Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
06:09 Kaiepi left 06:10 Kaiepi joined 07:19 patrickb joined
patrickb .tell Kaiepi Can you have a look at github.com/rakudo/rakudo/pull/2857 ? I can't add you as a reviewer, but it'd be good if you could have a look if the change is sane. 07:41
yoleaux patrickb: I'll pass your message to Kaiepi.
lizmat Files=1262, Tests=107912, 411 wallclock secs (30.41 usr 7.09 sys + 2930.99 cusr 254.50 csys = 3222.99 CPU)
Kaiepi looks good to me patrickb 07:43
yoleaux 07:41Z <patrickb> Kaiepi: Can you have a look at github.com/rakudo/rakudo/pull/2857 ? I can't add you as a reviewer, but it'd be good if you could have a look if the change is sane.
Kaiepi actually wait
nqp::execname() doesn't return an empty string on openbsd, usually it returns one of the flags passed to moarvm 07:44
|Tux| Rakudo version 2019.03.1-284-gb0dd44b00 - MoarVM version 2019.03-110-g081823ffc
csv-ip5xs0.696 - 0.708
csv-ip5xs-205.904 - 5.948
csv-parser21.539 - 21.547
csv-test-xs-200.424 - 0.426
test6.774 - 6.871
test-t1.651 - 1.683
test-t --race0.761 - 0.772
test-t-2028.209 - 29.148
test-t-20 --race8.823 - 8.860
 08:18
lizmat is it correct to say that modules.perl6.org is not getting updated? specifically modules.perl6.org/search/?q=Git%3A%3ABlame 10:10
does not seem to be in there 10:11
which was uploaded 3 days ago ?
timotimo kicks off a rebuild 10:18
there's not yet a cron job 10:19
btw where do the apocalypses and exegeses live? the design repo only has the synopses 10:20
lizmat github.com/ugexe/Perl6-ecosystems/.../cpan.json appears to be empty, which is bad I think :-) 10:21
github.com/ugexe/Perl6-ecosystems/...b2a4c6a114 10:22
1 addition, 42,933 deletions not shown because the diff is too large. Please use a local Git client to view these changes.
timotimo: maybe not a good idea to run the job now :-( 10:23
.tell ugexe seems we have an empty ecosystem: github.com/ugexe/Perl6-ecosystems/...b2a4c6a114 10:24
yoleaux lizmat: I'll pass your message to ugexe.
lizmat continues to work on P6W 10:25
timotimo oh, it is? 10:29
[2019-05-06 10:29:14.35700] [31398] [info] Processing dist 613 of 1382
looks like at least that is going fine 10:30
modules should be updated 10:59
lizmat hnmm modules.perl6.org/dist/Git::Blame:cpan:JMERELO is still a 404 ? 11:17
modules.perl6.org/search/?q=Git%3A%3ABlame still 0 results ?
timotimo [2019-05-06 10:18:55.34486] [31398] [info] rsyncing CPAN dists from cpan-rsync.perl.org::CPAN/authors/id 11:24
rsync: failed to connect to cpan-rsync.perl.org (207.171.7.135): Connection timed out (110)
i'll try a mirror 11:26
11:53 llfourn left 12:13 Geth left, Geth joined 12:23 llfourn joined 12:28 llfourn left
lizmat reportable6: 2019-04-22T12:00:00Z 2019-04-29T00:00:00Z 12:38
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
lizmat, gist.github.com/53c5185d19b4e28979...2f475e5132 12:39
AlexDaniel lizmat: but why 12:00 ?
lizmat copy/pasto 12:40
reportable6: 2019-04-22T00:00:00Z 2019-04-29T00:00:00Z
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
lizmat, gist.github.com/7cca5cf7dbc1e3dc8b...7abce19bf3 12:41
lizmat reportable6: 2019-04-29T00:00:00Z 2019-05-06T00:00:00Z
reportable6 lizmat, Couldn't find a snapshot for 2019-05-06T00:00:00Z (try ļ½¢listļ½£ command to see what's available).
lizmat :-( 12:42
reportable6: 2019-04-29T00:00:00Z 2019-05-05T00:00:00Z
reportable6 lizmat, Couldn't find a snapshot for 2019-05-05T00:00:00Z (try ļ½¢listļ½£ command to see what's available).
lizmat hmmm
AlexDaniel: ^^ ??
