[00:44] <poolie> hi all
[02:43] <poolie> hi spiv
[02:44] <spiv> Hi poolie
[02:44] <poolie> how are you today?
[02:45] <spiv> The cold is lingering more than I'd like, but basically good.
[02:46] <spiv> My kanban board is gradually having more things move to the right, which is nice, and http://webnumbr.com/bzr-in-progress-bugs and http://webnumbr.com/bzr-active-reviews are both trending in a good direction lately :)
[02:47] <poolie> good
[02:47] <poolie> oh, how did you get webnumbr to work for that?
[02:48] <poolie> i tried but failed because the numbers weren't in spans/divs by themselves
[02:48] <spiv> poolie: evil
[02:48] <spiv> poolie: I hand-hacked the xpath expression
[02:48] <poolie> can xpath pull substrings out of strings?
[02:48] <spiv> Oh, actually I don't think I did for those
[02:49] <spiv> But I did need to use https://bugs.launchpad.net/bzr/+bugtarget-portlet-bugfilters-stats as the URl
[02:49] <spiv> Which is also a bit evil...
[02:49] <poolie> ah, that would do nicely
[02:50] <spiv> http://webnumbr.com/bzr-branches-for-ubuntu uses a hand-hacked xpath to get a substring
[02:50] <poolie> ah so you can
[02:50] <poolie> nice
[02:51] <poolie> two things i',m going to do today
[02:51] <poolie> in fact maybe now
[02:51] <spiv> It'd be nicer if webnumber didn't need me to write the xpath myself, but I guess it writes half of it for me, so I could be worse :)
[02:51] <poolie> are to make the kanban script a bit faster
[02:51] <poolie> let it run anonymously
[02:51] <poolie> and then run it from peopel.c.c
[02:51] <poolie> that may be more than two
[02:51] <spiv> Heh, well all of those, whatever the count is, does sound nice.
[02:52] <spiv> In a vaguely related vein, have you seen http://twistedmatrix.com/highscores/?
[02:52] <poolie> i did
[02:52] <poolie> that was neat
[02:52] <poolie> somewhere down my personal-hacking yak list is to fix http keepalive in twisted
[02:52] <poolie> (specifically to try to finish a patch that adds it)
[02:52] <poolie> which will help wrested
[02:52] <poolie> which will help suck stuff out faster
[02:53] <poolie> it's not a strictly necessary depedency
[02:53] <spiv> And might get your name on the highscore list ;)
[03:04] <poolie> maybe on page 4
[03:05] <lifeless> spiv: how easy is it to do those?
[03:05] <lifeless> spiv: I'd like to graph timeout , oops and critical bug counts for lp
[03:06] <lifeless> (on the launchpad-project context)
[03:09] <poolie> lifeless, it's pretty easy if you can hand-hack an xpath expr that will work
[03:09] <spiv> lifeless: webnumbrs?  Login with OpenID, give it a URL, then click on a number on that web page.
[03:09] <poolie> if lp put all numbers into tagged spn
[03:09] <poolie> *spans it would be trivial
[03:10] <spiv> lifeless: mostly it works well, unless the numbers are loaded via js or embedded in a paragraph without a span or something to isolate it.
[03:11] <spiv> lifeless: for critical bug counts you'll probably find https://bugs.launchpad.net/launchpad/+bugtarget-portlet-bugfilters-stats is a good URL to use :/
[03:12] <spiv> lifeless: I'm not sure how you'd feed it timeout and oops numbers, I don't think we publish those publically do we?
[03:13] <lifeless> spiv: I'm not sure if we do either
[03:13] <spiv> lifeless: you could also get the Critical bug count off https://bugs.launchpad.net/launchpad/+bugs?field.importance=Critical&batch=1 ;)
[03:14] <lifeless> ah, we can do tags that way
[03:14] <spiv> (batch=1 because otherwise when count < 75 I don't think the number is explicit on the page)
[03:15] <poolie> lifeless,  so https://code.launchpad.net/~mbp/bzr-email/242262-mail-deadlock/+merge/4430 is ok with you?
[03:16] <lifeless> poolie: last I tried to merge it it was an old format branch needing an upgrade
[03:16] <poolie> i haven't actually seen that bug for ages but it's be good to flush the inventory
[03:17] <poolie> and i think it's an improvement
[03:17] <poolie> oh, that's what you mean by 'update'; ok
[03:20] <lifeless> poolie: 'upgraded' :P but yes. I think.
