[09:55] <lifeless> gmb: hi
[09:55] <lifeless> since you're here...
[09:55] <gmb> Morning lifeless. What can I do for you?
[09:55] <lifeless> your timeout bug for prod
[09:55] <lifeless> is it ready to CP ?
[09:56] <gmb> lifeless, I believe so; let me check...
[09:57] <lifeless> have you seen the hourly oops graph :P
[10:00] <gmb> lifeless, So yes, it's been merged, but into db-devel (I must've chosen the wrong target for the mp), which is annoying. So:
[10:00] <gmb> 1) I need to prepare a branch to merge it into devel
[10:00] <lifeless> yes, yes you do :)
[10:00] <gmb> 2) I need to do the same for production-stable
[10:00] <gmb> (Or whatever our CP target is these days)
[10:01] <lifeless> have you checked on staging too see if its faster ?
[10:01] <lifeless> thats the really interesting thing for me
[10:01] <gmb> lifeless, It's not on staging yet. We'll have to cowboy it there. I'll grab mthaddon in a few minutes and get that done.
[10:02] <lifeless> \o/
[10:08] <gmb> lifeless, Oh, hang on. It *is* in devel. I guess LP gets confused by the auto-merge from devel to db-devel.
[10:08] <gmb> So, that's a win.
[11:56] <jml> review wanted: https://code.edge.launchpad.net/~jml/launchpad/remove-interface-import/+merge/30375
[11:59] <leonardr> jml, i got it
[12:00] <jml> leonardr, thanks.
[12:00]  * jml off to lunch. back soon.
[12:07] <jtv> gmb, leonardr: time for this one?  Just extracting something to a file of its own.  https://code.edge.launchpad.net/~jtv/launchpad/xpi-dtd-parser/+merge/30383
[12:08] <leonardr> jtv, i'm here all day
[12:08] <jtv> leonardr: great, thanks
[12:33] <leonardr> jtv, just to satisfy my curiosity, what do we use the xpi dtd parser for?
[13:06] <leonardr> jtv, r=me
[13:21] <jtv> leonardr: thanks for the review
[13:21] <leonardr> np
[13:55] <rockstar> Anyone around for a review?
[13:55] <rockstar> gmb, leonardr?
[13:56] <lifeless> rockstar: yo
[13:57] <rockstar> lifeless, you want to do a review?
[13:58] <lifeless> I've done 25 or so in the last 24 hours
[13:58] <rockstar> lifeless, https://code.edge.launchpad.net/~rockstar/launchpad/unsupported-series-recipe-api/+merge/30397
[13:58] <lifeless> so why not :P
[13:58] <lifeless> rockstar: check the +activereviews page :)
[13:59] <rockstar> lifeless, that's funny, because I've been thinking of changing the way I do reviews.
[13:59] <lifeless> is 400 the right error to return ?
[14:00] <rockstar> lifeless, that's what everything else returns.
[14:00] <lifeless> oh!
[14:00]  * lifeless adds that to the list of things to look at
[14:00] <rockstar> I don't even know what 400 is.
[14:00] <rockstar> Ah, Bad Request, I think that's right.
[14:00] <lifeless> 'The request could not be understood by the server due to malformed syntax. The client SHOULD NOT repeat the request without modifications'
[14:00] <lifeless> its not malformed syntax
[14:01] <rockstar> lifeless, hm, maybe.  I guess REST does try and re-use HTTP descriptions for itself, so I can see the reasoning there.
[14:02] <lifeless> 409/412 maybbe
[14:02] <lifeless> anyhow
[14:02] <lifeless> you're following the pattern
[14:02] <lifeless> thats fine
[14:03] <rockstar> 412 makes more sense than 409
[14:03] <rockstar> Also, it's funny that if I google for "http 40x" the fourth item returned is xkcd...
[15:02] <jtv> gmb, leonardr: I need a taker for a cleanup branch that's over twice the review limit.  Any volunteers?
