Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-07-08

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

All times shown according to UTC.

Time Nick Message
00:04 BenGoldberg joined #perl6-dev
03:02 TimToady joined #perl6-dev
07:46 hankache joined #perl6-dev
08:16 Geth ¦ rakudo/nom: a2090821b4 | (Timo Paulssen)++ | src/core/IO/Socket.pm
08:16 Geth ¦ rakudo/nom: make sure IO::Socket passes a list to set-line-separators
08:16 Geth ¦ rakudo/nom:
08:16 Geth ¦ rakudo/nom: should unbreak Net::SMTP
08:16 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a2090821b4
08:16 FROGGS joined #perl6-dev
08:16 FROGGS o/
08:20 hankache joined #perl6-dev
08:33 nine Good morning!
08:37 timotimo greetings nine
08:39 Geth ¦ roast: ab290effa1 | (Timo Paulssen)++ | S32-io/IO-Socket-INET.t
08:39 Geth ¦ roast: test setting nl-in before the first get in IO::Socket::INET
08:39 Geth ¦ roast:
08:39 Geth ¦ roast: this used to explode in Net::SMTP
08:39 Geth ¦ roast: review: https://github.com/perl6/roast/commit/ab290effa1
08:44 hankache_ joined #perl6-dev
08:53 lizmat Files=1209, Tests=64165, 220 wallclock secs (13.29 usr  5.03 sys + 1338.71 cusr 136.92 csys = 1493.95 CPU)
09:00 nine I somehow need to find my Perl 6 mojo again. Haven't done anything useful in weeks :/
09:18 Geth ¦ nqp: 8453fac89e | (Stefan Seifert)++ | src/HLL/Compiler.nqp
09:18 Geth ¦ nqp: Factor out part of compile for easier re-use.
09:18 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/8453fac89e
11:24 Geth ¦ rakudo/nom: e1e9091f74 | (Timo Paulssen)++ | src/core/CompUnit/Repository/FileSystem.pm
11:24 Geth ¦ rakudo/nom: stop looking after finding a matching module file
11:24 Geth ¦ rakudo/nom:
11:24 Geth ¦ rakudo/nom: if both a .pm6 and a .pm exist with a matching name, only the
11:24 Geth ¦ rakudo/nom: .pm would be used, which is clearly not right.
11:24 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e1e9091f74
12:10 * lizmat is contemplating https://perl6.party/post/The-Hot-New-Language-Named-Rakudo over the weekend
13:15 Geth ¦ rakudo: MasterDuke17++ created pull request #1109: Only create Exception messages when gisting
13:15 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1109
13:24 Geth ¦ rakudo/nom: 1a4d94930c | (Daniel Green)++ | src/core/Exception.pm
13:24 Geth ¦ rakudo/nom: Only create Exception messages when gisting
13:24 Geth ¦ rakudo/nom:
13:24 Geth ¦ rakudo/nom: With the recent change to suggest similar method names when throwing
13:24 Geth ¦ rakudo/nom: X::Method::NotFound, it got a lot slower to try possible methods (e.g.,
13:24 Geth ¦ rakudo/nom: `try Int.sin`) even if you didn't care about the exception message. This
13:24 Geth ¦ rakudo/nom: commit removes the .message call from .throw and puts it in both
13:24 Geth ¦ rakudo/nom: branches of .gist.
13:24 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1a4d94930c
13:24 Geth ¦ rakudo/nom: 60b9acef30 | lizmat++ (committed using GitHub Web editor) | src/core/Exception.pm
13:24 Geth ¦ rakudo/nom: Merge pull request #1109 from MasterDuke17/only_create_Exception_messages_when_gisting
13:24 Geth ¦ rakudo/nom:
13:24 Geth ¦ rakudo/nom: Only create Exception messages when gisting
13:24 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/60b9acef30
13:25 jnthn I was just reviewing that
13:25 jnthn It has the slight issue that we're now throwing exceptions with a message unset
13:26 jnthn Which means that if they're ever unhandled to the degree that they're spat out by the VM, then they'll all have empty messages
13:26 MasterDuke how would that happen?
