[08:35] <frankban> good morning gmb
[08:42] <gmb> Morning frankban
[09:11] <frankban> gmb: could you please review https://code.launchpad.net/~frankban/launchpad/bug-1012799/+merge/110270 ?
[11:09] <frankban> gmb: you there?
[11:10] <gmb> frankban: Sort of. In a cafe, suck Internet. What's up?
[11:10] <frankban> gmb: could you please take a look at https://code.launchpad.net/~frankban/launchpad/bug-1012799/+merge/110270 ?
[11:11] <gmb> frankban: I will do shortly, sure.
[11:11] <frankban> thanks
[11:46] <gmb> frankban: r=me, but I can't log in to lp at the moment; think there's a proxy in the way. Feel free to self review and note that I gave it my approval.
[11:47] <frankban> ok, thanks gmb
[11:57] <bac> hi gmb, is anyone reviewing your branch yet?
[11:58] <gmb> Bac nope, feel free. :)
[11:58] <bac> gmb: have you run it against trunk since my change landed?
[11:58] <gmb> No.
[11:58] <bac> would be nice to know if i got them all
[12:01] <gmb> bac: I'll spin up a slave this afternoon for it.
[12:02] <bac> gmb: does your implementation have the limitations we discussed yesterday, such as '-m' doing pre-filtering so that not all tests are checked?
[12:03] <gmb> bac: If I understand correctly, it sits before filtering takes place, but that might only be for -t; I'll check.
[12:04] <bac> gmb: if that is the case, do you think we should warn when imcompatible options are chosen?
[12:05] <gmb> bac: yes, that sounds sane.
[12:05] <gmb> gary_poster: Might be late to standup; wifi is crummy here.
[12:06] <gmb> Am going to switch locations.
[12:06] <gary_poster> ack gmb
[12:09] <gary_poster> bac benji frankban gmb call in 2 https://plus.google.com/hangouts/_/dc2755d503ade939d419e49991619c2a6a3ba408?authuser=1&hl=en-US
[12:34] <gmb> gary_poster: G+ crashed. Nothing to add beyond longish lunch.
[12:34] <gary_poster> gmb, cool :-)
[12:34] <bac> gmb: good news, on a version of trunk before my changes, --require-unique correctly flags duplicates.  on trunk it finds none.
[12:34] <gmb> Woot.
[12:35] <gary_poster> benji, may I move your call to 2:00 PM?
[12:35] <bac> gmb: when the first dupe is found it raises an exception and quits.  did you consider making it find and report all dupes?
[12:35] <benji> gary_poster: sure
[12:35] <gary_poster> thank you
[12:36] <bac> gmb: i'd like to suggest that unless you rejected it for good reason
[12:36] <gmb> bac: hmm, no. But that wouldn't be too hard; I'll do that.
[12:36] <bac> yay
[12:36] <gmb> bac: no, I just like to prolong developers' pain.
[12:36] <bac> gmb: and finally, do you have a companion LP branch that incorporates that flag into one of the makefile targets?
[12:37] <gmb> bac: yes but I can't push it yet due to proxy issues
[12:37] <gary_poster> bac, which would you prefer, 1:30-2:00 or 4:30-5:00?
[12:37] <bac> 1:30
[12:37] <gary_poster> cool
[12:37] <gmb> bac: Will be doing that this afternoon.
[12:37] <bac> gmb: great
[12:37]  * gmb -> bureaucracy
[12:44] <bac> gmb: review done
[13:02] <gary_poster> oh poo.  I forgot to have the panel on bug 994752.  I'll add it to tomorrow's list
[13:02] <_mup_> Bug #994752: lxc-start-ephemeral's use of dhcp lease table is fragile <patch> <verification-done> <lxc (Ubuntu):Fix Released> <lxc (Ubuntu Precise):Fix Released by stgraber> <lxc (Ubuntu Quantal):Fix Released> < https://launchpad.net/bugs/994752 >
[13:08] <gary_poster> benji, are you working on the EBS snapshot thing?  If so, I will drag your card over to Active: Coding.  Also has anyone stepped up to bug 1012171 yet to your knowledge?
[13:08] <_mup_> Bug #1012171: Make captured stdout and stderr available within the subunit stream <paralleltest> <Launchpad itself:Triaged> < https://launchpad.net/bugs/1012171 >
[13:09] <benji> gary_poster: not quite yet; I'm trying to figure out the right way to set the ignores on my re-add-logs-directory branch; re. 1012171: not yet
[13:09] <gary_poster> benji ok.  Do you want me to make a bug + card for that thing you are working on?
