunately" they're just too simple. They become simple thanks to your scripting language or anything else. However you do want to share the code, document, solve bugs, etc. _But_: in a easy way, the cost of doing this being proportional to the simplicity of the "project".
[05:57] <lifeless> well, I think +junk should be permitted bugs
[05:57] <RenatoSilva> wgrant: can you get what I mean? separate code by 'junk' and 'project', there is too much blank space between them. I think there's a problem with the word "project". Your software don't need to be a prooooject to be useful and worth to share, it doesn't mean that it is a 'junk' tough. Can you get what I mean...
[05:57] <lifeless> but thats a different storey
[05:57] <RenatoSilva> I think I'll ahve to create a project named after my login name
[05:57] <wgrant> RenatoSilva: Yeah, I see what you mean.
[05:58] <wgrant> In Ubuntu, we have lots of little scripts. They all end up in the ubuntu-dev-tools project, although they're mostly independent.
[05:58] <spiv> RenatoSilva: personally, I'd just make a project per thing, even if they are just single files then.
[05:58] <wgrant> But that's a bit different from your situation, since it makes sense to have ubuntu-dev-tools grouped together.
[05:58] <wgrant> whereas your grouping would be by author, which is probably not a good idea.
[05:58] <wgrant> So, I agree with spiv.
[06:01] <RenatoSilva> spiv: project sounds like a project, something you're planning to do and is doing, something big.
[06:01] <RenatoSilva> to me
[06:01] <spiv> RenatoSilva: then don't think of it as that, think of it as "free thing I need to have to have a Launchpad bug tracker for my software"
[06:01] <RenatoSilva> wgrant: what is dev-tools for? utilities for developing ubuntu?
[06:02] <spiv> Easy ;)
[06:02] <wgrant> RenatoSilva: They're useful utilities for Ubuntu developers, yes.
[06:07] <RenatoSilva> spiv: it's like having to choose between coffee and milk, when you just want coffe _with_ milk :)
[06:08] <spiv> RenatoSilva: would that be a latte or a cappuccino or a macchiato or ... ;)
[06:09] <RenatoSilva> wgrant: project in my case would mean just a meta-project, to get the infra-structure to support the utilities. Its name would work just like a bucket...
[06:09] <RenatoSilva> or a tag...
[06:11] <RenatoSilva> Github has some work on this idea... it's a versioned paste bin I can't recall the name
[06:12] <RenatoSilva> you paste code there, and it becomes kind of mini-project
[06:13] <RenatoSilva> Oh, it's here: http://gist.github.com/
[06:15] <wgrant> That is interesting.
[06:26] <RenatoSilva> when I open a bug, can I select the branch where it happens, if I know where? For example, imagine I write a patch to Ubuntu, apnd put it into a branch. Then someone download the branch to test, but find a problem. The bug report must be associated to a project, obviously the one associated with the branch. Can you get what I mean? How to report bugs and request support (answers) related to a specific branch?
[06:32] <lifeless> you can link bugs to branches
[06:33] <wgrant> But since Bug<->Branch link statuses have been removed, doesn't such a link imply the *fix* is in the branch?
[06:33] <RAOF> But that's really the other way 'round, isn't it?  IE: this branch _fixes_ this bug?
[06:34] <lifeless> RenatoSilva: I think you are thinking about the problem in a different way than launchpad presents it
[06:34] <RenatoSilva> doesn't the link feature stand that?
[06:34] <lifeless> lp treats the project as the central core that people consolidate around
[06:34] <wgrant> Bugs do have branch-specific task, but only if your branch is a series branch.
[06:35] <wgrant> I don't know what you're meant to do if you need to track bugs on an in-progress branch.
[06:35] <lifeless> and branches are short lived entities that new code is developed in
[06:35] <RenatoSilva> lifeless: likely. So what would be the right way?
[06:35] <wgrant> lifeless: In Launchpad, maybe.
[06:35] <lifeless> RenatoSilva: file a bug on the project, done
[06:35] <wgrant> lifeless: But most projects don't have an 800-line limit.
[06:35] <wgrant> And branches can go on for months.
[06:35] <lifeless> wgrant: as I said, how lp treats it.
[06:35] <RenatoSilva> lifeless: even if the problem is with a branch?
[06:36] <lifeless> wgrant: lets not confuse this discussion; we can come back to modelling terms in a sec
[06:36] <lifeless> RenatoSilva: if your branch is so long lived that its an entity of its own its really a fork isn't it?
[06:36] <lifeless> a fork, or a different flavour - what launchpad calls a 'series'
[06:37] <wgrant> A series isn't a fork.
[06:37] <wgrant> Forks generally have different ownership.
[06:37] <lifeless> wgrant: I know, thats why I said or
[06:38] <lifeless> wgrant: which in english is the xor operator.
[06:38] <RenatoSilva> lifeless: a temporary fork
[06:38] <wgrant> lifeless: I took the 'or' to be an 'in other words' - sorry.
[06:38] <lifeless> RenatoSilva: could you expand on that?
