[00:00] <thumper> chx: stacking is done based on the development focus branch of the project
[00:00] <chx> ah! good
[00:00] <thumper> chx: also why do it this way?
[00:01] <chx> thumper: what would be a better way?
[00:01] <thumper> well...
[00:01] <thumper> I have trunk branched locally
[00:02] <thumper> when I create a new branch I branch it locally in my shared repo
[00:02] <thumper> and pushing it to lp is as simple as "bzr push"
[00:02] <thumper> as I have the public locations specified in ~/.bazaar/locations.conf
[00:02] <thumper> this branch is then stacked and fast to push
[00:02] <luke-jr> pushing a branch initially takes forever regardless of lp-to-lp or local-to-lp in my experience :(
[00:02] <thumper> luke-jr: not always
[00:03] <thumper> luke-jr: I can push a new launchpad branch to lp very quickly
[00:03] <thumper> from NZ even
[00:03] <luke-jr> also, you get free backup by branching lp-to-lp and working with a checkout :)
[00:03] <thumper> as long as it is stacked
[00:03] <thumper> luke-jr: I get free backup by pushing as needed
[00:03] <rockstar> luke-jr, then you don't get the distributed part of DVCS...
[00:03] <persia> a pathological case for stack-on-trunk is for new branches that add bugfixes to earlier releases, where there's vast divergence from trunk, but not so much divergence from a known non-trunk branch.
[00:03] <luke-jr> rockstar: you can always unbind if you need to later :)
[00:04] <thumper> persia: true
[00:04]  * persia has encountered this several times
[00:04] <thumper> persia: you can specify --stacked-on if you really care
[00:04] <mwhudson> however
[00:04] <mwhudson> --stacked-on doesn't really work too well
[00:04] <mwhudson> (i'm fixing that!)
[00:05] <luke-jr> more important to me IMO is the ugliness of the overview page :P
[00:05] <thumper> luke-jr: which overview page?
[00:05] <luke-jr> https://launchpad.net/armagetronad
[00:05] <luke-jr> the download list especially
[00:05] <persia> Plus, I never remember that I can do something fancy until I'm already incredibly annoyed at how long it takes to push.
[00:06] <luke-jr> not that "Series and milestones" gives any meaningful representation of releases
[00:06] <merbit> has anyone noticed "MemoryError" in ubuntu package branches? https://code.edge.launchpad.net/~ubuntu-branches/ubuntu/lucid/sun-java6/lucid
[00:06] <rockstar> persia, what formats are your branches?  It never takes much to push for me.
[00:07] <persia> rockstar: Whatever they were in LP.  I'm spoiled by ftp being incredibly fast, so impatient with any more intelligent protocol.
[00:07] <luke-jr> ...
[00:07] <luke-jr> the whole point of a custom protocol is always the speed :p
[00:08] <thumper> merbit: where are you getting MemoryError?
[00:09] <merbit> thumper: on that link, it has a triangle error and shows MemoryError
[00:09] <merbit> I'll take a screenshot
[00:09] <persia> luke-jr: bzr+ssh is faster for some purposes.  I have a somewhat different set of constraints, so that computation takes more time than wire transfer.  This isn't generally the case.
[00:09] <thumper> merbit: I see it
[00:09] <merbit> ah ok :)
[00:10] <merbit> thumper: I was wondering if that's normal, since lucid isn't stable yet
[00:11] <thumper> merbit: no, I think it is the java6 package in particular
[00:11] <merbit> ok thank you
[00:12] <merbit> oh right.. sun-java6 has been moved to partner repo
[02:14] <nnewton> hi all, I'm having some issues with login...mainly that I'm getting the "Oops" error on login
[02:14] <nnewton> wondering if there is a contact email to..email..or if I should file a bug (which maybe hard considering I can't login)
[02:34] <doctormo> Hey rockstar, how's development going?
[02:34]  * rockstar looks up
[02:34] <rockstar> doctormo, is there context here that I'm unaware of, or are you just asking in general?