[03:20] <lifeless> anyhow the diff does look fine
[03:21] <poolie> ok, it's running now
[03:23] <poolie> spiv, i like the 'like' button
[03:24] <poolie> well
[03:24] <poolie> more just i'm interested that it's there
[03:24] <poolie> on the webnumbr pages
[03:24] <poolie> 'like' is a bit blunt
[03:25] <spiv> poolie: I'd never noticed!  I think I've trained myself mentally filter out facebook cruft.  I guess it "likes" the individual webnumbr, rather than the site in general?
[03:25] <poolie> i think so
[03:25] <poolie> bit blunt meaning: does it imply "i like that spiv set this graph up" or "i like bzr having 50 bugs in progress" or "i like there being only 50 rather than 80" etc
[03:26] <spiv> "I like how the graph resembles \m/"
[03:29] <lifeless> spiv: no, the 1 of 245 results returns 1 as the number
[03:31] <spiv> lifeless: you need to hack the xpath expression :/
[03:31] <spiv> lifeless: see e.g. http://webnumbr.com/bzr-branches-for-ubuntu
[03:31] <spiv> substring-after(..., 'of')
[03:32] <poolie> \m/ ?
[03:32] <spiv> poolie: fist of rock!
[03:33] <poolie> heh :)
[03:33] <lifeless> http://webnumbr.com/launchpad-critical-bugs
[03:33] <spiv> lifeless: that's one heck of an xpath  :)
[03:34] <poolie> lifeless, ok, landed
[03:35] <lifeless> http://webnumbr.com/launchpad-timeout-bugs
[03:37] <lifeless> http://webnumbr.com/launchpad-oops-bugs
[03:37] <spiv> ubot5: nice try
[03:41] <lifeless> spiv: thanks for the help
[04:26] <mwhudson> lifeless: you probably want to dial the frequency down from 1 hour
[04:27] <mwhudson> ... if i remember what this site is like
[04:27]  * mwhudson vanishes
[05:49] <spiv> Heh, I've *finally* trained firefox's url bar to suggest https://code.launchpad.net/bzr/+activereviews over https://code.edge.launchpad.net/bzr/+activereviews
[05:56] <poolie> :)
[06:12] <lifeless> \o/
[07:13] <vila> hi all !
[07:14] <poolie> hi vila
[07:14] <spiv> Hey vila
[07:15] <vila> hey hey (hehe) :)
[07:17] <fullermd> yeh yeh?
[07:22] <spiv> Argh!  _pretty_need_write_lock breaks symbol_versioning.deprecated_passed
[07:24] <spiv> Which in turn breaks Repository.search_missing_revision_ids in trunk if you aren't using the fast version of the decorators
[07:28] <vila> breaks how ?
[07:30] <spiv> https://bugs.launchpad.net/bzr/+bug/718569
[07:30] <spiv> vila: see that bug report
[07:33] <vila> eeeek
[07:33] <spiv> Yup.
[07:40] <spiv> I can think of some workarounds, but they aren't very pretty.  Possibly the best approach is to have a dedicated class for those sorts of sentinels that reprs to its python dotted name.
[07:42] <spiv> i.e. "DEPRECATED_PARAMETER = Constant(__name__, 'DEPRECATED_PARAMETER')" or something
[07:42] <spiv> Or 'IdentitySensitiveConstant' or something...
[07:42] <spiv> Anyway!
[07:43] <spiv> What led me there is that I can now reproduce https://bugs.launchpad.net/udd/+bug/653307 locally, with trunk bzr
[07:44] <vila> poolie: regarding bug #718483 , did you requeue the packages today ?
[08:00] <poolie> vila, yes, i did
[08:00] <poolie> spiv, well that's good
[08:00] <vila> poolie: ha, good to know
[08:01] <vila> poolie: bug updated
[08:14] <poolie> good night vila
[08:14] <vila> poolie: g'night
[08:14] <fullermd> I'll try and keep him in line while you're gone...
[08:21] <vila> hehe
[08:32] <dknight> I want to integrate bazaar into enlightenment file manager. Where can I get suggestions about the integration?
[12:08] <spiv> vila: just proposed a fix for bug 718569
[12:09] <vila> spiv: targeting trunk only ? Does this mean we need trunk for p-i ?