[06:39] <RenatoSilva> lifeless: it's a separate work you _want_ to merge into main code, but in the meantime your own branch has a life, bugs can happen, you should have to maintain, to document, to support while other people are testing your branch/fork
[06:40] <RenatoSilva> *you could have
[06:40] <lifeless> so lp doesn't directly model this at the moment. What I'd do is tag bugs of that nature with the name of the reason-for-that-branch
[06:40] <lifeless> remembering that many people could be working on the same featureset
[06:40] <RenatoSilva> branches are also projects themselves
[06:41] <lifeless> not in lp
[06:41] <RenatoSilva> then I come back to my question
[06:41] <lifeless> I appreciate how you're modelling it though. Its just not how lp approaches the problem.
[06:41] <RenatoSilva> to my point
[06:42] <lifeless> you're modelling branch implies project
[06:42] <RenatoSilva> you'll have to use the global infra-structure (the project) to report a bug on a specific branch
[06:42] <lifeless> but project is heavyweight
[06:42] <lifeless> it implies separate wikis developers mailing lists and all sortsof other things
[06:43] <lifeless> if you imagine that we didn't have DVCS systems and just collaborated with mailing lists and patches, you wouldn't routinely ask for a new mailing list to have a long lived development task for a project
[06:43] <lifeless> instead you'd say on the list that this issue is specific to the feature-FOO development
[06:44] <lifeless> anyway; in lp you have some choices. You could:
[06:44] <lifeless>  - create a series for the branch. File bugs on the series.
[06:44] <lifeless>  - use tags
[06:45] <lifeless>  - link the branch [though the LP machinery may do confusing things here as it assumes 'fixed in' is the relationship nowadays.
[06:45] <wgrant> Creating a series for a branch that is going to be merged into trunk is a very bad idea.
[06:45] <lifeless> I'd like to note though, that this is the first time I've seem someone ask for the ability to file bugs on a branch, over uhm, 5 years now.
[06:46] <RenatoSilva> I'm not used to series yet, but aren't they related to the project?
[06:46] <lifeless> they are
[06:46] <RenatoSilva> so project is a list of series, which are lists of branches?
[06:47] <wgrant> lifeless: I've seen quite a number of bugs in Launchpad itself about features that I didn't know existed yet, and are some time away from landing.
[06:48] <lifeless> RenatoSilva: a project is well, a project. Like 'squid'. Or 'grep'.
[06:49] <RenatoSilva> lifeless: sorry
[06:49] <lifeless> RenatoSilva: a series is a well known branch of the project. Like 'trunk'. Or '2.3'.
[06:49] <RenatoSilva> I'll read more about series later I think
[07:04] <RenatoSilva> A possible solution if you're about creating a patch and want it to "sound" like a project, would be maybe create a project???
[07:05] <RenatoSilva> Thsi way people can test your branch/fork and report bugs etc, because it is a full-featured project
[07:05] <RenatoSilva> when you work is finished, you delete the project (is it possible) and merge the branch into the main project
[07:06] <wgrant> Creating a project doesn't seem like a good solution for a temporary branch.
[07:06] <wgrant> As projects can't really be deleted.
[07:06] <RenatoSilva> :(
[07:11] <RenatoSilva> can you guys understand? Imagine you start a patch which is refused by main developers, but is yet useful for other people, so that they start to report bugs and try to collaborate and all tasks realted to projects. Then you'll tell people to use the main project page to report bugs and etc., but the main developers would not like this (supporting a refused branch). Then you decide to create a project, a fork. Then it becomes popular, and then main d
[07:12] <spiv> RenatoSilva: you got cut off at "and then main d"
[07:12] <RenatoSilva>  and then main developers now find it nice and want to merge the branch into main code. Then the fork project has no sense anymore...so it would be nice if you could delete the project, because it's a temporary project.
[07:13] <spiv> RenatoSilva: but if you want to track bugs etc. in code rejected by the main developers then at that point you're maintaining a fork, so you may as well create a project for it.
[07:13] <spiv> RenatoSilva: I don't think that scenario happens very often
[07:14] <RenatoSilva> ok
[07:14] <spiv> RenatoSilva: usually because the main devs don't reject work like that outright
[07:14]  * RenatoSilva actually can't remember why he's thinking about this...
[07:14] <spiv> i.e. if it's something they might merge later if more improvements are made to make it acceptable, then generally they'll say so.
[07:15] <RenatoSilva> spiv: maybe they rejected it initially because it was porr code or so...but it impoves in the future
[07:16] <spiv> RenatoSilva: sure.  So they'll say that.  That's not a good reason to make a fork.
[07:16] <spiv> Also, managing long-lived branches with large delta against the upstream branch is a hard problem in general.
[07:17] <spiv> You're continually trying to integrate divergent work, and when bugs occur you have to work extra hard to figure out if it's just your branch or if it's due to the upstream branch, etc.
[07:17] <RenatoSilva> Anyway, about my not-so-junk repo. Is it better to put my few/single-file utilities into one single branch, or create one branch for each file?
[07:18] <RenatoSilva> spiv: ok forget about it. I can't even recall the reason why I was thinking about this o.O
[07:18] <spiv> It's hard to say.  It really depends on how you want to interact with Launchpad, etc.  It's probably easiest to start with an all-in-one project for now while you get a feel for how managing projects and branches in LP works.
