Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-release, 2016-02-22

| Channels | #perl6-release index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:01 [Coke] oh, guess what I forgot about!
00:01 * [Coke] kicks off some tests...
01:08 japhb joined #perl6-release
01:57 [Coke] nqp-2016.02 tagged and pushedtarballed.
01:58 [Coke] weird, wonder why that ^W didn't erase a word.
02:07 [Coke] liz - you broke tools/update-tai-utc.pl, I think.
02:08 [Coke] ah. no, just a doc update missing.
02:37 [Coke] again I rant about the fact that we have Inline::Perl5 tests in roast.s
02:49 ilbot3 joined #perl6-release
02:49 Topic for #perl6-release is now »r̈« - Discussions about Perl 6 and Rakudo release strategies - http://irclog.perlgeek.de/perl6-release/today
03:09 [Coke] release isn't happening today; there are failing tests.
03:22 [Coke] https://rt.perl.org/Ticket/Display.html?id=127589 is the meta ticket; there are four failing test files (no surprise, I think these were noted a few days ago); There was one that was a passing TODO, which I've unfudged in the errata branch, since that's one of the things that's for.
07:55 FROGGS joined #perl6-release
10:06 jnthn [Coke]++
11:14 FROGGS joined #perl6-release
11:42 hoelzro joined #perl6-release
11:57 [Coke] jnthn: so lizmat && moritz & I have updated some tests in the 6.c-errata branch; Can you verify that these changes appear reasonable for the purpose of that branch?
11:58 [Coke] (commit messages have lots of detail)
12:42 jnthn [Coke]: Just gone through them, I'm fine with all of them
13:19 [Coke] Ok. I'll pull the trigger on a tarball sometime this morning, Eastern time.
13:20 [Coke] kicking off a fine stresstest now
13:23 * lizmat hopes the release will be in time for the next P6W, due in ~ 9 hours
13:24 [Coke] lizmat: yup
13:24 [Coke] mainly because you fixed all those test! thanks!
13:30 [Coke] /spec/S17-procasync/no-runaway-file-limit.t did not like being run under high TEST_JOBS
13:31 [Coke] (just hung; never finished)
13:31 [Coke] slow enough running solo.
13:45 lizmat [Coke]: fwiw, I've seen that happening a few times
14:20 [Coke] do we care to deal with "Missing test file"? New tests have been added to roast/master (and therefore rakudo/nom) that don't exist in the 6.c-errata branch. I'm happy to leave it alone, and perhaps add a warning to the errata branch readme.
14:23 lizmat perhaps we should add empty test files ?
14:24 lizmat [Coke]: I think that would be best: I mean, the original tests are still in the 6.c-errata branch
14:24 [Coke] ?
14:25 [Coke] I don't understand your last send.
14:25 lizmat some test file got split into these new testfiles by moritz (because of memory hogging)
14:25 [Coke] I mean, I get all the words, but not your intent.
14:25 [Coke] that's not the only change, though. there are 3 missing files.
14:25 [Coke] and as we add new 6.d stuff, those files will be missing.
14:26 [Coke] we could track moritz's split-test-file-only changes into the branch, sure.
14:27 lizmat so, all new test files should be added as dummies to 6.c-errata, I would guess
14:27 lizmat to make sure we don't get a "cry wolf" effect on the "missing test file" messages
14:28 [Coke] I wouldn't do that, no.
14:28 lizmat well, we can't change spectest.data from roast, can we ?
14:28 [Coke] that fixes the symptom (the warning), but then we have "why are these test files added to the spec?"
14:28 lizmat we could leave a message in the stubs, indicating the reason ?
14:29 [Coke] It's more busy work.
14:29 lizmat so what's your suggestion?  just ignoring the "missing test file" message ?
14:29 jnthn Just either ignore the missing test file message, or only spew it out when the current branch of t/spec is master
14:29 jnthn That'll catch our development mistakes
14:30 jnthn I'd like to find a way to quiet passing todo tests when testing 6.c too
14:30 jnthn So we don't have to do that busy work either.
14:30 [Coke] actually - that's a good point; we can't use t/spectest.data in general, because what if a file is -removed- from it for a later version
14:30 jnthn Yeah...we'll need a better solution there
14:30 nwc10 instead of "only on master" is there an easy way to programatically detect that the branch is "on a stable release" ?
14:30 [Coke] jnthn: I don't mind fixing the errata branch when something passes a formerly failing todo. that seems ok to me.
14:31 nwc10 is my question moot given [Coke]'s observation about deleted files?
14:31 nwc10 anyway, I don't have a good answer for the "bigger" picture
14:31 [Coke] We could add a VERSION (or something similar to what I put in the README on the 6.c-errata branch)
14:31 [Coke] so for the stable version, it really should be "run everything"
14:31 nwc10 but I'm wary of anything that is "only on master" as it means that development branches can have mistakes that aren't spotted prior to integration
14:33 jnthn [Coke]: It may be is a bit of over-reach for -errata to untodo stuff. Not really now, but in the future if somebody is trying to target the errata version of a certain langauge relesae, but not implementing anything beyond it.
14:34 [Coke] jnthn: ok. we have at least 2 in the branch doing that right now.
14:34 jnthn 6.c-errata isn't meant to mean "and stuff that was in todo'd tests that we don't really consider supported until a later release"
14:34 [Coke] I mean, if we had our act together, we'd probably have stripped all the todo tests as part of 6.c
14:34 jnthn Right, that's what I'd like us to try and do for 6.c.1 an onward
14:34 jnthn *and
14:34 jnthn *if* we do that
14:35 jnthn Then the spectest.data issue goes away
14:35 jnthn You ignore spectest.data
14:35 [Coke] well, now's the time to back that ouf the errata branch, if you like.
14:35 jnthn And run everything
14:35 jnthn (That is, for a released version)
14:35 [Coke] as I just cut a tarball: http://rakudo.org/downloads/rakudo/
14:43 masak joined #perl6-release
15:21 cognominal joined #perl6-release
15:35 [Coke] 2016.02 compiler release cut
15:37 jnthn \o/
15:37 jnthn [Coke]++ # thanks!
18:30 crux joined #perl6-release
20:20 jnthn joined #perl6-release
20:25 crux joined #perl6-release

| Channels | #perl6-release index | Today | | Search | Google Search | Plain-Text | summary