[02:35] <doctormo> rockstar: Asking in general, have you never had someone care about your well being without context?
[02:35] <rockstar> doctormo, not anyone that isn't my wife, no.  :)
[02:35] <rockstar> (that includes my manager, thumper)
[02:35] <rockstar> :)
[02:36] <thumper> doctormo: you may be interested to know that we will be trying some build package from branch on dogfood RSN
[02:36] <rockstar> doctormo, things are going well.
[02:36] <rockstar> YESSIR!  Last few branches should be landing shortly.
[02:36]  * thumper is hurt
[02:36] <thumper> I ask how things are going
[02:36]  * rockstar kids
[02:36]  * thumper has kids
[02:36] <doctormo> thumper: Interesting, are there some good API changes?
[02:37] <thumper> abentley: is working on the api
[02:37] <rockstar> doctormo, you should hear the cursing right now...
[02:37] <thumper> there are challenges
[02:37] <doctormo> Brilliant, did we get the addition of SSH key management yet?
[02:37] <thumper> doctormo: additions rather than changes I hope
[02:37] <rockstar> doctormo, as far as I know, that's registry's thing.
[02:37] <rockstar> Abuse sinzui
[02:37] <thumper> doctormo: ask sinzui
[02:37] <thumper> ask with a stick
[02:38] <doctormo> thumper: For some reason you remind me of a fluffy bunny, I think I'm in bambi's forest.
[02:39] <thumper> doctormo: I have that picture on some things
[02:39] <doctormo> I was noticing a consistancy problem before, who's the fella that deals with UI?
[02:39] <thumper> doctormo: check https://launchpad.net/~thumper
[02:39] <thumper> doctormo: lots of people do
[02:39] <thumper> doctormo: hence the consistancy problem
[02:39] <mwhudson> hence the awesome consistency
[02:39]  * rockstar lulz
[02:39] <thumper> doctormo: which exact problem?
[02:40] <thumper> doctormo: we are sprinting right now, hence the collective answers
[02:40] <rockstar> (also, it's Friday, and our brains are mush)
[02:41] <doctormo> I don't supose there is a web devel / design position going spare? ;-D
[02:41] <thumper> doctormo: two actually
[02:41] <thumper> doctormo: kinda
[02:42] <thumper> doctormo: there is a web developer for launchpad registry, and I'm after someone as a general dev
[02:42] <sinzui> doctormo, sshkeys were exposed by several people over one week
[02:43] <thumper> hi sinzui
[02:43] <rockstar> See? sinzui responds to abuse.
[02:43] <thumper> rockstar: I think he responds more nicely to asking
[02:43] <sinzui> I usually respond in kind
[02:43] <thumper> heh
[02:43] <rockstar> :)
[02:43]  * sinzui is writing a lot of abuse into the already insane CSS file
[02:44] <doctormo> thumper: Interesting, well I have both design and developer xp. Level 15 python wizard, Level 20 perl ninja.
[02:44] <thumper> sinzui: want to come to NZ?
[02:44] <sinzui> yes
[02:44] <doctormo> I don't suppose there'll be many people at UDS from lp?
[02:44] <thumper> sinzui: ACLs could be a good reason
[02:44] <thumper> doctormo: there should be quite a few
[02:45] <thumper> doctormo: mwhudson, abentley, all of soyuz
[02:45] <sinzui> doctormo, one or two people form each team. I think that means about 8
[02:45] <mwhudson> i won't be launchpad by then officially
[02:46] <thumper> doctormo: here is my position - http://webapps.ubuntu.com/employment/canonical_LP-SEC/
[02:46] <thumper> doctormo: where do you live?
[02:49] <doctormo> Boston, MA, USA
[02:49] <thumper> ah, been there
[02:49] <thumper> one of the few places in the USA I've been to
[02:49] <doctormo> OK back to that problem, took me a min to get a screenshot: http://imagebin.ca/view/GpIJMg6E.html
[02:49] <thumper> the ui problem?
