AlexDaniel releasable6: next 00:25
releasable6 AlexDaniel, Next release in 3 days and ≈18 hours. Blockers: github.com/rakudo/rakudo/issues?q=...%9A%A0%22. Unknown changelog format
[Coke] does he ever not complain about the changelog format? 01:20
AlexDaniel releasable6: changelog raw.githubusercontent.com/rakudo/r.../ChangeLog 01:28
releasable6 AlexDaniel, Successfully fetched the code from the provided URL.
AlexDaniel, Changelog for this release was not started yet
AlexDaniel, Details: gist.github.com/d41a8c344866c53e77...66400d503b
AlexDaniel releasable6: changelog gist.githubusercontent.com/AlexDan.../changelog 01:29
releasable6 AlexDaniel, Successfully fetched the code from the provided URL.
AlexDaniel, 0 out of 249 commits logged (⚠ 1 warnings)
AlexDaniel, Details: gist.github.com/8986cb3eaf404e64f4...d1f752238b
AlexDaniel [Coke]: yes ↑
at this moment it's only useful right before the release when the actual changelog is committed 01:30
I can change it so that it understands the wiki page github.com/rakudo/rakudo/wiki/ChangeLog-Draft 01:32
but given how popular this idea is… maybe not 01:33
timotimo oh, i just didn't write anything in the changelog because i didn't do anything this month 01:36
okay, i jitted isprime_I on moarvm 01:37
oh, microoptimizing the DON'T_DIVIDE_NUMBERS function was this month, too 01:38
AlexDaniel timotimo: the JIT change should be here I think github.com/MoarVM/MoarVM/wiki/ChangeLog-Draft 01:39
timotimo it's also almost completely worthless ;)
[Coke] \o/
AlexDaniel FWIW I'm OK with changelogging all of rakudo changes, my goal was to stimulate microcontributions to the release process but this is not necessarily required 01:41
Geth nqp: jstuder-gh++ created pull request #390:
Rakudo Issue 1304: [js] Allow 'chain' op to use child as callee		 04:46
jnthn Got a fix for github.com/rakudo/rakudo/issues/1410 and testing it 10:53
Geth rakudo: e072474efd | (Jonathan Worthington)++ | src/core/Supply.pm
Sanitize one-whenever supply input

Previously it was serialized, but this makes sure it follows protocol also, so we can safely rely on that internally.		 11:01
rakudo: 2192ddd170 | (Jonathan Worthington)++ | src/core/Supply.pm
Fix duplicate sending of done message

The one-whenever supply/react optimized path was vulnerable to sending the `done` message twice, which it must not do if it's claiming to be a protocol-following Supply. This fixes the double-vow crash reported in issue #1410.
synopsebot RAKUDO#1410 [open]: github.com/rakudo/rakudo/issues/1410 [regression][⚠ blocker ⚠] supply/react “already vowed” regression
roast: ccff418444 | (Jonathan Worthington)++ | S17-supply/syntax.t
Test to cover #1410 (double-done bug)
 11:02
lizmat Files=1232, Tests=76026, 327 wallclock secs (14.88 usr 5.32 sys + 2251.59 cusr 230.84 csys = 2502.63 CPU) 11:10
nine t/spec/S17-supply/syntax.t seems to flap a lot more 12:52
lizmat saw it flap in her last spectest 12:53
running one again
jnthn wonders if it's the newly added test 12:55
lizmat You planned 83 tests but ran 69 12:56
so I guess it's test 70
jnthn Ah, then no
lizmat ah, it seems to consistently fail now 13:00
but apparently only under load
jnthn Haven't seen it crash at all here 13:01
Neither in spectest nor run on its own a bunch of times
lizmat Unhandled exception in code scheduled on thread 12
Access denied to keep/break this Promise; already vowed
in block at t/spec/S17-supply/syntax.t line 543
jnthn Do you have commit 2192ddd170?
lizmat doublechecks 13:02
yes
jnthn grmbl
That fixed another case with the same error
lizmat gist.github.com/lizmat/cafd72f89ee...b7d9f4f900
what I do is run the spectest, and then repeatedly run that single file with --ll-exception 13:03
in another window
masak .oO( "...don't do that, then!" ) :P 13:11
lizmat jnthn: I've now got two stacktraces that differ (added it as a comment in the gist) 13:12
where the trace differs, theres an ominous: # XXX Workaround for optimizer bug comment 13:13
also: SupplyOneWheneverTappable is in the trace, so looks like the new code is tickling a maybe already existing bug ? 13:14
timotimo we should have a tiny piece of code that saves a backtrace of when a promise was vowed, when it was kept/rejected, and when it was attempted to be kept/rejected or vowed again 13:36
for example with a wrap call 13:37
lizmat clickbaits www.perl.com/article/an-open-lette...community/ 13:40
tadzik ++! 13:45
jnthn Will look at that bug a bit later, fwiw; am working on something else at the moment :)
lizmat twitter.com/perlfoundation/status/...3467768833
jnthn lizmat++ 13:48
stmuk_ 13:45:37 lizmat | twitter.com/perlfoundation/status/...3467768833 14:59
oops
lizmat oops?
stmuk_ trying to config a new terminal :)
lizmat ah, ok :-)
Geth nqp/master: 4 commits pushed by pmurias++ 15:34
rakudo/js: 4 commits pushed by pmurias++ 15:46
jnthn Righty, I've a little time to look at that syntax.t thing 16:05
jnthn deletes code 16:14
Code that doesn't exist can't be broke
Geth rakudo: 2dbc2bcbd6 | (Jonathan Worthington)++ | src/core/Supply.pm
Further hardening against double-done sending