13:26 jnthn MasterDuke: When something went very wrong, usually
13:27 jnthn So "normally it won't".
13:27 llfourn MasterDuke: did that pass spectests?
13:27 MasterDuke llfourn: yep
13:27 llfourn oh nice
13:28 jnthn I think we can leave the change in and see what happens
13:29 jnthn But if you have a situation where you are seeing a crash with no exception message, just a backtrace, then this commit will be worth looking at.
13:29 MasterDuke put in a placeholder message in .throw that doesn't call .message?
13:29 MasterDuke or is that no better than no message?
13:29 jnthn Not really
13:29 jnthn tbh I'd probably have just overridden gist
13:30 jnthn And left message as a cheap "Method 'foo' not found on invocant of type Bar"
13:30 jnthn And put the espensive stuff in the .gist
13:30 jnthn Though that probably has trade-offs too
13:30 jnthn Maybe let's leave it as it is now and see if it actually bites us
13:31 MasterDuke cool
13:31 jnthn Because it only will if we manage to miss all of the Perl 6 installed exception handlers
13:31 jnthn The place I fear it biting us is when compiling CORE.setting
13:31 jnthn But even maybe then it could be OK
13:32 jnthn Guess we've a week until the next release; we'll keep an eye on it :)
13:32 MasterDuke and i didn't time anything other than the snippet in the PR, but this should speed up all exceptions that aren't gisted, though probably none quite so much as that Method::NotFound
13:33 jnthn Indeed
13:45 BenGoldberg joined #perl6-dev
14:11 Geth ¦ roast: 0b36afb8b4 | (Elizabeth Mattijsen)++ | S15-unicode-information/uniprop.t
14:11 Geth ¦ roast: Untodo now passing high codepoint test
14:11 Geth ¦ roast: review: https://github.com/perl6/roast/commit/0b36afb8b4
14:12 dogbert17 if anyone is insterested in taking a peek at some new ASAN barfage then look no further :-) https://gist.github.com/dogbert17/db210a75ab12862c46c19233a0579fd9
14:19 jnthn dogbert17: Please file that as a MoarVM issue so it doesn't get forgotten :)
14:21 dogbert17 do you really think that Zoffix is innocent :)
14:22 * dogbert17 creates issue ...
14:24 jnthn This time, maybe... :)
14:26 dogbert17 done, https://github.com/MoarVM/MoarVM/issues/611
14:27 jnthn Thanks :)
14:27 MasterDuke .oO(who needs automated continuous integration when they have a dogbert17++?)
14:28 dogbert17 btw, don't know if you have backlogged today but samcv's font program works when spesh is disabled
14:29 dogbert17 MasterDuke: at $work I'm far from being a role model when it comes to writing tests
14:30 MasterDuke heh, same here. in fact, not sure if i've ever written a test at $work
14:30 dogbert17 totally OT, I have received most components for my new computer, only the PSU is still on order
14:31 MasterDuke i'm hoping i'll only need a MB+CPU+RAM and can re-use everything else
14:32 dogbert17 reuse is good, my current PSU should probably work fine, also a Seasonic, but it's more than 10 years old, it's already been reused once
14:36 MasterDuke i think my current setup finally ditched the last part carried along from when i first built it in 1999 when i did the most recent CPU+MB+RAM upgrade in 2010
14:37 dogbert17 cool, my current system is also from 2010
14:38 lizmat .ask ab5tract do we agree that [(^)] $a, $b, $c is the same as ($a (^) $b) (^) $c ?
14:38 yoleaux lizmat: I'll pass your message to ab5tract.
14:52 dogbert17 if I have a small program, ~10 lines, which crashes during a normal run but works when MVM_SPESH_DISABLE=1, is there anything I can do to narrow things down?
14:54 llfourn dogbert17: what does it crash with?