[02:49] <doctormo> yes
[02:50] <rockstar> doctormo, bugs team's problem.  :)
[02:50] <doctormo> The text and the icons do different things for each of them, some popup selectors, some are links
[02:50] <doctormo> worst is the Target to milestone which is the oposite of the affects
[02:51] <sinzui> doctormo, you missed my greatest grievance in the bug widget. If you assign the milestone the link and js-action reverse.
[02:51] <thumper> doctormo: file a bug :)
[02:51] <thumper> upload the pic as an attachment :)
[02:52] <doctormo> Aye will do
[03:29] <doctormo> thumper: I don't suppose you'd make me a quote for the ground control website on what you think of it as a project?
[03:30] <thumper> doctormo: perhaps you should ask someone that uses gnome :)
[03:30]  * thumper uses kubuntu
[03:30] <doctormo> Although you don't have to be a gnome user to understand the workflow, I'll ask er, persia, you know ground control right?
[03:31] <persia> I've heard of it, yes.
[03:31] <doctormo> thumper: Some dolphin devs wanted to know how easy it'd be to make into a kde thing.
[03:31] <thumper> doctormo: that'd be interesting
[03:31] <persia> doctormo: You're using pygtk, right?
[03:31] <doctormo> Yes
[03:32] <thumper> doctormo: it ground control abstracted enough to split out the gui work so there could be a pyqt version?
[03:32] <persia> If you've abstracted the core as is advised by the docs, adding a python-qt frontend ought be a couple days effort.
[03:32] <thumper> persia: snap
[03:33] <persia> Mind you, "ought be" is almost never true, because invariably one discovers refactoring opportunities whilst doing so, but still.
[03:33] <persia> thumper: You're a KDE guy: maybe you want to toss up a quick framework?
[03:33] <thumper> persia: no... I'm a KDE user
[03:33] <doctormo> It's fairly abstracted, especially bzr / launchpadlib intergration, but the gui stuff is interesting, it's got 10 UIs etc.
[03:33] <persia> thumper: Ah.
[03:34] <thumper> persia: and my spare time is taken up with another project right now
[03:34] <thumper> persia: I'm writing yet another wiki
[03:34] <persia> Why?
[03:34] <persia> Or I should ask: which feature drove you to this extreme?
[03:34] <doctormo> It's a shame there isn't a way to get glade ui files working with pyqt
[03:35] <persia> You'd have to use Qt Designer to achieve the rough equivalence.  I suspect you'd do best to have someone else work on that, and collaborate on the abstraction.
[03:37] <thumper> persia: I want it bzr backed, and modular enough to be integrated with launchpad
[03:39] <doctormo> thumper: You want the wiki to be backed by bzr/lp?
[03:39] <thumper> doctormo: I want the wiki to be backed by a bzr branch
[03:40] <thumper> doctormo: and have it able to be integrated with LP
[03:40] <thumper> doctormo: so you should be able to go "bzr branch lp:project/wiki" to get the wiki for the project
[03:40] <doctormo> If it means you can start doing propper custom tables
[03:40] <doctormo> The problem with the wiki is far too many people are using it like a database.
[03:40] <persia> Hrm?
[03:41] <persia> thumper: That's a good set of reasons to write a new one :)
[03:41] <thumper> lp:wikkid
[03:41] <thumper> it is still really early
[03:41] <persia> doctormo: How do folks use the wiki like a database?
[03:41] <thumper> I'm working on it in the evenings
[03:41] <thumper> I'm hoping for an initial release soon
[03:41] <thumper> soon meaning within a month or two
[03:41] <luke-jr> meh
[03:42] <luke-jr> just drop the pygtk and switch to pyqt only
[03:42] <persia> thumper: That's hugely exciting.
[03:42] <thumper> persia: I think so
[03:42] <persia> luke-jr: dropping isn't best for GNOME folk, but maybe you could help make it work well for KDE folk?
[03:43] <luke-jr> GNOME folk are fictional :)
[03:43] <persia> ground-control upstream are GNOME folk: consider the power of introducing them to the beauty of pyqt code as a means of advocacy :)
[03:44] <luke-jr> what's ground control anyhow?