[12:09] <spiv> vila: the triggering change is only on trunk
[12:09] <spiv> vila: it's a lurking issue in all the older releases potentially, but so far it hasn't bitten anyone...
[12:10] <vila> bah, forgive me, was mixing up this bug and the missing chk one
[12:10] <spiv> Possibly because no-one uses DEPRECATED_PARAMETER + deprecated_passed in a needs_{read,write}_lock method until I did it recently.
[12:11] <spiv> Anyway, I think I found a tasteful and robust fix.
[12:11] <vila> spiv: approved
[12:12] <spiv> vila: yay!
[12:12]  * spiv sends it off to pqm and heads to bed
[12:12] <vila> spiv: such fixes should be left un-reviewed before the submitter goes to bed or risk of insomnia are far too high ;)
[12:13] <vila> grr. s//not be left/
[12:14] <spiv> Well, perhaps mutating a value from the formatvalue function is a bit sneaky, but otherwise tasteful... ;)
[12:15]  * spiv finds himself increasingly tempted to make a version of http://twistedmatrix.com/highscores/ for bzr...
[12:18] <Tak> there needs to be a penalty for submitting a bad ticket as well
[14:59] <fullermd> Sure, vila's a little odd sometimes, but that doesn't mean we need to rm him, does it?   :p
[14:59] <vila> tsk
[15:00] <fullermd> In other news, I just found a grumpy-making side effect of using reverse merges to undo changes   :|
[15:01] <jelmer> fullermd, what's that?
[15:02] <fullermd> The path you hand merge serves the double purpose of "branch to merge from" and "subtree of branch to merge".
[15:02] <fullermd> So if you're in a subdir and "bzr merge -r5..4 .", it only reverse merges that subdir, not the whole tree.
[15:03] <vila> meh, the path is just the subdir you want to merge from, the branch is inferred no ?
[15:03] <vila> worth documenting but does it deserve a warning ?
[15:04] <fullermd> Well, we really should have a 'bzr reverse-old-rev' command anyway.
[15:04] <jelmer> fullermd, agreed
[15:05] <fullermd> But this means that using merge of '.' isn't just "slightly confusing and unintuintive idiom", it's also "mildly quietly dangerous <previous>"
[15:06] <fullermd> If I hadn't just made the rev I just tried to reverse, and carefully examined the diff beforehand and noticed the output of 'merge' only listed 1 of the 2 files, I'd have never noticed I didn't actually roll back everything I wanted to  :|
[15:07] <fullermd> And if I came that close to missing it, imagine all the other poor people out there nowhere near as brilliant (not to mention handsome) as myself, and the trouble they can wind up in.
[15:08] <vila> > bash.org
[15:09] <vila> bug 718795
[15:16] <mgz> launchpad launchpad launchpad...
[15:17] <vila> desperate and enthusiast ?
[15:17] <vila> s/and/or/
[15:17] <mgz> why are some bug h1 tags now grey for me when they used to all be red...
[15:17] <vila> ...
[15:17] <vila> O_o
[15:18] <mgz> I'd like it if they just had a changelog posted every time they do something, as it is I always wonder if I'm going mad
[15:20] <vila> Decide you're mad, problem solved
[15:21] <jelmer> vila: can we have a bug status for that?
[15:21] <vila> for being mad ?
[15:21] <vila> or for the reporter being mad ?
[15:22] <jelmer> for the reporter, I don't see the value in telling the rest of the world *I'm* mad
[15:23] <fullermd> We're all mad here.  I'm mad.  You're mad.
[15:23] <vila> . o O (Didn't he just said he was seeing no value in telling that ?)
[15:24] <mgz> it's just some of us are madder than others.
[15:24] <vila> Time for a tea ? Where is Alice ?
[15:24] <fullermd> Well, if there's no value in it, obviously you'd have to be mad to mention it.
[15:25]  * jelmer thinks coffee is a great idea
[15:26]  * fullermd perks up.
[15:27] <vila> perks up as in perkolate upwards ?
[15:28] <fullermd> Upward, downward, sideways, inward...   pretty much any direction that gets me coffee.
[15:41] <mgz> bug 330590, meet <http://bugs.gentoo.org/show_bug.cgi?id=343355>
[20:33] <beuno> jam1, FWIW, LH has a different color scheme on LP than on trunk so people can tell them apart
[20:33] <beuno> that was the directive at the time
[20:34] <jam> beuno: I'll post that to the bug. Can you give any more background? Who gave the directive, etc?