AlexDaniel reportable6: list
reportable6 AlexDaniel, gist.github.com/4df6565ed8cae23e69...c85465f290
lizmat reportable6: 2019-04-01T00:00:00Z 2019-05-01T00:00:00Z
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
AlexDaniel reportable6: :( yeah sometimes it hangs and therefore stops doing snapshots 12:43
lizmat argh... probably another Perl NOC related outage :-(
reportable6 AlexDaniel, I cannot recognize this command. See wiki for some examples: github.com/perl6/whateverable/wiki/Reportable
lizmat, gist.github.com/24e46ea11249dd436e...c3567dbe97
AlexDaniel could be, but sometimes it's reportable itself that acts weird
12:43 reportable6 left
lizmat reportable6: 2019-04-29T00:00:00Z 2019-05-04T00:00:00Z 12:44
AlexDaniel I restarted itā€¦
lizmat yeah, got thta 12:45
does that imply it will do the missing ones again, or is there some manual process involved ?
AlexDaniel lizmat: it can't do the missing ones, it just fetches all the tickets regularly. If it missed some planned snapshots, then we don't know the state of things during these days 12:46
lizmat reportable6: 2019-04-29T00:00:00Z 2019-05-04T00:00:00Z 12:47
12:47 reportable6 joined
AlexDaniel still starting upā€¦ :) 12:47
lizmat reportable6: 2019-04-29T00:00:00Z 2019-05-04T00:00:00Z
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
lizmat AlexDaniel++
AlexDaniel reportable6: snapshot
reportable6 AlexDaniel, OK! Working on it. This will take forever, so don't hold your breath. 12:48
lizmat, gist.github.com/9d8ae56cae757cf03e...7ceccdc492
13:08 llfourn joined 13:13 llfourn left
Guest55475 hmm, a few scheduler related spectest suddenly fails 13:14
timotimo timing-related perhaps? 13:15
Guest55475 perhaps: not ok 22 - Passing :every as Inf never runs the given block
timotimo oh, that's weird! 13:16
i bet i broke things! :D
where does that test live?
Guest55475 t/spec/S17-scheduler/every.t 13:18
similar failures in t/spec/S17-scheduler/at.t and in t/spec/S17-scheduler/in.t
timotimo can you try turning spesh off 13:20
Guest55475 no improvement 13:23
timotimo OK, that's good so it's not my fault :) 13:24
can you run it without prove to see if there's any additional output?
i can't reproduce the issue locally
Guest55475 Failed test 'Passing :every as Inf never runs the given block' at t/spec/S17-scheduler/every.t line 120, expected: '0' got: '2445' 13:25
13:27 llfourn joined
timotimo ha 13:28
that file has only 107 lines on my system
Guest55475 it was updated today :)
timotimo does the "passing :at as NaN throws" test also fail? 13:29
Guest55475 test #25 fails as well yes 13:31
13:32 llfourn left
Guest55475 ha, I missed some relevant text. The following is also written: Minimum timer resolution is 1ms; using that instead of Infms 13:33
reportable6 AlexDaniel, No! It wasn't me! It was the one-armed man! Backtrace: gist.github.com/4d40f68d4c6b7658bd...c189d21b66
timotimo the thread pool scheduler code at least explicitly checks isnanorinf and if that's true only warns 13:38
hm, will :every(Inf) actually really never run the block? i'd imagine it'd run the block only once 13:40
github.com/rakudo/rakudo/pull/2819 - this should have been merged before or along with the tests 13:42
Guest55475 indeed, timotimo++ 13:44
timotimo commented on it
Guest55475 cool
lizmat oops... 13:50
13:51 llfourn joined
lizmat reverted the merge 13:52
Guest55475 and spectest is suddenly more or less clean 13:55
13:56 llfourn left
timotimo is someone spectesting with ASAN? 13:56
Guest55475 from time to time yes, shall I give it a spin? 13:59
14:00 lucasb joined
Guest55475 does so 14:01
timotimo thanks. make sure my latest terrible moarvm commits are in there :) 14:04
Guest55475 the latest seems to be 'MAST Ops: remove redundant range checks' 14:06
timotimo yeah 14:07
14:11 |Tux| left 14:18 |Tux| joined 14:28 llfourn joined
Guest55475 timotimo: spectest was clean with asan enabled 14:29
timotimo nice 14:30
14:33 llfourn left 14:55 llfourn joined, patrickb left 15:04 llfourn left
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/05/06/...s-to-riga/ 15:24
15:36 llfourn joined, ddddddd joined
ddddddd "I'm feeling lucky" on modules.perl6.org gives a 404 15:37
lizmat ddddddd: please file an issue :-) 15:39
15:41 llfourn left
ddddddd The error page said file an issue or report it here 15:44
I guess I can file an issue too
Kaiepi can someone on linux build moar with --optimize=0 --debug=3, run ./Configure.pl on rakudo, edit the makefile to use -O0 -g3 with cc/ld flags, run make install, look at the coredump generated, and tell me what line in src/vm/moar/runner/main.c the double free is happening in with github.com/Kaiepi/rakudo/tree/moar-runner ? 15:45
15:49 llfourn joined 15:54 llfourn left 15:57 ddddddd left 16:25 llfourn joined 16:30 llfourn left
timotimo i'll give it a try 16:37
Kaiepi: with just make install, i get no coredumps or crashes 16:42
Kaiepi run make spectest 16:43
timotimo,
the double frees start happening then on travis
timotimo how many test files does it usually take? 16:44
it's kind of not crashing at all 16:53
Kaiepi weird 16:56
timotimo yeah 16:57
ugexe m: sub foo(@arr where .elems = [1,2]) { say @arr }; foo([1,]); # bug? 17:01
yoleaux 10:24Z <lizmat> ugexe: seems we have an empty ecosystem: github.com/ugexe/Perl6-ecosystems/...b2a4c6a114
camelia [1]
ugexe er 17:02
m: sub foo(@arr where .elems = [1,2]) { say @arr }; foo([]); # bug?
