[00:56] <mtaylor> a private team can apparently not review merge-proposals
[02:02] <wgrant> I hate doctests.
[02:03] <spm> 3 words. so much ... emotion unsaid yet understood.
[02:05] <wgrant> See, I want to fix a trivial ordering edge case (bug #604842). But the only tests for the method are doctests where I can't really fit a retry. If only they were unit tests :(
[02:05] <_mup_> Bug #604842: BinaryPackageBuildSet.getBuildsByArchIds orders inconsistently <Soyuz:New> <https://launchpad.net/bugs/604842>
[04:19] <mtaylor> what do I need to do to make a branch private?
[04:19] <mtaylor> sorry - wrong channel
[04:30] <lifeless_> wgrant: change em to unit tests ;)
[04:31] <wgrant> lifeless_: I was going to ask if it was impolite to bloat a one-line branch into a few hundred lines by rewriting tests, but I presumed everyone would be asleep.
[04:32] <lifeless> wgrant: I'd spin a new branch for it
[04:33] <wgrant> lifeless: One branch to redo the tests, and a followup with the one-line change plus new test? Perhaps.
[04:44] <lifeless> yes
[04:44] <lifeless> its a little more work but - JFDI
[04:44] <lifeless> its not a lot more
[04:44] <wgrant> Yep.
[04:44] <wgrant> I replaced several hundred lines of doctests yesterday. It seems to be becoming a bit common.
[04:45] <wgrant> Nasty things.
[05:23] <lifeless> the meme has caught on
[05:23] <wgrant> Oh?
[05:23] <lifeless> and muhaha, its now my job to start memes ;)
[05:23] <lifeless> wgrant: that doctests are actively harmful
[05:23] <wgrant> The anti-doctest meme?
[05:23] <wgrant> Right.
[05:23] <wgrant> I now hate them with a passion.
[05:23] <lifeless> \o/
[05:23] <wgrant> A few months ago I merely mildly disliked them.
[06:23]  * cody-somerville has slowly been seeding an SOA meme.
[06:24] <wgrant> Has it caught on yet?
[06:24] <cody-somerville> Seem to be
[06:24] <wgrant> Excellent.
[06:24] <cody-somerville> Although I think lifeless might extinguish my efforts :/
[06:24] <wgrant> :(
[06:24] <wgrant> Why?
[06:25] <cody-somerville> I remember briefly skimming minutes for a meeting I missed that noted that lifeless thought splitting launchpad up into smaller, discrete services was a distraction.
[06:26] <wgrant> It is a distraction. But probably a beneficial distraction.
[06:26] <cody-somerville> Indeed. Which is why I hold out hope that the comment isn't as fatal sounding as it does.
[06:39] <mtaylor> cody-somerville: fwiw, I support your interest in splitting
[06:39] <cody-somerville> :)
[06:40] <mtaylor> of course, I have ZAROO vote
[06:40] <mtaylor> but smaller discrete units is pretty much a best practice, especially when you want to scale (hi facebook)
[06:41] <cody-somerville> or hello amazon
[06:42] <mtaylor> or really anyone
[07:06] <thumper> mtaylor: morning
[07:06] <thumper> mtaylor: how's your summit going?
[07:27] <spm> thumper: we have a project that has been renamed. we're trying to reset the lp:<project> alias; but that failed. now, whenever we try and re-create 'trunk' we keep getting 'you are being stacked on <junk branch as placeholder>' which is painful. suggestions?
[07:28] <thumper> hmm...
[07:28] <thumper> I'm just relocating, back in a few minutes
[07:28] <thumper> pastebin more details please
[07:28] <spm> kk
[07:57] <poolie> rockstar, https://dev.launchpad.net/LEP/DynamicConfiguration
[07:58] <rockstar> poolie, thanks
[08:43] <mwhudson> bigjools: where did you go?
[08:45] <bigjools> mwhudson: haydn
[08:45] <mwhudson> bigjools: want to talk about buildd-manager now, or busy with other things?
[08:45] <bigjools> we're talking about derived distros, might be useful for you?
[08:45] <mwhudson> ah sure
[08:45] <mwhudson> be along in a sec
[08:45] <bigjools> but yes I'd like to talk about that too :)
[08:46] <mwhudson> bigjools.fork()
[09:56] <wgrant> jml: We still can't renew our memberships. Maybe you could reset everyone to not expire?