[07:19] <RenatoSilva> spiv: it would be nice if you could delete or close projects tough
[07:19] <spiv> Oh, you were asking about one branch vs. multiple branches.  That doesn't really matter a whole lot either way if it's just a handful of small things.
[07:19] <RenatoSilva> spiv: ok (about my repo)
[07:19] <spiv> Do whatever seems easiest, it's not a big deal to change your mind later.
[07:20] <RenatoSilva> wouldn't it be weird one branch for a single file?
[07:20] <spiv> Not really.
[07:20] <RenatoSilva> oO
[07:20] <spiv> No more so than one branch for several unrelated pieces of code, certainly ;)
[07:20] <lifeless> if there isn't a common theme between the files, one branch with unrelated stuff is more weird.
[07:21] <RenatoSilva> ok
[07:30] <RenatoSilva> hum, "you don't earn karma from +junk branches. "
[07:30] <RenatoSilva> that's a good definition for +junk branches :)
[07:59] <RenatoSilva> what's the diff between name and title when registering a project?
[08:07] <lifeless> name == url
[08:08] <lifeless> or url element actually
[08:11] <wgrant> lifeless: Not any more, on the new project page.
[08:11] <lifeless> wgrant: no?
[08:11] <wgrant> lifeless: That's the only place in Launchpad where it's different (it was changed with the new guided registration stuff a couple of months ago)
[08:11] <lifeless> wgrant: I made a new project on saturday.
[08:12] <wgrant> https://edge.launchpad.net/projects/+new
[08:12] <wgrant> Name, URL and title are requested.
[08:13] <lifeless> heh. well then
[08:13] <lifeless> no idea why I saw what I did in the weekend
[08:13] <wgrant> That happens sometimes.
[08:14] <wgrant> I'm fairly concerned that it was rebranded URL there, but nowhere else.
[08:14] <wgrant> (no other views, nor the model...)
[08:21] <robin_> Hi, guys. What is the files in .bzr/packs/*? I found them are very large.
[08:22] <RenatoSilva> so what's the diff
[08:23] <robin_> I want to create new branch from existing branch in launchpad. Then I get the existing branch bzr branch lp:**********.
[08:23] <robin_> But there are lots of bzr files in the directory. In my new branch, can I delete these bzr files ?
[08:24] <wgrant> RenatoSilva: Taking bzr as an example, its name is 'Bazaar', but its title is 'Bazaar Version Control System'. Some Launchpad components follow that convention, with the title including a few words describing what it is.
[08:24] <wgrant> robin_: Lots of bzr files?
[08:24] <wgrant> robin_: What do you mean?
[08:25] <RenatoSilva> wgrant: ah ok
[08:25] <robin_> wgrant: like .bzr/repository/packs/some.pack files.
[08:26] <wgrant> robin_: Those contain the actual bzr version control data.
[08:26] <RenatoSilva> what good name and title would be for my utilities?
[08:26] <wgrant> robin_: In general, poking inside .bzr is a Very Bad Idea.
[08:26] <RenatoSilva> could the name be just RenatoSilva?
[08:26] <wgrant> Delete those packs, and you break your history.
[08:27] <robin_> wgrant: But I found they are very large. For example, I get the branch sized 871M, but the bzr files sized 600M.
[08:28] <wgrant> robin_: They store diffs to reproduce every file in the history of the branch, so they can be very large. But you can sometimes make them a bit smaller by running 'bzr pack'
[08:30] <robin_> wgrant: OK. But if I want to create new branch from existing branch in launchpad. The bzr files of existing branch should be reserved?
[08:31] <wgrant> robin_: Yes. But you can share that data between multiple local branches using a shared repository.
[08:31] <wgrant> robin_: And if the branch is in a new enough format, pushing the new branch up to Launchpad will only push the new data (using the stacked branches feature).
[08:33] <robin_> wgrant: I do it like this: bzr branch lp:~****. I get the existing branch in a directory, say, existing-branch. And I want to push it to new branch, say, new-branch.
[08:34] <robin_> wgrant: Should I run "bzr init-repos" before I push the new branch "bzr push lp:~my-id/****"?
[08:34] <RenatoSilva> confusing: http://bazaar-vcs.org/Branch.
[08:34] <wgrant> robin_: I would create a directory where I'm going to keep the project's branches, and run 'bzr init-repo .'. Then 'bzr branch lp:someproject'. Then I can 'bzr branch someproject new-branch'.
[08:35] <wgrant> robin_: Then you hack away in new-branch, and 'bzr push lp:~user/project/branch'
[08:35] <wgrant> You've only got one copy of those huge packs locally, because of the shared repo.
[08:35] <robin_> wgrant: The help of launpad.net does not explain bzr init clearly.
[08:36] <wgrant> robin_: Probably because it's more a bzr question than a LP question.
[08:36] <wgrant> robin_: Which page are you looking at?
[08:37] <robin_> wgrant: https://help.launchpad.net/Code/
[08:38] <wgrant> robin_: There are a few bzr documentation links there. Did you look at them?