[03:44] <doctormo> persia: https://wiki.ubuntu.com/UDS-M/Attendees <- tell me that isn't being used as a giant database table
[03:45] <persia> doctormo: I consider that wiki abuse, especially since LP *already* exposes a bundle of that information.
[03:45] <doctormo> luke-jr: Demonstration http://blip.tv/file/3141629
[03:45] <doctormo> persia: The problem is that it isn't exposing enough and a lot of what goes on is custom tables.
[03:45] <doctormo> persia: It'd be nice to have a wiki-db
[03:45] <persia> The right answer is to add travel details input to the LP sprints interface.
[03:46] <doctormo> It is, that sort of thing should be formal and nailed down, but there is plenty of other instances of tables which are experimental and use once.
[03:46] <luke-jr> my daughter says ground control is about an air balloon
[03:47] <thumper> persia: patches accepted :) (well - reviewed anyway)
[03:49] <persia> thumper: I'm the wrong person to ask for patches :)  My preferred language is make, and nearly all the time I have for development is spent fixing stuff that gets in the way of what else I do.
[03:50] <doctormo> persia: Make as a language is liek alien abductions, there are many stories about it but I'm critical that it exists at all.
[03:51] <persia> doctormo: http://people.ubuntu.com/~persia/ubisim is a quick hack script, http://bazaar.launchpad.net/~persia/+junk/moblin-analysis/annotate/head:/Makefile is a more powerful tool.
[03:52] <spiv> doctormo: something a bit like freebase.com?
[03:52] <doctormo> Holy smokes, it's like someone coding Quake 3 in DOS Batch
[03:53] <doctormo> luke-jr: Did the video make sense?
[04:05] <luke-jr> no
[04:05] <luke-jr> I had no sound from it
[04:05] <luke-jr> looked like a GNOME addon for LP
[04:07] <persia> It's definitely a tool for manipulating LP from the desktop.  That it's GNOME is a bug :)
[04:09] <mwhudson> persia: real programs just talk native x11 protocol stuff over a socket?
[04:11] <luke-jr> real programs are Qt
[04:11] <luke-jr> :)
[04:12] <persia> mwhudson: No, real programs are sufficiently abstracted that they can be trivially made to work with arbitrary desktop environments and have a sufficiently diverse development community that this just works.
[04:12] <luke-jr> L(
[04:12] <luke-jr> :)*
[04:13] <mwhudson> persia: i think by that criteria no real programs exist
[04:13] <luke-jr> actually, in all seriousness, real programs are UI-less and have UI implementations independent from the purpose
[04:13] <thumper> real programs only have command line interfaces
[04:14] <thumper> and aren't used by real people
[04:14] <thumper> only geeks
[04:17] <persia> mwhudson: ubiquity is a nice example of a real program by those criteria, just to pick one.
[04:17] <persia> luke-jr: Well, are architected in a way that enables that, yes.
[04:17] <persia> thumper: No.  A CLI is a UI.
[04:17] <spm> real programs don't even have cmd line interfaces and are handled by batch scheduling systems :-P
[04:18] <luke-jr> on another note, all this abstraction makes things bloated
[04:18] <luke-jr> Windows NT ran with 16 MB RAM
[04:19] <doctormo> persia: It'll probably make you happy to hear that ground control is also cli, so you get a choice between cli and gnome :-D
[04:20] <persia> spm: A batch scheduling system is an interface: if it *only* works with batch scheduling systems, that's a bug.
[04:20] <luke-jr> doctormo: isn't that self-contradicting?
[04:20] <persia> doctormo: That's a good step towards solving the bugs :)
[04:20] <luke-jr> I thought ground control was supposed to be a UI for bzr/lp
[04:20] <doctormo> luke-jr: Do you think I'd test it by quiting and reloading nautilus several million times?