14:55 llfourn (the error)
14:56 dogbert17 ==6370== ERROR: AddressSanitizer: heap-use-after-free on address 0x8ad139c0 at pc 0xb58a1400 bp 0xbf8a40a8 sp 0xbf8a409c
14:56 dogbert17 READ of size 4 at 0x8ad139c0 thread T0
14:56 dogbert17 #0 0xb58a13ff in gc_mark /home/dogbert/repos/rakudo/nqp/MoarVM/src/6model/reprs/MVMCallCapture.c:50
14:57 llfourn oh the one above. Sorry I have no idea, but would also love to know.
14:57 MasterDuke that looks like the same thing the whateverables are hitting and timotimo was trying to figure out yesterday
14:59 dogbert17 here's the code, it's a slightly slaughtered test from a file called .../stress.t https://gist.github.com/dogbert17/0b0baed4ac8aaf762a3e26f8c74110f1
14:59 moritz dogbert17: you can try with leaving out SPESH_DISABLE, and instead only run with MVM_SPESH_INLINE_DISABLE or MVM_SPESH_OSR_DISABLE
14:59 dogbert17 moritz: will do
14:59 moritz dogbert17: if you 'git grep getenv' in moarvm, you get an idea of what env variables are supported
15:02 dogbert17 moritz: it fails with both MVM_SPESH_INLINE_DISABLE and MVM_SPESH_OSR_DISABLE
15:02 moritz dogbert17: that kinda suggests it's a generic spesh bug
15:03 moritz and likely lucky, because inlining is rather tricky :-)
15:03 dogbert17 I generated a spesh log, it doesn't tell me anything and besides, it's 25 megs
15:06 [Tux] This is Rakudo version 2017.06-176-g60b9acef3 built on MoarVM version 2017.06-49-g6011f876
15:06 [Tux] csv-ip5xs        2.529
15:06 [Tux] test            12.381
15:06 [Tux] test-t           4.133 - 4.138
15:06 [Tux] csv-parser      12.072
15:06 MasterDuke dogbert17: maybe trying running it in rr, saving the recording, and seeing if timotimo can make anything from it
15:08 AlexDaniel joined #perl6-dev
15:09 dogbert17 MasterDuke: rr: command not found :( where can I find it?
15:10 AlexDaniel should be in your distro repos
15:10 AlexDaniel (e.g. apt install rr)
15:11 AlexDaniel MasterDuke: “that looks like the same thing the whateverables are hitting” I'm pretty sure [Coke] was hitting it yesterday as well
15:11 AlexDaniel although he didn't give any details
15:14 AlexDaniel dogbert17: also, “Fails when run normally but works with MVM_SPESH_DISABLE=1” is probably not entirely true
15:14 AlexDaniel maybe bump ^80 to something higher to see it fail again
15:15 AlexDaniel buggable: status
15:15 AlexDaniel buggable: help
15:15 buggable AlexDaniel, tags | tag SOMETAG | eco | eco Some search term | author github username | speed | testers CPANTesters report ID
15:15 AlexDaniel NeuralAnomaly: status
15:15 NeuralAnomaly AlexDaniel, [✘] Next release will be in 6 days and 13 hours. Since last release, there are 22 new still-open tickets (1 unreviewed and 0 blockers) and 179 unreviewed commits. See http://perl6.fail/release/stats for details
15:16 Geth ¦ rakudo/nom: 80b3e89b53 | (Elizabeth Mattijsen)++ | src/core/set_symmetric_difference.pm
15:16 Geth ¦ rakudo/nom: Replace different type handling by multiple candidates
15:16 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/80b3e89b53
15:16 dogbert17 it wasn't in the repo, not with that name anyway, trying to build from src
15:16 AlexDaniel what distro?
15:17 dogbert17 Linux Mint
15:17 AlexDaniel oh, I see, it's not even in debian stable probably
15:18 moritz is that mozilla's tool for debugging stuff?