[13:10] <benji> gary_poster: sure (or we can reuse the last bug, either way is fine with me)
[13:11] <gary_poster> benji I assume you mean bug 996729?  if so, ugh, I want that one to be done!  is it the same cause though?
[13:11] <_mup_> Bug #996729: zope.testing --subunit allows bad output on stdout, which can break subunit processing and tests <paralleltest> <qa-untestable> <Launchpad itself:Fix Committed by gmb> < https://launchpad.net/bugs/996729 >
[13:12] <benji> gary_poster: no, not 996729; let me find the bug number
[13:12] <benji> gary_poster: bug 1011793
[13:13] <_mup_> Bug #1011793: lib/lp/services/profile/profiling.txt fails rarely/intermittently in parallel tests <paralleltest> <qa-ok> <Launchpad itself:Fix Released by benji> < https://launchpad.net/bugs/1011793 >
[13:18] <gary_poster> ack
[13:19] <gary_poster> benji I decided just to make a card.  It's in Active Coding with your face on it
[13:19] <benji> gary_poster: my face appreciates it
[13:19] <gary_poster> :-)
[13:32] <gary_poster> frankban, hey.  https://plus.google.com/hangouts/_/d8192b69d811f42c71018b1e307a9fa8ae1a1333?authuser=1&hl=en-US when you are ready
[14:16] <gmb> bac, Just looking at adding an error after all the tests have been scanned... it would mean adding the error handling outside the generator, which in turn would mean looping running through the generated results... Obviously this is going to be less performant than the current incarnation, but I'm not sure how much we need care about that.
[14:16] <gmb> Any opinions?
[14:22] <bac> hi gmb.  let me look
[14:31] <bac> gmb: what if tests_from_suites had another parameter that was a list of duplicates found.  tests_from_suites could populate that list without complaining and then find_tests could warn about the whole lot of them if any exists.
[14:32] <gmb> bac, Right, that's what I've done, but tests_from_suites is a generator and is recursive, so just doing something like:
[14:32] <gmb> tests = tests_from_suites(..., duplicate_ids=dupes)
[14:32] <gmb> won't have dupes fully populated
[14:32] <gmb> Until someone loops over `tests`
[14:32] <gmb> Or at least, I don't think it will.
[14:33] <gmb> My tests might be too shallow to prove one way or the other;
[14:34] <bac> gmb: but it looks like find_tests does loop over the results of tests_from_suites.
[14:34] <bac>         for test, layer_name in tests_from_suite(suite, options):
[14:34] <gmb> bac, Right, but then it does other stuff that we don't want to care about if there are duplicates.
[14:34] <gmb> Although I suppose pragmatically it doesn't actually matter.
[14:34] <bac> ah
[14:35] <bac> gmb: i'd call my request a 'nice to have it not too complicated'
[14:36] <gmb> bac, Right, I think I can see a way to do it... and actually, this being zope.testing,  I don't think "Ooh, that's a bit ugly" is actually a valid complaint.
[14:36] <gmb> It's not exactly going to win any pageants.
[14:37] <gmb> Thanks :)
[14:49] <bac> gmb: sorry, i couldn't resist having a go at it.
[14:49] <bac> http://paste.ubuntu.com/1040891/
[14:49] <gmb> bac, That's exactly what I've done :)
[14:49] <bac> excellent
[15:02] <gary_poster> gmb, sorry, went over with francesco.  ready at https://plus.google.com/hangouts/_/d8192b69d811f42c71018b1e307a9fa8ae1a1333?authuser=1&hl=en-US ?
[15:02] <gmb> gary_poster, Okay, will be with you in a minute or two.
[15:03] <gary_poster> cool gmb
[16:54] <gmb> gary_poster, lp:~gmb/launchpad/devel/.bzr/branches/zope.testing-p13 has an up-to-date requirement on my zope.testing patch, and is based on the latest devel.
[16:55] <gmb> If you coudl run that in a buildbot of some sort, I'd be grateful
[16:57] <gary_poster> on call and lunch, will do soon
[16:57] <gmb> gary_poster, No huge rush; I EOD in 3 :)
[16:58] <gary_poster> :-0
[16:59] <gary_poster> :-) do you want me to land if it passes, gmb?
[17:01]  * bac reboots
[17:03] <gmb> gary_poster, Please; I'll be landing the zope.testing branch shortly.
[17:03] <gary_poster> great
[17:13] <gary_poster> gmb, do you have an MP for that already?  https://code.launchpad.net/~gmb/launchpad/devel/.bzr/branches/zope.testing-p13 doesn't take me anywhere...
[17:14] <gary_poster> gmb maybe you meant lp:~gmb/launchpad/zope.testing-p13 ?