[08:39] <robin_>  wgrant: Could you give some examples? What I want to do is: 1. Get existing branch;   2. Start my branch from existing branch;   3. Working on my new branch. The most problem confused me is bzr init-repo.
[08:40] <robin_>  wgrant: Or, could you explain bzr init-repo? Thank you very much!
[08:41] <wgrant> robin_: 'bzr init-repo .' will create a shared repo in the current directory. Do you understand what a shared repo does?
[08:42] <robin_> I am sorry. no:)
[08:43] <wgrant> robin_: Since the revision history data can get very large (which is what caused you to ask in the first place), bzr needs some way to share this data between multiple local branches.
[08:44] <wgrant> Or we'd all need very big hard disks.
[08:44] <wgrant> Shared repos do this - if I create a branch in a shared repo, its revisions are in fact stored in the shared repo.
[08:44] <wgrant> So if I branch from that branch into the same shared repo, the revision data doesn't have to be copied.
[08:46] <robin_> wgrant: Got it. Should I run bzr init-repo before getting  the existing branch of after getting the existing branch?
[08:46] <wgrant> robin_: Before - otherwise when you get the existing branch, bzr won't know to put the data in the shared repo.
[08:48] <robin_> wgrant: Does the things are: I run bzr init-repo to set up the shared repo; And I download the existing branch as my new branch: bzr branch lp:~existing-branch my-new-branch; Then I can work in my-new-branch  directory?
[08:50] <wgrant> robin_: Since you are likely to want to make more than one branch, and it's useful to be able to look at the existing branch, I'd first get a copy of lp:~existing-branch in my repo.
[08:50] <wgrant> So, just 'bzr branch l:~existing-branch trunk', or something like that.
[08:51] <wgrant> Then to create your branch: 'bzr branch trunk new-branch'
[08:54] <robin_> wgrant: I got it. But what is the different between bzr init and bzr init-repo?
[08:55] <wgrant> robin_: init creates a branch. init-repo creates a repository.
[08:55] <wgrant> robin_: A repository just stores data for one or more branches.
[08:55] <wgrant> robin_: A branch is the thing that has a history of code.
[08:57] <robin_> wgrant: Should I run bzr init when creating my new branch of another branch?
[08:58] <wgrant> robin_: No. 'bzr init' starts a new branch, with empty history. 'bzr branch' basically copies a branch, so the init is not required.
[09:00] <robin_> wgrant: OK! I am clearly about it. Thank you very much!!
[09:01] <wgrant> robin_: No problem.
[09:27] <RenatoSilva> I'm trying to run this:
[09:27] <RenatoSilva> bzr push lp:~renatosilva/+junk/solenoid-moin-theme
[09:27] <RenatoSilva> command-line: line 0: Bad configuration option: ClearAllForwardings
[09:27] <RenatoSilva> bzr: ERROR: Connection closed: please check connectivity and permissions
[09:29] <RenatoSilva> What could be wrong? Pagean is running with my key
[09:29] <RenatoSilva> Pageant
[09:37] <LarstiQ> RenatoSilva: that seems like your ssh client config is borken
[09:42] <RenatoSilva> it seems that I must unblock some process in the firewall but it is not bzr.exe
[09:42] <RenatoSilva> should be pageant, putty?
[09:46] <RenatoSilva>  I deactivated firewall and it didn't worked
[10:38] <tymofiy> hi there.
[10:39] <tymofiy> is there any way to submit translation to launchpad besides web-interface ?
[10:40] <lifeless> possibly via API's. And you can upload files too.
[10:40] <dpm> tymofiy: you can translate files locally and then upload the translated PO file to launchpad.
[10:41] <dpm> I don't think there is an API for that yet, but henninge might be able to tell you more on that
[10:43] <tymofiy> ok, thanks.
[10:44] <tymofiy> through I expected some bazaar integration
[10:44] <lifeless> there is, at the moment though I think its output only
[10:44] <wgrant> Launchpad can import translations and templates from bzr.
[10:44] <wgrant> lifeless: *input* only, in fact.
[10:44] <tymofiy> what a pity.
[10:45] <wgrant> http://blog.launchpad.net/translations/import-translations-from-bazaar-branches
[10:45] <james_w> wgrant: bear in mind that that might be output from lifeless' perspective ;-)
[10:45] <dpm> tymofiy: there is a planned feature for automatically committing translations to bzr branches as well
[10:45] <wgrant> james_w: Ah, that's true.
[10:46] <tymofiy> I have seen it, is there any plans for output? for localizers to be able to commit translations?
[10:46] <lifeless> actually, I thought lp could only write to bzr branches as I was discussing mechanics of that with abentley a few weeks ago
[10:47] <wgrant> lifeless: Ah, so sounds like it's coming soon, then. I wasn't aware of more than the fact it was planned.
[11:01] <henninge> wgrant, lifeless, tymofiy: That feature is currently under development and is scheduled for this cycle.
[11:01] <henninge> if other things don't get in the way ...
[11:04] <henninge> tymofiy: It is already possible for translators to commit, if I understand correctly what you mean by that.
[11:28] <RenatoSilva> thank you guys, I'm leaving, bye.