15:18 dogbert17 Fails => Fails with ASAN barfage
15:19 AlexDaniel moritz: yes
15:20 AlexDaniel ah, by the way, do we have a ticket already?
15:21 AlexDaniel .tell timotimo https://irclog.perlgeek.de/perl6-dev/2017-07-08#i_14844351 same issue? But with a lot less code to reproduce
15:21 yoleaux AlexDaniel: I'll pass your message to timotimo.
15:22 AlexDaniel MasterDuke: hey, let's bisect it?
15:22 AlexDaniel MasterDuke: dogbert17's code doesn't require anything
15:22 AlexDaniel except no restricted setting, but that we can do
15:24 moritz well, you might need ASAN to reproduce it reliably
15:29 AlexDaniel well, whateverables are reliably not working, so I'm hoping that it should be ok…
15:33 * dogbert17 blah, rr doesn't work under VirtualBox it seems :(
15:40 timotimo AlexDaniel: different issue
15:40 yoleaux 15:21Z <AlexDaniel> timotimo: https://irclog.perlgeek.de/perl6-dev/2017-07-08#i_14844351 same issue? But with a lot less code to reproduce
15:40 AlexDaniel hm ok
15:41 timotimo the other one, though, the one in MVMCallCapture, that's probably the same issue
15:51 timotimo i'm glad i was able to at least fix two little things today
15:52 dogbert17 perhaps you can fix a third as well :)
15:54 AlexDaniel dogbert17: well, with your code, I got it bisected to 92c187d2dc. I can't get the segfault on any commit before that
15:54 AlexDaniel dogbert17: this makes sense in a way because it touched Proc::Async, but obviously that's not the reason why it segfaults
15:57 dogbert17 AlexDaniel: perhaps you should try using the bots with a version preceeding that commit then
15:57 * AlexDaniel laughs
15:57 dogbert17 :)
15:57 AlexDaniel well, it may work :D
15:58 AlexDaniel dogbert17: but really, I bumped it to “for ^1000”, segfaults reliably on 92c187d2dc but not on commits before it
16:01 dogbert17 I could try reverting that change in my environment and see if the example program works
16:03 AlexDaniel pretty sure it will
16:03 AlexDaniel but that doesn't really help :|
16:04 dogbert17 but it might give jnthn and timotimo a clue as to what's going on
16:07 nine Well destructuring a slurpy arg is probably not used all that often.
16:08 timotimo is this about the MVMCallCapture explosion?
16:09 dogbert17 yes, i.e. https://gist.github.com/dogbert17/0b0baed4ac8aaf762a3e26f8c74110f1
16:09 dogbert17 the explosion vanishes if MVM_SPESH_DISABLE=1
16:09 timotimo i believe this will happen whenever there's a callsite that's not interned, for example because it has some kind of flattening going on
16:10 dogbert17 AlexDaniel suspects this commit: https://github.com/rakudo/rakudo/commit/92c187d2dc395f7ec31ada38ee0724a9dba6f13e
16:10 AlexDaniel well I don't really suspect it, but the segfault appeared after that commit, yes
16:11 timotimo we might want to hotfix this by having @args again but passing path => @args[0]
16:12 dogbert17 my rebuild takes forever, suits me right, have both ASAN and --no-optimize enabled
16:14 nine How can I run rakudo with ASAN?
16:14 timotimo asan is something you have to compile your stuff with
16:14 timotimo so you'd have to pass --asan to moarvm's Configure.pl
16:15 nine I did that. Anything else?
