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.
lizmat Files=1255, Tests=75814, 367 wallclock secs (15.82 usr 6.06 sys + 2351.48 cusr 260.35 csys = 2633.71 CPU) 07:54
this is with the js branch, and is significantly slower than before :-( 07:55
Geth rakudo/js: a24c37b3b4 | (Paweł Murias)++ | src/vm/js/make-release.js
[js] Stop using hardcoded paths in our release helper
 08:27
rakudo/js: f731c81466 | (Paweł Murias)++ | src/core/Lock/Async.pm6
[js] Re-indent code
[Tux] Rakudo version 2018.09-72-g687a3cbb0 - MoarVM version 2018.09-123-g0191bd9cf
csv-ip5xs0.904 - 0.922
csv-ip5xs-207.516 - 7.659
csv-parser22.308 - 23.763
csv-test-xs-200.431 - 0.446
test8.019 - 8.869
test-t1.801 - 1.890
test-t --race0.847 - 0.962
test-t-2032.858 - 33.647
test-t-20 --race11.724 - 13.179
 10:42
(probably high because I'm actually working :) 10:43
Geth rakudo/js: 350a4bd46e | (Paweł Murias)++ | 3 files
[js] Make .hyper and .race return a modified HyperSeq and RaceSeq so we can type check against it
 11:15
rakudo/js: 3cc3574437 | (Paweł Murias)++ | src/core/JavaScriptScheduler.pm6
[js] Remove dead and commented out code
 11:20
Kaiepi when should $scheduler.queue(:hint-affinity) be used over $scheduler.queue? 11:21
dogbert17 ZOFFLOP: t/spec/MISC/misc-6.d.rakudo.moar - ===SORRY!=== An exception occurred while evaluating a constant at /home/dogbert/repos/rakudo/t/spec/MISC/misc-6.d.rakudo.moar:123 - You cannot create an instance of this type (Int:D) 11:22
jnthn Kaiepi: When you're going to have a sequence of events and they're going to be processed sequentially, and so there's no point having them sprayed over a bunch of threads only for them to stumble into each other. 11:24
Kaiepi: Typically, if the thing you're scheduling will feed a Supply, you want to give that hint to the scheduler 11:25
Kaiepi ah 11:27
thanks
Geth rakudo/js: 92c504ba55 | (Paweł Murias)++ | src/Perl6/Metamodel/BOOTSTRAP.nqp
[js] Share a #?if with the jvm backend
 11:39
pmurias how do you measure that? time make m-spectest? 12:13
pmurias I'll see if I can duplicate and investigate the slowdown 12:16
lizmat pmurias: just the last line of 'make spectest' :-) 12:22
pmurias: hmmmm... maybe something else is going on
if I compare it to a later (warm) spectest of master, the js branch one actually uses significantly less CPU, just way more wallclock 12:23
pmurias: so please defer looking into this until tomorrow, where I have done another make spectest on the js branch on a cold machine 12:24
AlexDaniel ===SORRY!=== 12:52
P6opaque: no such attribute '@!dispatchees' on type Routine in a Scalar when trying to get a value
ehhh wish me luck
oh, that was quick 12:53
AlexDaniel it's something to do with EXPORT magic… 12:53
Geth rakudo: be2d4cb4b3 | (Timo Paulssen)++ | lib/NativeCall.pm6
make nativecall's `postcircumfix:<[ ]>` return containers