camelia []
timotimo why would you .elems = [1, 2]?
oh haha 17:03
ugexe using @arr [$,*@] works, but still...
timotimo m: sub foo(@arr where .elems) { @arr.perl.say }; foo([])
camelia []
timotimo m: sub foo(@arr where *.elems) { @arr.perl.say }; foo([])
camelia Constraint type check failed in binding to parameter '@arr'; expected anonymous constraint to be met but got Array ($[])
in sub foo at <tmp> line 1
in block <unit> at <tmp> line 1
timotimo that's how, hm?
m: sub foo(@arr where *.elems = [1,2]) { @arr.perl.say }; foo([]) 17:04
camelia Constraint type check failed in binding to parameter '@arr'; expected anonymous constraint to be met but got Array ($[])
in sub foo at <tmp> line 1
in block <unit> at <tmp> line 1
timotimo m: sub foo(@arr where *.elems = [1,2]) { @arr.perl.say }; foo()
camelia [1, 2]
ugexe hmm. i just started operating under the assumption i didnt need to use * in signatures for most things :P
17:05 robertle joined
ugexe i reverted the cpan ecosystems back to their last state, and updated ecogen to not update an ecosystem if there is no meta data. This does not yet address adding mirrors; there is a PR but I don't think it will work as-is 17:15
17:20 llfourn joined
timotimo oh no! 17:24
that was my PR
17:25 llfourn left
ugexe im not sure the mirrors will help or not at this point 17:30
since it seems like maybe PAUSE itself was/is down? (so no new uploads)
now we just need a metaecosystem ecosystem for aggregating the ecosystems and their mirrors 17:31
timotimo no 17:32
only the one rsync server was down
the other one had the new package that wasn't available on the other one
hm, actually
hm, which mirror did i put in? 17:33
ugexe blogs.perl.org/users/mohammad_s_anw...again.html
timotimo the german rsync mirrors actually are a bit late
netcologne is at 29 hours, rwth-aachen is 22 hours old
hs-esslingen is a whooping 1.5 days old 17:34
ugexe hmm, must just be a coincidence then
timotimo but the cpan-rsync.perl.org one is 23 hours out of date and its data seems to be 23 hours old, too?
i wonder how that works
vrurg nine: would you mind if I copy your make code into nqp-configure repository? 17:57
.tell nine would you mind if I copy your make code into nqp-configure repository? 17:59
yoleaux vrurg: I'll pass your message to nine.
nine vrurg: not at all 18:09
yoleaux 17:59Z <vrurg> nine: would you mind if I copy your make code into nqp-configure repository?
vrurg nine: thanks. NQP builds with make.nqp, but Rakudo doesn't. Looks like m-clean target is lost on Makefile object. 18:10
nine make.nqp is about a year old and I implemented exactly what was needed for a rakudo build. So requirements may have upped in the mean time 18:12
vrurg nine: Not sure. It's a simple case of clean: m-clean // m-clean: 18:15
I'm investigating. 18:16
18:33 vrurg left 18:41 llfourn joined 18:48 llfourn left 19:00 lucasb left, brrt joined 19:02 vrurg joined 19:16 gfldex left 19:35 gfldex joined 19:39 Geth_ left, b2gills left 19:40 b2gills joined 19:41 llfourn joined 19:47 llfourn left 20:08 ufobat_ joined 20:11 ufobat__ left 20:28 brrt left 20:38 lucasb joined 20:55 llfourn joined 21:01 llfourn left
Kaiepi timotimo, i think i figured out what was going on with the double free in the moar runner on travis 21:07
on linux dirname returns the same pointer passed to it so freeing dir_path_temp was causing a double free since dir_path was already freed 21:08
21:29 llfourn joined 21:34 llfourn left 22:03 llfourn joined
timotimo ooh 22:08
22:09 llfourn left, llfourn joined 22:40 lucasb left

Powered by Cro and the RakuĀ® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!