[12:38] <abentley> lifeless, dpm, tymofiy: Some implementation has been done, but the feature's not finished yet, AIUI.
[13:34] <didrocks> hi! Is there any way to check the permission level with launchpadlib the user chose when launching "get_token_and_login" (I don't find anything related to this in https://help.launchpad.net/API/launchpadlib)
[13:38] <intellectronica> didrocks: i'm pretty sure there isn't, but it's worth asking leonardr when he's around
[13:39] <leonardr> didrocks: you can only check that from the website
[13:39] <didrocks> leonardr: so, if the user click on the wrong button, I can't know it before first crash in editing, for instance?
[13:40] <leonardr> didrocks: i believe so
[13:42] <didrocks> hum. I was thinking about using manage-credentials to do that without opening the browser
[13:42] <didrocks> it's working well on edge, but crashing on staging
[13:43] <didrocks> not good if I want to test and don't create a tons of projects automatically :D
[13:46] <leonardr> didrocks, what is manage-credentials, a script?
[13:47] <didrocks> leonardr: yes, in ubuntu-dev-tools package
[13:47] <james_w> leonardr: yeah, it's in ubuntu-dev-tools so that every tool there doesn't have to deal with creating credentials
[13:48] <james_w> it has a mode where it will just not open the browser and grab the token for you, which I guess is what doesn't work on staging
[13:49] <leonardr> how does it grab the token without opening the browser?
[13:49] <james_w> do you really want to know?
[13:50] <leonardr> james_w: yes, because you might be phishing
[13:50] <james_w> it takes --email --password and --level options and does the login and post itself
[13:50] <leonardr> yeah, you're phishing
[13:50] <james_w> I'm not very keen on having it there
[13:50] <leonardr> you're negating the security advantages of oauth
[13:50] <didrocks> james_w: that'w why I was asking a cleaner method to check access level once the token is given :)
[13:51] <didrocks> that's the only part which interests me in my case :)
[13:51] <didrocks> (my tool need write level on LP given by the user, otherwise, it's unusefull)
[13:51] <james_w> negating some of the advantages at least
[13:52] <leonardr> didrocks: feel free to file a bug asking for introspection on the token within the web service. i've argued for that in the past but it's been shot down. if you show that people want it it's more likely to happen
[13:52] <leonardr> james_w: what's left? you're acting the same as a website that asks for your facebook password
[13:52] <james_w> leonardr: can we extend the interface so that an app can say in the OAuth request "I need write permission", and then LP provides a different interface
[13:52] <james_w> ?
[13:52] <leonardr> james_w: that'd be one way to solve it
[13:53] <james_w> leonardr: but I can (and have) verified that the tool doesn't store the password
[13:53] <didrocks> james_w: it's exactly what I want "write access" or "deny"
[13:53] <james_w> it's not the default behaviour of this tool or anything, and as I said, I'm not that keen on it
[13:53] <leonardr> james_w: you're training users to enter their launchpad password into any app that asks for it
[13:53] <james_w> no I'm not
[13:54] <leonardr> whoever uses manage-credentials is
[13:54] <james_w> and I encouraged didrocks to not do it this way, but with the interface that LP currently provides it requires a lot more work for the person writing the app
[13:55] <didrocks> that's why I wanted to change this. But I really need a way to check or restrict the user choice for granting access
[13:55] <leonardr> james_w: i want to make the interface as easy as possible, but the whole point of oauth is to let us tell users "you should NEVER enter your launchpad password unless you are IN YOUR WEB BROWSER and AT THE LAUNCHPAD SITE"
[13:56] <james_w> if it's possible to do this in the OAuth request then I will file a bug listing the cases that I think we should cater to
[13:56] <leonardr> if there's anything i can do to make things easier within that constraint, let me know and salgado and i will be happy to do it
[13:56] <james_w> leonardr: I have the same aim
[13:56] <james_w> great
[13:56] <james_w> I'd be looking for bugs under launchpad-foundations?
[13:57] <leonardr> james_w: yes, i think so
[13:57] <james_w> I'll file one now then
[13:57] <james_w> oh, it's already there
[13:58] <james_w> bug 199555
[13:59] <leonardr> james_w: there are some new credential-getting methods in launchpadlib you might not be aware of
[13:59] <leonardr> and once i can land the code you'll be able to get public read credentials anonymously, without making the user log in
[13:59] <didrocks> leonardr: I completed this bug report
[14:00] <james_w> that's cool
[14:00] <james_w> I've not looked at the new lplib
[14:01] <james_w> I'm still working my way through packaging the new/updated dependencies
[14:01] <leonardr> didrocks: what's the bug number? i'm probably going to merge it into 199555
[14:01] <didrocks> leonardr: I added a comment to 199555
[14:02] <leonardr> ok, great
[14:03] <didrocks> so, I'm reverting to the traditional way, waiting for a better way to ask level access and not treating the exception meanwhile :)
[14:03] <leonardr> cool
[14:08] <james_w> added my comment to the bug as well
[14:24] <leonardr> james_w: i filed bug 387297 against ubuntu-dev-tools
[14:39] <sinzui> barry_: ping
[14:39] <barry_> sinzui: pong
[14:40] <sinzui> barry: I don't think you are on #launchpad-code, are you ready for our standup?
