[09:55] <frankban> hi gmb, could you please take a look at https://code.launchpad.net/~frankban/launchpad/bug-1014907/+merge/110987 ?
[09:55] <gmb> frankban, I'll look shortly
[10:54] <frankban> thanks gmb
[12:04] <gmb> gary_poster, My 12:30 just showed up, so I'll likely not make standup.
[12:04]  * gmb -> shooting people
[12:04] <gary_poster> gmb, heh
[12:04] <gary_poster> ok ttyl
[12:08] <gary_poster> bac benji frankban (gmb not here): https://plus.google.com/hangouts/_/e7129ba442a5def36f39fb51ce8543146f1cdad1?authuser=1&hl=en-US in 2
[12:15] <gary_poster> gmb, your card is 24 hours old, so pls ask for pair when you return
[12:15] <gary_poster> bac's is too but he's not here
[12:17]  * bac is trying
[13:16] <gmb> gary_poster, The card is only just 24 hours old (I moved it to WIP during standup yesterday, didn't start on it until after that ) and I'm making good progress. I'm happy to hang fire on pairing for now, but will re-evaluate in an hour or so.
[13:18] <gary_poster> gmb, cool.  if it is not to review/landing by your EoD, please make sure you hand it off before your EoD, since it is blocking another card.  (and pair before then if it makes sense, yes, thanks!)
[13:18] <gmb> gary_poster, Of course, will do.
[13:18] <gary_poster> cool ty
[13:57] <bac> hi gmb -- i just ran bin/test and unexpectedly got the following:
[13:57] <bac> ...ome/bac/launchpad/lp-branches/1012171> bin/test -vvt TestWebServiceErrors
[13:57] <bac>         You specified a module along with --require-unique;
[13:57] <bac>         --require-unique will not try to enforce test ID uniqueness when
[13:57] <bac>         working with a specific module.
[13:58] <bac> i didn't specify a module using -m
[13:58] <gmb> bac, Um. Well, that's... unexpected.
[13:59] <gmb> bac, If you poke around in zope.testing.testrunner.options you'll find the check for if options.module and options.require_unique_ids near the bottom; you could poke a pdb in there and see what's going on.
[13:59] <bac> gmb: ok, will do.
[14:16] <gmb> benji, gary_poster: Do either of you know if - ahaha - mess about with sys.stdout?
[14:16] <gmb> Um
[14:16] <gmb> I mean "doctests - ahaha -"
[14:16] <gary_poster> lol
[14:16] <gary_poster> it was mysterious for a moment there
[14:16] <gary_poster> I don't offhand.  benji might
[14:17] <gary_poster> oh wait a sec
[14:17] <gary_poster> yeah. I think doctests generically mess with stdout
[14:17] <gary_poster> I thought you meant a specific doctest
[14:17] <gary_poster> but yeah
[14:17] <benji> yes, doctests capture stdout (that's how they know what the "actual" bit is to compare with the "expected" bit)
[14:17] <gary_poster> right ^^^
[14:20] <gmb> Right.
[14:21] <gmb> benji, So, the problem in some of the subunit doctests in zope.testing is that they appear not to see the subunit output. I'm guessing that that's because we force subunit to __stdout__ and doctest is looking at something else. Is that sane-sounding?
[14:21] <benji> gmb: yep, that's a good hypothesis
[14:23] <gmb> Haha.
[14:23] <gmb> benji, gary_poster:
[14:23] <gmb> class _SpoofOut(StringIO):
[14:23] <gmb> ...
[14:23] <gmb> This is going to be fun...
[14:23] <gary_poster> tunder stdout!
[14:26] <gmb> if not streams_munged and sys.stdout != sys.__stdout__:
[14:26] <gmb>     stab_stab_stab_stab = True
[14:47] <gary_poster> benji, could you take a glance at the "Unnecessary cronscripts?" thread and reply to flacoste's comment about the rosetta cronscript please?
[14:47] <benji> gary_poster: sure
[14:51] <gary_poster> ty
[14:58] <bac> here's what the stdout/stderr subunit output look like: http://paste.ubuntu.com/1049226/
[15:26] <gmb> gary_poster, Is my card blocking because it's in the WIP lane or because something depends on it?
[15:27] <gmb> Because with this docstring problem we can pretty much say that we're back to square one for now (my previous good progress turned out to be a mirage)
[15:27] <gary_poster> gmb, because something depends on it (I don't want to start bug 1013921 until zope.testing is back in reasonably good order)
[15:27] <_mup_> Bug #1013921: our zope.testing fork needs to emit subunit time immediately before test start and immediately before test completion <paralleltest> <Launchpad itself:Triaged> < https://launchpad.net/bugs/1013921 >
[15:28] <gmb> gary_poster, Okay. Give me a little while, and I can pair / hand-off with A.N.Other on what I know so far ( which is not a huge amount, except that global state hates me)
[15:28] <gary_poster> gmb, pair with someone then, maybe an American so you can hand off?  I could be ready if Liam can't tell me where to find prune.py
[15:28] <gary_poster> or benji might be ready soon
[15:28] <gmb> Okay.
[15:28] <gmb> Makes sense.