[17:14] <gmb> gary_poster, Yes, I did. Damn colo branches
[17:14] <gmb> I'll do the mp now.
[17:14] <gary_poster> thank you
[17:15] <gmb> gary_poster, Ah, wait, hang on... I haven't updated the make target to use the new switch. Durr.
[17:15] <gmb> So, which make target needs updating?
[17:16] <gary_poster> gmb, I'd expect that you'd want to change buildout-templates/bin/test.in
[17:16] <gmb> Aah.
[17:17]  * gmb looks.
[17:18] <gary_poster> gmb, the defaults dict maybe?
[17:18] <gmb> gary_poster, Something like:
[17:18] <gmb> require_unique_ids: True,
[17:18] <gmb> ?
[17:18] <gmb> Er. With quotes
[17:19] <gary_poster> gmb, maybe.  you got me :-P
[17:19] <gmb> Heh.
[17:20] <gmb> Well, let's see what happens....
[17:21] <gmb> gary_poster, Yep, that seems to work.
[17:21] <gmb> I'll update it.
[17:21] <gary_poster> awesome
[17:31] <bac> gary_poster: chat?
[17:31] <gary_poster> bac, yes was making hangout https://plus.google.com/hangouts/_/5cfec4c6014cb96f4037ea07d4f0744fa15d3e92?authuser=1&hl=en-US
[19:59] <bac> hey benji, i'm looking at bug 1012171 and trying to figure it out.  you have time for a quick call?
[19:59] <_mup_> Bug #1012171: Make captured stdout and stderr available within the subunit stream <paralleltest> <Launchpad itself:Triaged> < https://launchpad.net/bugs/1012171 >
[19:59] <benji> bac: sure
[20:00] <bac> cool, let me grab some tea and set up a hangout...
[20:00] <benji> k
[20:05] <bac> benji: i invited you but the window doesn't show me the url
[20:05] <benji> bac: I got the invitation, one sec
[20:15] <bac> benji: it looks like it is a dict with a testtools.content.Content object as the value
[20:16] <benji> hrm, it's a shame but you'll have to complicate things a little; I hope those are easy to build
[20:17] <bac> should be
[20:36] <gary_poster> We have 12 in a row green.  The last time we had that--the only other time we had that--was May 4-8
[20:37] <benji> yay!
[20:37] <benji> (note use of exclamation mark)
[20:37] <gary_poster> lol
[20:37] <gary_poster> Also for the last three days
[20:37] <gary_poster> we have a statistic of 75%
[20:38] <gary_poster> green
[20:38] <gary_poster> which is pretty good
[20:38] <benji> what was our goal again?  95%?
[20:39] <gary_poster> yeah
[20:39] <benji> so we need 7 more to be 1 failure in 20
[20:39] <benji> quick, spin up 7 instances
[20:39] <gary_poster> :-) We are not supposed to switch to stretch goals till 90% but we don't have any other known bugs to work on atm
[20:48] <bac> benji: shouldn't the zope.testing subunit tests pass, i.e.  bin/test -t subunit  ?
[20:51] <benji> bac: unfortunately I have no idea; the failing tests on our fork have conditioned me not to run the tests
[20:51] <bac> gah, i want to strangle my mechanic.  end of week five might have it done by middle of next week.
[20:52] <bac> benji: when i finished my branch there were only 4 failures.  i thought subunit worked.
[20:52] <gary_poster> benji, bac, we do have instructions on what tests are supposed to pass in versions.cfg
[20:52] <benji> right, I forgot about that
[20:53] <benji> we really need to at least squelch the failing tests
[20:53] <gary_poster> I forgot to ask about it
[20:53] <gary_poster> we could do that pretty easilyI woud guess
[20:53] <gary_poster> would
[20:54] <gary_poster> I will add card
[21:02] <bac> the expected failures per versions.cfg (reallly an odd place to track it) are http://pastebin.ubuntu.com/913757/
[21:02] <bac> i'm seeing this in trunk http://pastebin.ubuntu.com/1041435/
[21:05] <bac> hmm, that looks mighty funny
[21:06] <bac> so i'm picking up zope.testrunner from my ~/.buildout.  is that right?  is that what we use when running tests for zope.testing?
[21:08] <gary_poster> um, I don't think so
[21:08] <gary_poster> I think that may be from dependencies
[21:08] <gary_poster> and their tests
[21:08] <gary_poster> we use older version
[21:08] <gary_poster> I have to run
[21:08] <gary_poster> but--13 green runs!  new record
[21:09] <gary_poster> sorry must run; baby bored
[21:11] <bac> benji: can you try grabbing our 3.9.4 trunk and running tests?  i think i'm getting weird, extra bad results