[14:40] <barry> sinzui: yep. could be irc disconnected me
[14:55]  * RenatoSilva thinks that +jungle would be better than +junk.
[15:20] <RenatoSilva>  I tagged a specific revision, but when I commit or push, it stands that there's nothing to change. Why?
[15:20] <RenatoSilva> It seems that new tags are not considered as changes
[15:24] <kiko> RenatoSilva, I'm not entirely sure, hmm
[15:24] <kiko> abentley, what do you think of the above?
[15:26] <abentley> kiko, RenatoSilva: tags aren't part of revisions, which is the main thing "push" considers when it emits that message.  I'll have to check whether they're pushed anyway in that case.
[15:26] <RenatoSilva> I was talking in #bzr. Tags are pushed anyway.
[15:27] <RenatoSilva> I just can't find them in Launchpad tough.
[15:27] <leonardr> james_w: i just had a talk with flacoste and mars about bug 387297
[15:27] <leonardr> i'm writing an update
[15:28] <james_w> cool
[15:29] <RenatoSilva> it seems a loggerhead issue
[15:36] <RenatoSilva> abentley, kiko: bug 246739
[15:45] <kiko> RenatoSilva, are tags visible from anywhere on launchpad itself?
[15:47] <RenatoSilva> kiko: I don't think so
[16:17] <didrocks> can somebody create gtkhtml3.14 project in LP please? (it differs from gkthml)
[16:55] <fta> didrocks, can't you do it yourself?
[16:58] <didrocks> fta: I think, but the process was to ask to jcastro (I don't know if he setup particular things like upstream bug) and as he is busy, he told me to ask here :)
[16:59] <fta> oh, ok then. i thought it was open to anyone.
[17:00] <fta> ripps, is https://edge.launchpad.net/~ripps818/+archive/staging fully automatic now?
[17:00] <fta> oops, wrong channel
[17:01] <ripps> fta: Having put it in a cron job yet, but it still works everytime I run the script so far.
[17:01] <ripps> *Haven't
[17:01] <fta> excellent
[17:02] <kiko> didrocks, go ahead and create it and I can do any admin stuff on it that is necessary
[17:03] <didrocks> kiko: I just don't know what can of admin stuff jorge is doing, if he does :). I will ask him later, when he will be less busy
[17:04] <kiko> well, just show me -- it's likely to be checking the project, linking it to any ubuntu packages, reassigning it to ~registry etc
[17:08] <didrocks> kiko: ok, I created https://edge.launchpad.net/gtkhtml3.14
[17:08] <didrocks> this has to be related to this package: https://edge.launchpad.net/ubuntu/+source/gtkhtml3.14
[17:08] <kiko> let me check
[17:09] <didrocks> I tried to set the maintainer to "Registry Administrators", but as I'm not part of the team, I can't
[17:09] <didrocks> kiko: if you want a recent example jcastro has done, you can check there: https://edge.launchpad.net/libgtop
[17:18] <kiko> didrocks, can you confirm they use the gnome bug tracker?
[17:22] <didrocks> kiko: I'm looking, maybe pedro knows this
[17:23] <didrocks> kiko: I'm not sure if they use gtkhtml as a product, or gtkhtml2... I prefer to ask to pedro first
[17:28] <ahasenack> hi guys, the landscape "later" milestone is still timing out in LP (production and edge), any news about that? It has 287 bugs, most of them private, and from previous discussions I understand this is the issue (the security check)
[18:01] <kiko> ahasenack, that's interesting! do you have an OOPS ID for us?
[18:04] <stefanlsd> From the Launchpad API, what does 'Link to a file resource. ' mean?
[18:06] <intellectronica> stefanlsd: it's a link to a file, hosted on the launchpadlibrarion
[18:07] <stefanlsd> intellectronica: in this case, i would like to attach a file to a bug report...
[18:08] <intellectronica> stefanlsd: you can use the addAttachment (iirc) method on Bug
[18:09] <stefanlsd> intellectronica: yeah. I am looking at the API doc for it, although the one required field is  'Link to a file resource'. I've tried a python file resource, but doesnt seem to work
[18:11] <intellectronica> stefanlsd: i think you just pass the data itself when using launchpadlib
[18:27] <Scurz> hi
[18:28] <Scurz> I've created a new mailing list on launchpad and would like to know who (if these personn are presents) accepts or refuses the new mailing lists ?
[18:29] <beuno> Scurz, approved
[18:30] <Scurz> beuno: oh, thanks to you !
[18:33] <Scurz> hmm, I have a question, can I subscribe someone to the mailing list who has no account on launchpad ?
[18:50] <beuno> Scurz, no, only launchpad memebers can subscribe
[19:05] <Scurz> beuno: ok
[19:06] <vadi2> Where can I go to unsubcribe from bug mail from a particular project?
[19:12] <vadi2> ah, got it. wasn't really intuitive to me, I looked everywhere in my personal preferences first, and 'Subscribe to bug mail' when I want to unsubscribe is eh :)
[19:35] <aboudreault> emm...