[04:21] <luke-jr> lol
[04:21] <thumper> you can fix any problem by adding another level of indirection
[04:22] <persia> Well, except insufficient computing resources :)
[04:22] <doctormo> thumper: It's the magic trick: Look over there, *alakzam!*
[04:22] <thumper> persia: if the indirection is to another machine, that can work too :)
[04:23] <doctormo> http://divajutta.com/doctormo/gcweb/ <- thoughts on the icon/logo?
[04:23] <luke-jr> hehe, deploy VNC clients :p
[04:23] <persia> distributed resources are still resources :)
[04:23] <doctormo> luke-jr: Just incase you ever get the urge to develop a Dolphin plugin for groundcontrol, now you know how it works.
[04:24] <persia> doctormo: Looks like a cyclops: consider a rocketship with 4 windows so that two show.
[04:24] <luke-jr> doctormo: I don't like Dolphin
[04:24] <luke-jr> I do all my file management from a CLI
[04:25] <doctormo> persia: With two visible it looks like a squid
[04:25] <doctormo> luke-jr: Well, you may have friends who you want to help you out with some copy editing for a branch, and then they'll _have_ to use gnome ;-P
[04:26] <luke-jr> doctormo: I doubt I'll ever trust someone who can't use a CLI with my code
[04:26] <doctormo> Actually it sort of looks like the rocket is wearing a bra... odd
[04:27] <doctormo> luke-jr: The key phraise was "copy editing", i.e. they're helping you with docs, design, something non-code.
[04:27] <luke-jr> let me rephrase:
[04:27] <luke-jr> someone who can't use a CLI should be banned from computers
[04:27] <luke-jr> :)
[04:30] <persia> Let's not ban anyone.  Instead, let's enable them to use and learn.
[04:30] <luke-jr> by taking away the GUIs
[04:30] <luke-jr> <.<
[04:30] <doctormo> persia: I've updated the icon, added a little more perspective
[04:31] <doctormo> luke-jr: No wonder there are so many tools that stink, did you design "find" ?
[04:31] <persia> doctormo: I like that better.
[04:31] <luke-jr> doctormo: no, sorry
[04:32] <luke-jr> I do concur find has a terrible interface tho
[04:32] <luke-jr> other than simple stuff, I almost always need the man page
[04:32] <doctormo> luke-jr: There are examples of sorry deisgn in CLI, and I know some developers who like the CLI because they believe that design isn't required at all.
[04:32] <doctormo> Or worse, they only do API because they think it needs no design.
[04:32] <luke-jr> http://forums3.armagetronad.net/viewtopic.php?f=1&t=20185 <-- lol
[04:33] <persia> design is critical to all interfaces.  The set of criteria that make an interface good varies hugely on the interface type though, so being good at "interface design" is in fact guaranteed to be false.
[05:30] <bernie> I've uploaded a package to my ppa with dput. the upload was successful, but I don't see it in the web page.
[05:31] <bernie> I waited 10 minutes... when should I start to get worried?
[05:33] <persia> Check your mail: you should have an ACCEPT or REJECT message.
[05:33] <persia> If you don't, check to make sure your .changes file was signed by a key that is registered with your account on launchpad.
[05:34] <bernie> persia: oh, right.
[05:34] <bernie> persia: the package was rejected indeed
[05:34] <bernie> Unable to find distroseries: unstable
[05:34] <persia> Right.  You need to specify an Ubuntu target for a PPA.
[05:35] <bernie> I guess I have to fix my changelog?
[05:51]  * lamont has a stupid question...
[05:52] <lamont> I have a no-tree branch on my machine, want to pull a branch from launchpad and have it use the local no-tree branch for everything it can, rather than pulling everything from launchpad...
[05:52] <lamont> how painful is that?
[05:53] <spiv> lamont: pretty easy.
[05:53] <lamont> trying the branch the local and merge the lp branch, since ultimately that's my goal anyway
[05:53] <spiv> Yep, that works fine.
[05:54] <lamont> oh, and long time no chat.  howdy
[05:54] <spiv> If you had a shared repo for that branch, then you could just "bzr branch lp:whatever" into that repo and get the same benefit.