16:15 timotimo nope. every invocation of rakudo will now be super slow
16:15 dogbert17 but much faster than valgrind
16:15 timotimo yup
16:17 nine Ok, now my moarvm just segfaults :_
16:17 nine :)
16:19 timotimo it might be necessary to recompile rakudo
16:19 timotimo because rakudo has that shared library it loads
16:19 timotimo the extops and the containers and such
16:19 nine I tried. But even a rakudo/install/bin/moar just segfaults
16:19 timotimo wow
16:19 dogbert17 reverting https://github.com/rakudo/rakudo/commit/92c187d2dc395f7ec31ada38ee0724a9dba6f13e seems to do the trick
16:19 nine #1  0x00007ffff38f31c1 in __asan::MaybeInstallSigaction (signum=signum@entry=11, handler=0x7ffff38f3140 <__asan::ASAN_OnSIGSEGV(int, siginfo_t*, void*)>) at ../../../../libsanitizer/asan/asan_posix.cc:39
16:20 nine Even tried a make realclean and built MoarVM again
16:20 AlexDaniel dogbert17: try a higher number of iterations please
16:21 AlexDaniel dogbert17: because I've seen things not segfault otherwise
16:21 AlexDaniel like 10000 or something :)
16:22 dogbert17 works at 480
16:26 dogbert17 also works with 2480 iterations
16:28 * nine utterly failed at golfing this issue
16:29 nine oooooh!
16:29 nine The nqp::force_gc makes a huge difference :)
16:30 nine Maybe I should have looked at the failing example instead of writing one from scratch :P
16:31 nine m: use nqp; class Foo { multi method new(*@args ($p, *@a), *%_) { self.bless(:$p, :@a, |%_) } } ; my @a = <foo bar baz quz>; my int $i = 0; sub foo(*@a) { with Foo.new(|@a) { $i++ }; $i--; $i++; say $i.Int; nqp::force_gc; }; for ^10000 { foo(|@a); }
16:31 camelia rakudo-moar 80b3e8: OUTPUT: «(signal SEGV)1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤…»
16:31 nine Smallest example so far...
16:32 dogbert17 nine: and if you MVM_SPESH_DISABLE=1
16:34 dogbert17 AlexDaniel: 5000 iterations also works
16:34 nine dogbert17: then it succeeds
16:34 dogbert17 that should be quite a clue for our speshialists :-)
16:35 nine Actually it sometimes even runs to completion with spesh enabled
16:35 nine Though most of the time it segfaults after 79 or 258 iterations
16:35 dogbert17 I bet that it would stop if you had ASAN working
16:36 nine Well if ASAN didn't completely kill my moarvm I'd be happy to test :/
16:36 dogbert17 you could do valgrind ofc but it will be sloooow
16:36 nine Need to leave for dinner now. If someone would be so kind to add my golfed version to the bug report, please :)
16:37 dogbert17 timotimo, your hot fix is beginning to sound tempting :-)
16:38 dogbert17 *hotfix
16:40 dogbert17 hmm, do we have bug report?
16:47 MasterDuke asan output of nine's example: https://gist.github.com/MasterDuke17/fc12f33fd8e271b0b15bf2acd0d8ef55
16:52 MasterDuke the summary and consumed views in heaptrack of running Zoffix's perl6-Toaster webapp and doing a couple gets (leaks a lot of memory, believed to be due to DBIish): http://i.imgur.com/N8mi3e2.png http://i.imgur.com/FkKudKN.png
16:54 Zoffix wow that app looks really cool
16:57 dogbert17 Zoffix: how's your studies going, have you passed 25% ?
17:01 Zoffix m: printf "%.2f", 201/831
17:01 camelia rakudo-moar 80b3e8: OUTPUT: «0.24»
17:01 Zoffix Nope
17:01 Zoffix I didn't do any this week, though I started again about 15 minutes ago.
17:02 dogbert17 have you backlogged? one of your commits seems to have uncovered a spesh bug
17:03 Zoffix No. Which commit?
17:04 dogbert17 https://github.com/rakudo/rakudo/commit/92c187d2dc395f7ec31ada38ee0724a9dba6f13e
17:05 Zoffix ok
17:06 dogbert17 is it possible to get the effect you want without the destructuring?