This hopefully resolves the remaining cases where that can happen in the single-whenever supply/react. Along the way, it unifies two very similar code paths into one correct one.		 16:19
jnthn Hopefully that helps :) 16:20
|Tux| Rakudo version 2017.12-252-g2dbc2bcbd - MoarVM version 2017.12.1-36-g20a4a7976
csv-ip5xs1.092 - 1.094
csv-ip5xs-2012.466 - 12.641
csv-parser11.832 - 12.036
csv-test-xs-200.436 - 0.501
test10.874 - 10.974
test-t2.740 - 2.745
test-t --race1.119 - 1.134
test-t-2048.604 - 49.756
test-t-20 --race16.379 - 16.536
 17:31
Top 5
2018-01-14 19:39:12 test-t 2.758
2018-01-12 17:28:22 test-t 2.758
2018-01-15 11:08:15 test-t 2.756
2018-01-17 18:20:48 test-t 2.745
2018-01-17 18:24:01 test-t 2.740
timotimo nice to see that the 5 best times are all from the last week :) 17:36
[Coke] \o/ 17:37
Geth nqp: dwarring++ created pull request #391:
remove unreachable code in cursor_init		 17:38
nqp: 3a9a497db9 | (David Warring)++ | src/QRegex/Cursor.nqp
remove unreachable code in cursor_init

we're already in an else block from a previous nqp::isconcrete(self) && $!braid test		 19:12
nqp: a76455f2c4 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/QRegex/Cursor.nqp
Merge pull request #391 from dwarring/cursor_init-tweak

remove unreachable code in cursor_init
tbrowder lizmat: fwiw, i added CPAN5 tags to my two ported modules several hours ago but they don’t show up on the modules.perl6.org/t/CPAN5 search 20:33
lizmat maybe it takes a while, dunno :-) 20:34
Geth rakudo: a1fd0b7322 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/core/List.pm
Mention List.sink is as-designed

Closes github.com/rakudo/rakudo/issues/1393		 21:36
rakudo/post-release: ae697080d2 | (Zoffix Znet)++ | src/Perl6/Actions.nqp
Use target type as default default for coercers

This way we don't coerce optional parameters when they're not provided, avoiding uwanted crashes and warnings.
Fixes RT#130479: rt.perl.org/Ticket/Display.html?id=130479		 23:45
synopsebot RT#130479 [new]: rt.perl.org/Ticket/Display.html?id=130479 [BUG] Type coercion in signature gives warning on optional named arguments
Geth roast/post-release: 6457b4fab8 | (Zoffix Znet)++ | S12-coercion/coercion-types.t
Test coercers don't coerce their default default

RT#130479: rt.perl.org/Ticket/Display.html?id=130479 Rakudo fix: github.com/rakudo/rakudo/commit/ae697080d2		 23:46
nqp/post-release: 8 commits pushed by (Zoffix Znet)++, pmurias++, (David Warring)++, (Elizabeth Mattijsen)++ 23:54
roast/post-release: 14 commits pushed by usev6++, (Zoffix Znet)++, (Jeremy Studer)++, (Elizabeth Mattijsen)++, (Jonathan Worthington)++
review: github.com/perl6/roast/compare/645...df8a72ec2a		 23:55
rakudo/post-release: 32 commits pushed by (Jonathan Worthington)++, (Zoffix Znet)++, (Elizabeth Mattijsen)++, (Tom Browder)++
review: github.com/rakudo/rakudo/compare/a...d74010ba1b

Powered by Cro and the Raku® Programming Language.

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

Thank you!