[05:54] <spiv> Howdy :)
[05:55] <spiv> lamont: here, have a gratuitous happy baby photo: http://gallery.puzzling.org/v/parenting/vincent/month2/IMG_7702+_Developed+in+UFRaw_.jpg.html
[05:56] <lamont> most cute
[05:57] <lamont> I think I have a shared repo.  at least if I didn't screw it up when I built it
[05:57] <lamont> it would have helped if I hadn't deleted my local copy of the branch tree as "done" earlier...
[06:06] <spiv> You can also do "bzr checkout ." to make a tree in a no-tree branch, btw.
[06:07] <lamont> ok.  anyway, face -> pillow.  g'night
[06:08] <spiv> G'night!
[09:24] <allquixotic> Hi -- is there any way to allow a Launchpad Team to have read-only access to some (but not all) private branches associated with a commercial project? For instance: our dev team has read/write access to all our private branches, but we want to give a QA team the ability to check out and build the source, without making it *open* source, and without allowing them write access.
[09:27] <mwhudson> allquixotic: subscribe them to the branches you want them to access
[09:28] <allquixotic> mwhudson: And that will give them read-only access?
[09:28] <allquixotic> Only the maintainer of the project can write, I suppose?
[09:29] <mwhudson> allquixotic: the owner of the branch can write
[09:29] <mwhudson> allquixotic: subscribers have read only access
[09:29] <allquixotic> Good to know. Thank you!
[13:51] <mroos> hello. is this the right place to ask help about launchpad login problem?
[13:51] <mroos> I try to reset my forgotten password but it tells me my account is disabled, an I'm stuck
[13:56] <BlackZ> mroos: what's your LP ID?
[13:58] <mroos> mroos@linux.ee is the email I'm trying to use
[13:58] <mroos> and looks like I have used or tried to use it before
[14:00] <BlackZ> mroos: https://launchpad.net/~mroos/+claim
[14:04] <mroos> thanks, it works!
[14:12] <pmatulis> is there a sandbox where i can create a test bug?
[14:15] <maxb> pmatulis: staging.launchpad.net
[14:16] <pmatulis> maxb: thanks
[14:19] <pmatulis> maxb: "Sorry, there was a problem connecting to the Launchpad server."  Reloading doesn't help.  :(
[14:21] <maxb> losa ping: staging is broken right now? please advise?
[14:22] <mthaddon> maxb: it isn't a service we advertise with any SLA - but as it happens there's an update in progress which is why it's down
[14:23] <maxb> mthaddon: Is there any way people can tell whether its down for routine updates vs. just being broken?
[14:23] <mthaddon> maxb: not currently - we're working on improving that
[14:24] <maxb> Because so long as it's linked from lpnet's frontpage, there's something of an implicit SLA of "we'll tell you when we break it" :-)
[14:25]  * mthaddon wasn't aware it was linked from the front page 
[14:25] <maxb> It's the first item under "Get Started"
[14:27] <mthaddon> maxb: hmm, am I missing something - I see the text, but no link
[14:27] <maxb> mthaddon: "What's this"
[14:27] <mthaddon> maxb: nm, I see it once you click on "what's this" - thx for letting me know about that
[18:34] <paultag_> Is there any page that outlines the launchpadlib API? I have found the LP API docs, and some basic launchpadlib walkthroughs, but I can't find documentation on how to get members of a given team. Any ideas, hackers?
[18:36] <james_w> paultag_: there's no page I know of that is a guide to the overall API
[18:36] <james_w> it's pretty much look at the API and guess what you need to do right now
[18:37] <blueyed> what's with OOPS-1574D2190 ?
[18:37] <paultag_> james_w, Humm. OK. Are contributions welcome for documentation?
[18:37] <james_w> paultag_: yes!
[18:37] <paultag_> james_w, OK. I might get to that this weekend. Just wanted to make sure it was not already done. Thanks :)
[18:38] <beuno> blueyed, a timeout
[18:39] <blueyed> beuno: I see.. but another bug I've reported did not timeout.. should I just retry (just doing so)? - but like forever?