17:06 Zoffix multi method new(*@args, *%_) { my $path = @args.shift; self.bless(:$path, :@args, |%_) }
17:06 Zoffix Probably faster too
17:06 dogbert17 cool
17:08 Zoffix We still have a kludge for bit shift ops: https://github.com/rakudo/rakudo/blob/80b3e89/src/core/Int.pm#L400-L415
17:08 Zoffix https://rt.perl.org/Ticket/Display.html?id=131306
17:09 Zoffix m: use nqp; class Foo { multi method new(*@a, *%_) { my $p = @a.shift; self.bless(:$p, :@a, |%_) } } ; my @a = <foo bar baz quz>; my int $i = 0; sub foo(*@a) { with Foo.new(|@a) { $i++ }; $i--; $i++; say $i.Int; nqp::force_gc; }; for ^10000 { foo(|@a); }
17:09 camelia rakudo-moar 80b3e8: OUTPUT: «1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51␤52␤53…»
17:09 Zoffix m: use nqp; class Foo { multi method new(*@a, *%_) { my $p = @a.shift; self.bless(:$p, :@a, |%_) } } ; my @a = <foo bar baz quz>; my int $i = 0; sub foo(*@a) { with Foo.new(|@a) { $i++ }; $i--; $i++; say $i.Int; nqp::force_gc; }; for ^10000 { foo(|@a); }
17:09 camelia rakudo-moar 80b3e8: OUTPUT: «1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51␤52␤53…»
17:11 dogbert17 what happens if you run it locally?
17:11 Zoffix m: class Foo { method x (*@ ($a, *@b)) {} }; for ^100_000 { $ = Foo.x: 1, 2, 3, 4 }; say now - INIT now
17:11 camelia rakudo-moar 80b3e8: OUTPUT: «3.1004012␤»
17:11 Zoffix m: class Foo { method x (*@b) {$ = @b.shift} }; for ^100_000 { $ = Foo.x: 1, 2, 3, 4 }; say now - INIT now
17:11 camelia rakudo-moar 80b3e8: OUTPUT: «0.324994␤»
17:11 Zoffix m: say 3.1/.32
17:11 camelia rakudo-moar 80b3e8: OUTPUT: «9.6875␤»
17:11 dogbert17 that's quite some difference
17:15 dogbert17 is the correct term for stuff like '*@args ($p, *@a)' parameter destructuring?
17:15 Zoffix ¯\_(ツ)_/¯
17:16 dogbert17 ok, I'll write a MoarVM issue for that
17:24 dogbert17 ok, https://github.com/MoarVM/MoarVM/issues/612
17:25 dogbert17 so, who should implement Zoffix optimized workaround ?
17:26 Zoffix I was about to commit, but it's wrong
17:26 Zoffix m: class Foo { method x (*@b where .so) {$ = @b.shift} }; for ^100_000 { $ = Foo.x: 1, 2, 3, 4 }; say now - INIT now
17:26 camelia rakudo-moar 80b3e8: OUTPUT: «0.5779238␤»
17:26 Zoffix Right way is still faster tho
17:27 dogbert17 Zoffix++, I suspect that your fix will make quite a few people very happy
17:27 Zoffix It's not a fix. It's a workaround a problem :)
17:28 dogbert17 :)
17:28 dogbert17 and you have an issue to refer to
17:29 dogbert17 I tried a sloppy grep of the Rakudo src but I only managed to find one case of parameter destructuring (the one you're changing), but there has to be more of them or ?
17:40 Zoffix Why has to be?