[20:34] <jam> also, if you have any idea how hard it would be to make it toggleable, so that LP could easily change the color scheme, that would be great
[20:35] <beuno> jam, yeah, Mark and Kiko at the time. I was never super convinced it was necessary, but, at the time, LP was still closed, so it may not apply anymore
[20:36] <beuno> jam, well, to do theming properly there is some work to be done, at least to split out the images into a separate "theme" dir, not too much work though, I think
[20:38] <jelmer> Hi Jam
[20:39] <jam> hi jelmer
[21:30] <pindonga> hey, I have a question on bzr-builder, is this the right place to ask?
[21:31] <poolie> pindonga, here's fine
[21:31] <pindonga> I see the dput/key-id option pair. However I wanted to ask about the rationale for requiring the key-id... I want to automate the building of source packages, and since I cannot find a way to automate entering the passphrase for the gpg key (which if possible would in some way defeat the purpose of the passphrase) I was suggested not to sign the package
[21:32] <pindonga> so I wanted to add support for this to bzr-builder
[21:32] <pindonga> this is for the dailydeb command
[21:32] <poolie> ok
[21:32] <poolie> i'm not sure but i would think you should make a key specifically for the daily debs, with no passphrase
[21:33] <pindonga> that was my other approach
[21:33] <pindonga> however, I'm not sure what's best tbg
[21:33] <pindonga> tbh
[21:34] <pindonga> having a key without a passphrase is as unsecure as not having a key
[21:36] <pindonga> poolie, I think bzr-builder should not be strongly opinionated on this one, unless there is a good reason
[21:36] <jam> pindonga: could you set up a gpg agent for your build bot?
[21:36] <jam> and that isn't entirely true
[21:37] <jam> one is having a key locally, that validates something you give to other people
[21:37] <jam> vs, not having a way to secure anything with 3rd parties
[21:37] <pindonga> jam, yes, I was a bit dramatic :)
[21:37] <poolie> i tentatively agree it should not be too opinionated
[21:37] <pindonga> jam, what's the thing with the gpg agent?
[21:37] <poolie> however, the signature does prove it came from a machine with the key vs somewhere else in the world
[21:37] <pindonga> poolie, I'll see if I can provide a branch with the change, and we'll see
[21:38] <jam> pindonga: man gpg-agent
[21:38] <poolie> seeing as there's no human present to stamp the package with "i approve this build" that seems to be as strong as you can get
[21:38] <poolie> if you want to make sure no one breaks into that machine that's really out of scope (though worth doing)
[21:38] <jam> basically, you set up a process that knows the key, for the lifetime of a run
[21:38] <jam> so that if you shutdown, restart, the information is lost
[21:38] <jam> you can also configure it to remember only for a specific time, but default of 1 hour doesn't work well with your "run this every day" idea
[21:38] <jelmer> pindonga: supporting unsigned packages seems reasonable to me
[21:39] <pindonga> jam, actually it's run this every merge, but yes
[21:39] <jelmer> pindonga: although its use is limited, what do you plan to do with the source packages built that way?
[21:40] <pindonga> jelmer, let's say I push them to private ppa's
[21:40] <pindonga> since this use case is for building packages from private branches
[21:40] <pindonga> which are not supported by lp's recipes
[21:40] <jelmer> pindonga: How would the private PPA's authenticate the packages that were uploaded?
[21:40] <pindonga> so it would make sense to push those to private ppa's
[21:41] <pindonga> jelmer, well, for starter's I guess you need the user/pass of the ppa to push something to it?
[21:41] <jelmer> pindonga: The transport isn't authenticated at the moment, the GPG key is used for that at the moment.
[21:42] <pindonga> ok, that's true for launchpad
[21:42] <pindonga> however bzr-builder could be used to build packages for debian, right?
[21:42] <jelmer> pindonga: Debian itself doesn't accept unsigned packages
[21:42] <pindonga> in which case (which is not my use case) it may make sense to push unsigned content to the serrs
[21:43] <pindonga> he,
[21:43] <pindonga> bad utopic use case
[21:43] <jelmer> you could use it for a private repository where you build the source packages and the resulting binary packages yourself
[21:43] <jelmer> but even then I think apt-get/aptitude would warn because the packages weren't signed
[21:43] <pindonga> ok, so the best shot here is to use the gpg-agent?