[09:56] <wgrant> (you just changed the default for new members, it appears)
[10:18] <deryck> gmb, https://code.edge.launchpad.net/~deryck/launchpad/move-bugs-js-files-again/+merge/29776
[10:29] <jml> wgrant, there isn't a control for that
[10:30] <wgrant> jml: It's not exposed through the API?
[10:30] <jml> wgrant, I don't know.
[11:09] <poolie> hi wgrant
[11:09] <wgrant> Hi poolie.
[11:36] <poolie> benji, leonardr: https://bugs.edge.launchpad.net/launchpadlib/+bug/604963 for this issue
[11:36] <_mup_> Bug #604963: shouldn't need to close window and press enter after authenticating <launchpadlib :New> <https://launchpad.net/bugs/604963>
[11:38] <jml> lifeless, noodles775: btw, if you want to see some tests that use fake objects rather than the database, lp.code.xmlrpc.tests.test_codehosting (tests for the fake & real implementations); lp.codehosting.vfs.tests (tests that use the fake); lp.codehosting.inmemory (the fake itself)
[11:39] <jml> running the test_codehosting tests makes the database overhead for testing really obvious
[11:40] <poolie> leonardr, http://sourcefrog.net/tmp/1y/closeme.html
[11:42] <noodles775> jml: will do, thanks.
[14:21] <jtv> jam: hi!  Partial work on dropping BranchRevision.id: http://paste.ubuntu.com/462977/  Still failing these tests or a subset thereof: http://paste.ubuntu.com/462978/
[14:21] <jam> jtv: thanks for the headsup
[16:40] <leonardr> mars, src/lazr/restful/tales.py:from epydoc.markup import DocstringLinker
[19:32] <mtaylor> morning all
[20:29] <lifeless> mtaylor: hai
[20:29] <lifeless> mtaylor: I'd like to suggest a warm up task for you, in lp stuff,if I may.
[20:32] <mtaylor> lifeless: yes!
[20:32] <lifeless> make it fast
[20:33] <lifeless> can you see http://people.canonical.com/~stub/ppr_test.html?
[20:33] <mtaylor> lifeless: ok
[20:33] <thumper> hi mtaylor
[20:33] <mtaylor> lifeless: I would like lp to be fast
[20:33] <mtaylor> hi thumper
[20:33] <mtaylor> thumper: did you see the bug I just filed? :)
[20:33] <thumper> um...
[20:33] <thumper> which one?
[20:34] <mtaylor> bug#605130
[20:34] <_mup_> Bug #605130: subscribed team cannot view private branch <Launchpad itself:New> <https://launchpad.net/bugs/605130>
[20:34] <thumper> mtaylor: probably because the branch is owned by a private team
[20:34] <mtaylor> thumper: probably
[20:34] <mtaylor> thumper: but bzr branching it works
[20:35] <thumper> heh...
[20:35] <mtaylor> thumper: I probably won't care anymore by the time you have time for it - but I thought I'd point it out :)
[20:35] <lifeless> so
[20:35] <lifeless> this is the same as PPPA access
[20:35] <thumper> that's because the code path for accessing the branch over bzr doesn't look at the team
[20:35] <lifeless> we're going to be adding a 'Traversal' permission.
[20:36] <lifeless> which will be used to make sure such pages can render reasonably. granting read access to something with a url path containing private things will imply access to traverse them
[20:36] <lifeless> and traversing means 'show the url and a human name for the thing'
[21:06] <lifeless> mtaylor: so, can you see that page?
[21:23] <mtaylor> lifeless: yup
[21:24] <lifeless> cool
[21:25] <lifeless> mtaylor: and see how its got blueprint bits in it ?
[21:26] <lifeless> mtaylor: and see how 99% of blueprint pages complete within 200 sql statements ?!??!1
[21:26] <lifeless> and 3.2 seconds ?
[21:26] <lifeless> thats pretty slow ;)
[21:27] <lifeless> mtaylor: and with that, gnight
[21:27] <mtaylor> lifeless: mk
[21:28] <mtaylor> lifeless: I will fix
[21:28] <lifeless> mtaylor: mean is 50 sql statements, which is still pretty terrible.
[21:28] <lifeless> mtaylor: yeah, if its fixed, be in a good position to add features.
[21:28] <lifeless> gnight.