17:40 Zoffix ZOFFLOP: t/spec/S32-io/open.t
17:42 Geth ¦ rakudo/nom: f6d4fbd24b | (Zoffix Znet)++ | src/core/Proc/Async.pm
17:42 Geth ¦ rakudo/nom: Work around SEGV on sig unpacking
17:42 Geth ¦ rakudo/nom:
17:42 Geth ¦ rakudo/nom: - Also makes invocation 6x faster
17:42 Geth ¦ rakudo/nom: - Bug worked around: https://github.com/MoarVM/MoarVM/issues/612
17:42 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f6d4fbd24b
17:42 Zoffix It would've been here, but it's already using the `where .so` thing: https://github.com/rakudo/rakudo/blob/nom/src/core/Proc.pm#L221
17:43 * dogbert17 rebuilds
17:51 dogbert17 samcv's Font program now works
17:52 dogbert17 perhaps the bots will work better as well
17:56 Geth ¦ rakudo/nom: b391692644 | (Elizabeth Mattijsen)++ | src/core/set_symmetric_difference.pm
17:56 Geth ¦ rakudo/nom: Fix a few edge cases of (^) wrt QuantHash type permutations
17:56 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b391692644
18:03 AlexDaniel timotimo: so should I just rebuild rakudo on the server and enjoy now working bots?
18:04 AlexDaniel .tell [Coke] https://irclog.perlgeek.de/perl6-dev/2017-07-08#i_14844908
18:04 yoleaux AlexDaniel: I'll pass your message to [Coke].
18:04 AlexDaniel .tell [Coke] this may help your Proc::Async issue
18:04 yoleaux AlexDaniel: I'll pass your message to [Coke].
18:10 dogbert17 AlexDaniel: Zoffix has worked around the problem
18:10 AlexDaniel yes, I see
18:16 Geth ¦ rakudo/nom: 05c255c14b | (Elizabeth Mattijsen)++ | src/core/set_symmetric_difference.pm
18:16 Geth ¦ rakudo/nom: Simplify [(^)] for 2+ with Baggy semantics
18:16 Geth ¦ rakudo/nom:
18:16 Geth ¦ rakudo/nom: Fixes the failing test in S03-operator/mix.t
18:16 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/05c255c14b
18:18 lizmat .tell ab5tract https://github.com/rakudo/rakudo/commit/05c255c14b
18:18 yoleaux lizmat: I'll pass your message to ab5tract.
18:18 dogbert17 afk &
18:21 MasterDuke huh, earlier today i built rakudo with asan, but now even nqp won't build
18:22 MasterDuke https://gist.github.com/MasterDuke17/3f461a964e0a92426a5a923543b6e454
18:23 lizmat afk again&
18:23 MasterDuke moar/nqp/rakudo all at HEAD
18:36 Zoffix damn, I'm still getting "Failed to rename" with precomp on Windows
19:35 ugjka joined #perl6-dev
19:49 BenGoldberg joined #perl6-dev
20:02 committable6 joined #perl6-dev
20:02 greppable6 joined #perl6-dev
20:04 committable6 joined #perl6-dev
20:12 greppable6 joined #perl6-dev
21:26 vendethiel joined #perl6-dev
21:55 stmuk joined #perl6-dev
22:00 stmuk_ joined #perl6-dev
22:11 BenGoldberg joined #perl6-dev
22:36 unicodable6 joined #perl6-dev
22:36 bisectable6 joined #perl6-dev
22:36 evalable6 joined #perl6-dev
22:36 greppable6 joined #perl6-dev
22:36 committable6 joined #perl6-dev
22:36 bloatable6 joined #perl6-dev
22:36 quotable6 joined #perl6-dev
22:36 coverable6 joined #perl6-dev
22:36 benchable6 joined #perl6-dev
22:36 statisfiable6 joined #perl6-dev
22:41 Geth ¦ roast: 20c194f0f4 | (Samantha McVey)++ | S17-procasync/stress.t
22:41 Geth ¦ roast: Let S17-procasync/stress.t be launched from roast folder
22:41 Geth ¦ roast:
22:41 Geth ¦ roast: Previously it had to be launched from t/spec/ assuming it was in
22:41 Geth ¦ roast: rakudo's folder. This fixes that so it can be started from both.
22:41 Geth ¦ roast: review: https://github.com/perl6/roast/commit/20c194f0f4
23:31 samcv wow. i just got all Emoji V4.0 passing. and no more of GraphemeClusterBreak.t failing
23:31 samcv 240 emoji in one swoop
23:32 samcv :D
23:52 Zoffix samcv++

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