[21:43] <poolie> i don't see the _harm_ in having them signed by a low-trusted key
[21:44] <beuno> jam, re: bug #718978
[21:44] <pindonga> in order to achieve the equivalent of lp's recipes but for private branche?
[21:44] <poolie> just make a key with no passphrase, and don't let people break into the box
[21:44] <poolie> hello beuno
[21:44] <beuno> jam, AFAIK, loggerhead does work with bzr-search if it's installed
[21:44] <beuno> heya poolie!
[21:44] <jam> beuno: sure. but...
[21:44] <jelmer> pindonga: just create a separate GPG key without a passphrase - there shouldn't be a need for gpg-agent in that case
[21:44] <pindonga> poolie, jelmer jam thanks
[21:44] <jam> beuno: launchpad's pqm version of loggerhead had the search box explicitly disabled
[21:45] <jam> presumably a scalability thing
[21:45] <jam> so this got propogated to trunk and 'experimental'
[21:45] <jam> we'd like to bring it back
[21:45] <beuno> jam, ah, right
[21:45] <jam> as a configurable thing
[21:45] <jelmer> also, g'morning beuno, poolie
[21:45] <beuno> morning(?) jelmer
[21:46] <jelmer> beuno: Sorry!  Afternoon?
[21:46] <beuno> jelmer, well, you're still in .nl, right?  so it's evening for you and afternoon for me  :)
[21:47] <vila> beuno: nope, it's morning in europe, you're jet-lagged again
[21:47]  * beuno blinks
[21:48] <beuno> vila, you confused me there for a second  :)
[21:48] <vila> yes !
[21:48] <jelmer> ha :)
[21:48] <beuno> you guys sprinting and confusing me, aren't you?
[21:49] <vila> not even ;)
[21:49] <jam> beuno: vila's just lying to you
[21:49] <jam> and probably still jetlagged or something
[21:49] <jam> he should certainly be in bed with his lovely gf by now
[21:49] <fullermd> Well, _I_ haven't gotten to bed yet.  I don't think he's allowed until I do.
[21:50] <jelmer> fullermd: You sleep?
[21:50] <vila> fullermd: but poolie is here now !
[21:51] <fullermd> Well, I close my eyse when I tpye sometimes to try and impresonate vila.  That's sorta like slepping.
[21:51] <vila> slapping in your sleep are you ?
[21:52] <vila> me ? noticing tyops ? without making one ?
[21:54] <poolie> hi jelmer
[21:55] <poolie> and vila
[21:55] <vila> beuno: truth is, jam moved to europe and is trying to make you think I'm the one lying
[21:55] <vila> poolie: hey ;)
[21:55] <jam> vila: not for 2 more weeks
[21:56] <vila> jam: ;)
[21:56] <vila> jam: just read your mail, welcome !
[21:58] <poolie> vila, thanks for (it seems) trying to update the doc site and web site
[21:58] <poolie> did you get stuck or are you still doing it?
[21:59] <vila> It still is on my radar but couldn't get to it today
[22:02] <jelmer> hi Noldorin
[22:08] <Noldorin> jelmer, hi there.
[22:13] <Noldorin> what's up? :)
[23:41] <jelmer> spiv, hi
[23:41] <jelmer> spiv, I'm seeing a weird issue with search_missing_revision_ids()
[23:42] <jelmer> spiv: revision_id is now deprecated so it's set to symbol_versioning.DEPRECATED_PARAMETER, but symbol_versioning.deprecated_passed() fails, as it's a different string (with the same contents) that's being passed in
[23:45] <spiv> jelmer: I just landed a fix for that (I hope) in lp:bzr
[23:46] <spiv> jelmer: it's a bug in the "pretty" version of needs_read_lock/needs_write_lock
[23:46] <spiv> jelmer: you can workaround it with bzrlib.decorators.use_fast_decorators()
[23:46] <jelmer> spiv: ah, cool :)
[23:47] <spiv> jelmer: bug 718569
[23:48] <spiv> jelmer: I'm a bit amazed this is the first time we've encountered that bug :)
[23:49] <jelmer> spiv: yes, me too :)
[23:52] <jelmer> spiv: Great, I can confirm your branch works :)
[23:53]  * jelmer now has tarmac working for lp:bzr-gtk