[19:36] <aboudreault> I selected "Copy binaries ..." in launchpad and my package is "waiting to be built"
[19:36] <aboudreault> And that build should fails at the upload ... it's the second time I got a similar error
[19:37] <aboudreault> Does that bug is known ?
[19:37] <kiko> aboudreault, it's more likely that there's something wrong with the package
[19:38] <aboudreault> What could be wrong ?
[19:38] <kiko> aboudreault, you can try asking cprov for help, or at least paste-bin the build failure so others can see it
[19:38] <kiko> it /could/ be a bug -- there is a bug around that is related to binary copies indeed -- Ursinha or matsubara can find the number for us
[19:39] <cprov> aboudreault: point me to the PPA page, I can take a look
[19:39] <aboudreault> the packge built successfully.... for jaunty. Then I just wanted to copy binaries to intrepid. And ... a rebuild is "forced" or something...
[19:39] <aboudreault> https://launchpad.net/~ubuntugis/+archive/ubuntugis-unstable
[19:39] <aboudreault> You can see the package gdal-ecw , which has a failure for intrepid
[19:40] <aboudreault> and I'm absolutly sure that I've selected "copy binaries" and not the rebuild option.
[19:41] <aboudreault> off course, I can solve this by creating a package-1-1~intrepid1 version... just wanted to report it.
[19:44] <cprov> aboudreault: fear not, I'm checking.
[19:45] <aboudreault> Ok, will wait
[19:51] <cprov> aboudreault: did you get an copy error before it succeed ?
[19:52] <cprov> aboudreault: which forced you to re-select the source you wanted to copy  ...
[19:52] <aboudreault> not for intrepid. it has been copied without error. after, tried to copy the package from jaunty to hardy, and get the error that the package was building for intrepid.
[19:57] <cprov> aboudreault: that's weird, let me play more with it in my ppa. You can just delete the intrepid copy and copy again from jaunty with binaries.
[19:59] <aboudreault> Ok... will try to recopy
[20:01] <cprov> aboudreault: okay, let me know if it works or crashes this time. Sorry for the inconvenience.
[20:03] <aboudreault> the build status of the old one is set to the new copy.
[20:05] <cprov> aboudreault: yes, that's a side-effect of having a built in intrepid that shouldn't be there, although the repository will be sane now
[20:05] <cprov> aboudreault: i.e, jaunty and intrepid will have the same binaries (the ones built on jaunty)
[20:07] <aboudreault> Ok, so the package will be published as normal ?
[20:08] <cprov> aboudreault: yes, that's what we can find out in 3 minutes
[20:09] <aboudreault> all right. thx
[20:09] <cprov> aboudreault: np.
[20:12] <aboudreault> cprov: same problem with hardy
[20:12] <aboudreault> will do the same thing after the build fail
[20:12] <cprov> aboudreault: but you copied it while we were talking ?
[20:13] <aboudreault> I just tried to copy it 2 minutes ago
[20:13] <aboudreault> when the intrepid one was marked as "published"
[20:14] <cprov> aboudreault: I'm trying to reproduce it in my PPA and I can't, copy fails
[20:15] <cprov> what kind of race condition is that ...
[20:15] <aboudreault> just a guess, try selecting the "Copy binaries" option before selecting the destination serie.
[20:18] <cprov> aboudreault: found it, 'copy binaries' before there are any binaries published.
[20:19] <aboudreault> hmm
[20:19] <aboudreault> Ok, will copy the binaries as soon as they are uploaded in the future
[20:20] <cprov> aboudreault: the backend thinks the source has failed in the series it was previously uploaded and retries it in the destination series.
[20:21] <aboudreault> Ok. my package for intrepid is marked "Published   14 minutes ago" but can't get it with apt-get update && install. I think I have to reupload a modified version.
[20:22] <cprov> aboudreault: that's the source, not the binaries.
[20:23] <aboudreault> Ok, will wait a while.
[20:27] <aboudreault> but if the build status is a failure, the binaries will not be published, no?
[20:28] <cprov> aboudreault: no, not this simple, your binaries are indeed not published. Let me digg more.
[20:29] <aboudreault> all right.
[20:36] <eagle00789> how to fix the Cannot find svn repository root. problem??
[20:37] <sinzui> barry:  ping
[20:38] <barry> sinzui: pong
[20:38] <sinzui> barry I think you fell off #launchpad-code
[20:38] <barry> sinzui: irc sucks
[20:38] <sinzui> barry: barry I'm reading https://bugs.edge.launchpad.net/launchpad-registry/+bug/350390
[20:39] <cprov> aboudreault: the binaries are not being copied to the other series (the UI might be saying that, only one record in the information-message rendered after the copy)
[20:43] <cprov> aboudreault: are you selecting the jaunty published record to copy over the other series ?
[20:44] <aboudreault> yes
[20:44] <aboudreault> jaunty to intrepid,hardy
[20:44] <cprov> aboudreault: the copy works on https://edge.launchpad.net/~cprov/+archive/sandbox, despite of the fact the builds are broken, the binaries end up being published
[20:45] <cprov> aboudreault: in your case there are only binaries for jaunty.