[15:28] <benji> yep
[15:41] <benji> gmb: I'm ready for talking if you want.  If not I'll take lunch and we can talk after.
[15:54] <gmb> benji, I'm ready now, let me just get a hangout started.
[15:57] <gmb> benji, https://plus.google.com/hangouts/_/d154498843db24fb425588dbc74afba32c57518d?authuser=0&hl=en-GB#
[16:02] <bac> benji, gary_poster: i'd like to pair up with either of you after lunch, too.  i'm about done but have gotten my knickers twisted wrt getting data once/only once from stdout/stderr StringIO objects
[16:03] <gary_poster> bac, cool
[16:04] <bac> gary_poster: will you be available later?
[16:04] <bac> if so, i'll eat now
[16:04] <gary_poster> bac, yes.  1:15ish?
[16:04] <bac> gary_poster: ok.
[16:13] <gmb> benji, lost you...
[16:13] <benji> gmb: trying to reconnect
[16:14] <benji> google really hates me of late
[16:15] <benji> gmb: no hangout (or termbeamer because it uses google talk) for me
[16:15] <benji> I'm starting to wonder if google has decided that I am a bad person and should be shunned.
[16:16] <gmb> Heh.
[16:17] <gmb> benji, Okay, let meeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
[16:17] <gmb> Hmm
[16:17] <gmb> Let me poke around those test and get back to you shortly. They do... interesting things.
[16:17] <gmb> (Also, I need fluids)
[16:45] <gmb> benji, So, I've not managed to get any further, and I'm about to be interrupted by mangerial types. If I push what we've got so far, will you be okay to take it from there? There are still six tests failing.
[16:45] <gmb> (Some of those might be subunit related)
[16:47] <gmb> gary_poster, around?
[16:51]  * gmb emails the team
[17:10] <gary_poster> gmb, sorry was lunching.  saw your email.  Hopefully one of us can run with it today.  Have a nice evening
[17:18] <bac> hi gary_poster, i just pushed a branch up to lp:~/zope.testing/1012171 -- let me know when you'd like to hangout
[17:23] <gary_poster> bac, cool, now is good.
[17:23] <gary_poster> you starting or shall I?
[17:23] <gary_poster> hangout I mean
[17:23] <bac> i will
[17:23] <gary_poster> thx
[17:24] <bac> gary_poster: https://plus.google.com/hangouts/_/c2ff86710d20766821883a7b65ce7b4663d43092?authuser=0&hl=en
[17:29] <benji> gmb: I was lunching.  I'll take a look.
[18:45] <gary_poster> banji, bac, hallyn's workaround for bug 1014916 not only makes machines start all about the same time, but as an expected/hoped corrollary, slices around three minutes off our test run time.
[18:45] <_mup_> Bug #1014916: simultaneously started lucid containers pause while starting after the first seven <lxc (Ubuntu):Confirmed> < https://launchpad.net/bugs/1014916 >
[18:45] <gary_poster> yay!
[18:45] <bac> nice
[18:46] <bac> go hallyn
[18:47] <gary_poster> With that, the differeence between the first and last worker stopping is more like 5 or 6 minutes, at least in my one sample so far
[18:47] <gary_poster> Which means that the round robin won't be quite as profitable, but the combination of that workaround and the non-roundrobin should still be significant
[19:31] <bac> gary_poster, benji: if either of you have time for a review: https://code.launchpad.net/~bac/zope.testing/1012171/+merge/111093
[19:31] <gary_poster> on call will check after
[19:33] <benji> bac: I'll take a look.
[19:33] <bac> benji: thanks
[19:33] <benji> bac: how many tests are still failing on your branch?
[19:34] <bac> benji: 7
[19:34] <bac> benji: i fixed one, the 'latin1' in test_subunit
[19:34] <benji> hmm, I would have thought it would be lower.
[19:35] <bac> it was trivial, just a \n that got moved in the output
[19:35] <bac> why would you think that?  my stuff shouldn't have affected those other failures
[19:37] <benji> I misunderstood part of the change, but capturing stdout may help in fixing the other failures I'm working on right now.
[19:39] <benji> bac: looks good
[19:41] <bac> benji: thanks
[19:44] <bac> hurrah, the NC house killed the sea-level rise bill.
[20:04] <gary_poster> bac, wow, something good actually happened in the legislature!  or at least non-bad.
[20:04] <gary_poster> benji, how goes test failure fun?
[20:04] <benji> gary_poster: pretty good; down to 4
[20:05] <gary_poster> benji, great!  what were we before?  /me goes to look
[20:05] <gary_poster> 3
[20:05] <benji> gary_poster: more ;)
[20:05] <benji> oh, you meant before we made them go up
[20:05] <gary_poster> :-) yeah
[20:06] <gary_poster> http://pastebin.ubuntu.com/913757/.
[20:12] <benji> down to the original 3 failures.  I want to think about this a little though, because it is a bit hacky (but not obscene)
[20:26] <gary_poster> :-)
[20:26] <gary_poster> but good news nonetheless
[20:33]  * bac -> optometrist.  \0/
[20:36] <gary_poster> bac :-) k have a good evening