so that you can `$my-carray[0]++` for example		 13:06
lizmat m: role A { method a() { ... } }; role B does A { has $!b handles <a>; }; class C does B { } # is this intentional or a NYI ? 13:40
camelia 5===SORRY!5=== Error while compiling <tmp>
Method 'a' must be implemented by C because it is required by roles: B.
at <tmp>:1
lizmat in other words: do "handled: methods not count for being implemented ?
Zoffix dogbert17: is it actually a flopping test or a consistent failure? 13:49
dogbert17: Also, how did you run that test file? 13:52
Zoffix .tell dogbert17 can you elaborate on this? colabti.org/irclogger/irclogger_log...-10-06#l72 Was it an actual flop or a consistent failure? What OS 64/32bit? How did you run that test file? That failure points to the 6.d switching mechanism being broken 14:04
yoleaux Zoffix: I'll pass your message to dogbert17.
dogbert17 Zoffix: back from your business trip? 14:20
yoleaux 14:04Z <Zoffix> dogbert17: can you elaborate on this? colabti.org/irclogger/irclogger_log...-10-06#l72 Was it an actual flop or a consistent failure? What OS 64/32bit? How did you run that test file? That failure points to the 6.d switching mechanism being broken
dogbert17 Zoffix: it's a flop but I can repro it (!) 14:22
.tell Zoffix, it happens on my 32 bit Linux VM 14:23
yoleaux dogbert17: What kind of a name is "Zoffix,"?!
dogbert17 grrr
.tell Zoffix it happens on my 32 bit Linux VM
yoleaux dogbert17: I'll pass your message to Zoffix.
dogbert17 running 'Rakudo version 2018.09-72-g687a3cbb0 built on MoarVM version 2018.09-123-g0191bd9cf'
Zoffix . 14:25
yoleaux 14:23Z <dogbert17> Zoffix: it happens on my 32 bit Linux VM
Zoffix I'm on 2018.09-73-gbe2d4cb4b
dogbert17: ok, I'll try to reproduce on my 32bit VM
(had it running on 64-bit for the last 30 minutes and not a single failure)
Yes, I'm back from the trip... Won $500 bucks at the games, but also got a cold -_- 14:26
dogbert17 Zoffix: cool, if I get it during spectest I will also get it *the first time I run* it separately i.e. './perl6 t/spec/MISC/misc-6.d.rakudo.moar' 14:27
congrats, perhaps you should spend some on the latest Tomb Raider game
Zoffix meh
dogbert17 not your kind of game then :) 14:28
Zoffix I still have unfinished earlier tomb raider 14:29
dogbert17 aha
Zoffix Filed R#2343 14:31
Geth: help
Geth Zoffix, Source at github.com/perl6/geth To add repo, add an 'application/json' webhook on GitHub pointing it to hack.p6c.org:8888/?chan=#perl6 and choose 'Send me everything' for events to send | use `ver URL to commit` to fetch version bump changes
Zoffix R##2343
R#2343
R#2342
dogbert17 added a comment to RH #2343 14:33
R #2343
hmm 14:34
Zoffix R#2343 14:35
synopsebot R#2343 [open]: github.com/rakudo/rakudo/issues/2343 [32-bit][6.d] Apparent failure in 6.d language switching mechanism on 32-bit boxes
travis-ci Rakudo build failed. Timo Paulssen 'make nativecall's `postcircumfix:<[ ]>` return containers 14:46
travis-ci.org/rakudo/rakudo/builds/437997150 github.com/rakudo/rakudo/compare/6...2d4cb4b301
buggable [travis build above] ☠ All failures are due to: failed make test (6 failures). Across all jobs, only t/05-messages/moar/01-errors.t test file failed. 14:46
Zoffix No one bothered to fix that test still? 15:02
Geth rakudo: 0d98607059 | (Zoffix Znet)++ | t/05-messages/moar/01-errors.t
Fix SEGV test

MVM_JIT_LOG no longer complains about failed opens. Cover for its SEGV only.		 15:06
timotimo zoffix, is there any need to cover MVM_JIT_LOG at all? it's been removed from moarvm, so it's not quite possible for it to have an effect liek a crash 15:43
Zoffix So why didn't people who had that information fix the test? 15:44
This is just comical.
timotimo i didn't actually look at the test failure :( 15:45
Geth rakudo: 2d7724f8bc | (Zoffix Znet)++ | t/05-messages/moar/01-errors.t
MVM_JIT_LOG has been removed
Geth nqp: Kaiepi++ created pull request #505:
Implement asyncsocket op		 16:00
rakudo: Kaiepi++ created pull request #2344:
Create async sockets before passing them to .listen/.connect		 16:01
travis-ci Rakudo build passed. Zoffix Znet 'Fix SEGV test 16:08
travis-ci.org/rakudo/rakudo/builds/438024588 github.com/rakudo/rakudo/compare/b...98607059b6
travis-ci Rakudo build passed. Zoffix Znet 'MVM_JIT_LOG has been removed' 17:07
travis-ci.org/rakudo/rakudo/builds/438033676 github.com/rakudo/rakudo/compare/0...7724f8bca3
pmurias it's good the the moar error message tests wheren't some mysterious problem with the js branch :) 17:46
Kaiepi does anyone know why the BOOTIO object returned by nqp::asyncsocket always has its handle pointing to 0xFFFFFFFF00000017 when i pass it to nqp::asynclisten? hastebin.com/focaxeroba.rb 22:16
i wish i hadn't kept screwing with it when i had it fully working
timotimo did you check the REPR is correct? i.e. no decont missing for example? 22:23
Kaiepi: ^
Kaiepi i tried using a decont and it didn't fix the problem 22:24
i haven't checked the repr though
timotimo could very well be an undefined object, too 22:27
Kaiepi it it possible to do from gdb? i can't check it from perl since it segfaults as soon as i try to call .tap
timotimo yes, you can "print IS_CONCRETE(obj)"
and "print REPR(obj)"
Kaiepi it's P6opaque 22:29
which isn't right
timotimo yeah, it's some kind of class
print STABLE(obj)->debug_name
Kaiepi Scalar 22:30
i guess i'll try deconting it again and checking what info shows up for it then
timotimo in that case a decont would have been right
you can try call MVM_dump_p6opaque or what it's called on the object 22:31
Kaiepi how do i debug an error like this when it doesn't tell me anything other than invalid argument? hastebin.com/tekasuwave.coffeescript 23:56

Powered by Cro and the Raku® Programming Language.

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

Thank you!