[15:02] <jtv> https://code.edge.launchpad.net/~jtv/launchpad/test-cruft-soyuz/+merge/30407
[15:02] <leonardr> jtv, i'll look at it
[15:02] <jtv> Great, thanks.  It's basically lint.
[15:02] <jtv> Just lots of it.
[15:02] <rockstar> jtv, out of curiosity, why didn't you do them in small increments?
[15:03] <rockstar> (I've been thinking about proposing a lower limit to review diff sizes, so I'm trying to find out the cause of outliers)
[15:03] <jtv> rockstar: it went rather faster than I expected—first you remove some boilerplate from the tests, then you run "make lint"...
[15:03] <jtv> ...and suddenly you're at 1,400 lines & growing.
[15:03] <rockstar> jtv, have you used bzr-pipes at all?
[15:03] <jtv> This is rather atypical I hope.
[15:04] <rockstar> jtv, yeah, but I think what sinzui talked about in his lightning talk is probably also the secret to my small amount of success.
[15:04] <jtv> Once or twice, but I was scared off by two things: one was some kind of version skew that broke the plugin for me locally, and the other was that I need to look up how it works every time—and then Aaron tells me what I find in the documentation isn't actually how it should be done.
[15:04] <rockstar> So maybe saying "No branch more than 500 lines of diff" would be helpful for everyone's velocity.
[15:05] <sinzui> That is what U1 did
[15:05] <rockstar> jtv, I'm happy to sit down with you and show you how I do it.  It's a lot easier than one might think.
[15:05] <rockstar> sinzui, really?  That's interesting.
[15:05] <sinzui> barry fixed his productivity by timeboxing.
[15:05] <rockstar> sinzui, I'm curious to know what you think about my idea.
[15:06] <jtv> rockstar: I know—I've done it.  But I don't do it _quite_ frequently enough to know that I won't screw it up next time.  If you could go over the online docs and make sure they're correct, that would be more valuable IMHO
[15:06] <rockstar> jtv, yeah, that's probably a good idea.
[15:06] <jtv> I think the limit of 800 was very helfpul, though sometimes you do just have to go over.
[15:06] <rockstar> jtv, yeah, I'm curious if those "you just have to" situations really could be avoided more if we planned for it.
[15:07] <sinzui> 2/4/6 hours should be enough to produce something valuable. also. I have take 2 hours to make 2000 lines of change moving something to lp.registry. The time was the correct limit
[15:07] <rockstar> sinzui, yeah, although I'd contend that 6 hours might be a little high for most cases.
[15:07] <rockstar> I divide my days up into 3 hour increments.
[15:08] <sinzui> I use 2 hours, and I get nervous when I reach 4. I start looking for ways to end the branch immediately
[15:09] <rockstar> sinzui, yeah, but I've noticed that if I'm doing anything that might touch soyuz stuff, it takes a bit longer, because I don't know all of their gotchas just yet.
[15:09] <rockstar> (I know all the code gotchas, and how to avoid them)
[15:09] <bigjools> rockstar: *we* don't know all of the gotchas yet
[15:10] <sinzui> rockstar, I factor that into getting something ready to code, and I use that time to refactor code so that it is clear. tech-debt amounts to a lot of bugs in Lp.
[15:11]  * bigjools is going to have a 1200 line branch soon.  No way around it.
[15:11] <rockstar> sinzui, huh.  That's interesting as well.  A prep branch to go before the actual branch.
[15:12] <sinzui> Salgado does a lot of them. He seems to use it to make the real task easier to do, or at least to remove any distractions from his primary task
[15:14] <rockstar> sinzui, yeah, it's the distraction thing that annoys me.
[15:14] <rockstar> sinzui, I've been performing an experiment with my branch workflow, and my biggest patch is ~450 lines, so I think we really can break them up more.
[15:15] <leonardr> jtv, does lint really complain that there should be a space in ('foo',) ?