[18:39] <blueyed> now OOPS-1574L2322
[18:40] <blueyed> now it's through.
[18:40] <beuno> blueyed, right, so the server must of been under heavy load when reporting that bug, or the dupe search
[18:41] <keffie_jayx> hey guys, I have a quick question. I founded a team for a proyect and used one of my emails a contact
[18:42] <keffie_jayx> I changed the contact email, but my other email is still associated with the team :S
[18:42] <keffie_jayx> https://launchpad.net/~vento-dev-team
[19:11] <dazwin> Hey folks. I just uploaded a customized version of Vala (GNOME's self-hosting compiler) to my PPA. The build failed because valac could not be found. Any ideas how I get around this cyclical dependency?
[19:11] <flacoste> dazwin: good question
[19:12] <flacoste> dazwin: is valac in universe or main?
[19:13] <dazwin> Let me check - I'm currently using a version from the vala teams own PPA, so not sure
[19:15] <dazwin> main
[19:15] <dazwin> If it helps, the build log is here: http://launchpadlibrarian.net/45116757/buildlog_ubuntu-lucid-i386.vala_0.8.0-0ubuntu1ppa1_FAILEDTOBUILD.txt.gz
[19:24] <flacoste> dazwin: hmm, it should have fullfilled the build dependencies from main
[19:24] <flacoste> let me look at the log
[19:26] <flacoste> dazwin: it's a problem in your packaging probably
[19:27] <flacoste> dazwin: valac isn't in the list of dependencies
[19:27] <dazwin> Except it seems to work Ok for the vala team ;)
[19:27] <dazwin> Yeah, but then does it make sense for a package to Build-Depends on itself?
[19:27] <flacoste> no
[19:27] <flacoste> dazwin: but a self-hosting compiler never relies on a previous version of itself to be available
[19:28] <dazwin> Except vala does!
[19:28] <flacoste> dazwin: that would not be self-hosting
[19:28] <flacoste> dazwin: that package build as is on the Ubuntu archive?
[19:29] <dazwin> According to their hacking page, the normal process is to build with a previous version, then rebuild using the latest compiler. The syntax is pretty stable now, so it seems this doesn't cause too many problems.
[19:29] <flacoste> dazwin: then it should build-depends on itself
[19:29] <flacoste> dazwin: i suggest you ask the vala ubuntu maintainer for help
[19:30] <dazwin> Yep, all I've done is apt-get source from main, add my source code patch, dch, and dput
[19:30] <dazwin> Ok, thanks - will do
[19:30] <flacoste> sorry to not being able to help more, but it's a little outside my dpkg-fu
[19:30] <dazwin> np
[19:31] <flacoste> dazwin: it's weird also that the autoconf doesn't complain from the go that valac isn't available
[19:31] <flacoste> if it's a dependency for building, they should check for it
[19:31] <dazwin> I saw that - maybe a configure.ac problem
[19:31] <flacoste> not sure
[19:31] <flacoste> the fact that the directory is called bootstrap
[19:32] <flacoste> and that a bunch of code is compiled
[19:32] <flacoste> using CC
[19:32] <flacoste> seems to hint that a valac used for bootstraping is being compiled
[19:32] <dazwin> It's there: AC_PATH_PROG(VALAC, valac, valac) - maybe I should try this without having a previous valac installed
[19:32] <flacoste> and that the second pass should use that valac interpreter
[19:33] <flacoste> dazwin: i see checking for valac... valac
[19:33] <flacoste> in the log
[19:33] <flacoste> that's weird
[19:34] <flacoste> so it must have foud it
[19:34] <flacoste> dazwin: i'd suggest trying to build without your patch
[19:34] <flacoste> dazwin: just to make sure
[19:55] <dazwin> Should I be bumping the version number, despite the build failing? (my dput just got rejected)
[20:14] <flacoste> dazwin: probably
[20:16] <dazwin> Thanks flacoste - even though it probably shouldn't need to be bumped, it actually helps to identify the upload as I'm still getting messages about build failures on other platforms.
[20:58] <rephormat> Greetings everyone.