[20:45] <aboudreault> k, so for that time, I'll have to reupload a modified version of the package
[20:47] <cprov> aboudreault: if you can wait few minutes more, the binaries can be copied.
[20:48] <aboudreault> Unfortunately, I gtg. I'll let the repository like it is... and reupload a new version (if needed) tomorrow morning.
[20:49] <aboudreault> If you have more info, keep me informed in private please. (to be sure I'll see the message)
[20:49] <aboudreault> Thanks again for your help.
[20:50] <cprov> aboudreault: deal.
[20:56] <eagle00789> how to fix the Cannot find svn repository root. problem??
[21:03] <eagle00789> this is my import url: svn://smartass.webhop.info:443/share/HDA_DATA/svn/ToolbarKiller
[21:03] <eagle00789> can someone check please what i did wrong so i can have launchpad do an import of the code??
[21:11] <eagle00789> anyone??
[21:11] <eagle00789> i'm struggling with this problem for days now...
[21:14] <fta> {code,bazaar}.lp should display the svn rev-ids when a branch is imported from svn. the bzr revs have no value upstream and are causing confusion. at least, the two should be display, like in bzr log
[21:15] <fta> +ed
[21:15] <eagle00789> and how to fix it??
[21:17] <eagle00789> the problem is that this still is a initial import wich doesn't run...........
[21:29] <eagle00789> this is the last problem i have to get launchpad working for me....
[21:51] <thumper> eagle00789: ask a question on launchpad-code project and include a link to the import branch
[21:52] <eagle00789> àlready done about 2 days ago. still no answer...
[21:53] <eagle00789> sorry, has been 4 days already
[21:55] <eagle00789> last post: https://answers.launchpad.net/launchpad-code/+question/73930
[22:02] <Saviq> hi all, any word on https://answers.launchpad.net/launchpad/+question/73756 ?
[22:30] <eagle00789> don't waste your breath here. i'm also still waiting for an answer on one of my questions
[22:33] <Saviq> yeah it seems there's not enough staff working @ launchpad recently
[22:34] <eagle00789> if there is any staff working at launchpad at al.
[22:35] <eagle00789> they are currently spending way to much time on open sourcing launchpad... (don't know how they can call it that, when the most used features of launchpad will remain closed source)
[22:40] <exarkun> enjoy your whine-fest
[22:40] <eagle00789> i'm not whining. it's just frustrating...
[22:41] <eagle00789> each and every question i ask everywhere lately is neglected. even Experts Exchange. have a question there open for more then 2 weeks now without ANY responce...
[22:46] <Saviq> eagle00789: still, this kind of attention won't get you any good
[22:46] <eagle00789> i know... you are right
[23:13] <jkakar> Bug folks: thanks for adding the lp: to the beginning of branch URLs.  It makes pasting them a little bit easier, which adds up when you do it 20 times a day.
[23:15] <RenatoSilva> Hey guys, do you remember our talk about branches and projects?
[23:15] <RenatoSilva> Can I create a project called "branches"?
[23:16] <wgrant> RenatoSilva: Why would you want to do that?
[23:16] <wgrant> Unless you have some software named 'branches', that seems like a bad idea.
[23:19] <RenatoSilva> wgrant: it's a meta project, just an alternative for +junk
[23:20] <wgrant> RenatoSilva: If you want bugs for your branch, you should probably just create a project for that branch.
[23:20] <wgrant> Anything else is probably going to be very ugly.
[23:20] <RenatoSilva> a place to put non-project branches which aren't junk either
[23:21] <wgrant> RenatoSilva: If it is non-project, it is by definition junk.
[23:21] <RenatoSilva> wgrant: actually, bugs is secondary, the main thing is to publish code
[23:22] <wgrant> RenatoSilva: What is the benefit of having that project, rather than just using +junk?
[23:22] <RenatoSilva> wgrant: junk is a strong word, it sounds like old stuff, gabbage
[23:22] <wgrant> RenatoSilva: That's not the intended meaning here.
[23:22] <wgrant> And you don't work around bad terminology by creating hacks.
[23:22] <RenatoSilva> wgrant: main benefit is the label
[23:24] <RenatoSilva> wgrant: I just think it's a better neutral level, I can't avoid associating +junk with sand boxes or so. Maybe it's not the sense, but the word is strong.
[23:24] <wgrant> RenatoSilva: File a bug, perhaps. But creating projects with the same purpose as +junk isn't right.
[23:24] <RenatoSilva> s/level/label
[23:27] <RenatoSilva> wgrant: I do have some stuff here that 'junk' fits well, for example learning_java or so... but other stuff is more than junk, and less than project...
[23:27] <wgrant> RenatoSilva: Why is it less than a project?
[23:27] <wgrant> A project has no minimum size.
[23:39] <RenatoSilva> wgrant: I think it's annoying, imagine 50 projects of single files
[23:40] <RenatoSilva> some code just do a litte but useful work, I think it's weird calling them proooojects
[23:41] <RenatoSilva> can I really set up many projects of little stuff? Won't you find it annoying?
[23:59] <RenatoSilva> wgrant: bug 387552