[15:15] <leonardr> i never knew that
[15:15] <jtv> leonardr: yup... I discussed with sinzui who confirmed that PEP8 wants it that way.
[15:15] <jtv> (it surprised me as well :-)
[15:15] <sinzui> Maybe this is a conversation for the review's meeting. smaller branches or shorter development time.
[15:15] <rockstar> jtv, I've ignored that (it's stupid...)
[15:16] <sinzui> rockstar, you should submit a patch to pep8.py them
[15:16] <rockstar> sinzui, yeah, before I brought it to the reviewer's meeting, I wanted my own experiment.
[15:16] <sinzui> then
[15:17] <rockstar> Usually, the reviewer's meeting can turn into yak shaving/bikeshedding, so I wanted actual facts (similar to your lightning talk)
[15:27] <leonardr> jtv, the blank line on line 627 of your patch looks strange. is it correct?
[15:28] <jtv> leonardr: looking...
[15:29] <leonardr> jtv: ditto for line 811
[15:29] <jtv> leonardr: yeah... those are how "make lint" wants them.
[15:30] <leonardr> jtv: blank line before any function definition?
[15:30] <jtv> Yes, looks like.
[15:32] <leonardr> jtv, r=me
[15:32] <jtv> leonardr: thanks!  I'll try to bother gmb for the next one, because this just ain't fair.  :-)
[15:33] <gmb> Yes.
[15:47] <leonardr> jtv, for the record i got some test failures when i ran the soyuz tests, but they kind of look like a problem with my larger setup. i can paste them if you want
[15:48] <jtv> leonardr: if they're MissingCelebrityErrors, that's a devel change and "make schema" will help.
[15:48] <leonardr> jtv: no, it started as a GpgmeError ("unknown error code") and then cascaded from there
[15:49] <jtv> leonardr: I think I had those a bit longer ago, and rocketfuel-{setup,get} or something made it go away.
[15:49] <jtv> Do you get them in devel?
[15:49] <leonardr> jtv: i don't know. sorry--i'm fighting another fire and don't want to context-switch unless this is really important to you
[15:50] <jtv> leonardr: not particularly—ec2 will be my judge
[15:50] <leonardr> ok
[15:50] <jtv> gmb: here's one I'd like to ask your review for: https://code.edge.launchpad.net/~jtv/launchpad/exportstorage-cleanup/+merge/30418
[15:52] <gmb> jtv, Righto
[15:52] <jtv> fab
[15:56] <gmb> jtv, Is the VWS on line 61 really necessary?
[15:56] <jtv> gmb: make lint says so
[15:59] <gmb> jtv, Garh.
[15:59] <gmb> Okay.
[15:59] <gmb> I'm not wholly sure the new linter isn't a bit crackful.
[15:59] <jtv> It's a lot stricter—which by and large is a good thing, I think
[16:00] <gmb> jtv, True.
[16:00] <gmb> jtv, Anyway, r=me.\
[16:00] <jtv> \o/ thanks
[19:15] <jtv> leonardr: it's still highly unfair that your co-reviewer buggered off before I could be as demanding with him as I was with you...  got room for a few more?
[19:15] <jtv> I have one small lint cleanup: https://code.edge.launchpad.net/~jtv/launchpad/pre-export-cleanup/+merge/30445
[19:17] <leonardr> jtv, i see a conflict in potemplate.py
[19:17] <jtv> leonardr: thanks...  I'll update my branch & re-push.  Hang on.
[19:22] <jtv> leonardr: to keep you updated: I pushed a while ago but am still polling for the diff to update.  Will let you know.
[19:23] <jtv> leonardr: it's updated.
[19:28] <leonardr> jtv, r=me
[19:28] <jtv> leonardr: cool, thanks
[19:28]  * leonardr will be back in a couple hours in case no one else wants to do reviews
[19:29] <jtv> I'll ask around.  Thanks for all the reviews today!