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.
ugexe [SCHEDULER 99951] Will not add extra worker; hit 64 thread limit [branch with some total completed] 00:00
wonder if thats something like $finished.then({ #`(cant spawn a thread to reach here) }); cleanup()
quite a few "[SCHEDULER 99951] Heuristic queue progress deadlock situation detected" before that 00:01
right around when it would slowdown
gist.github.com/ugexe/5e1576f8cbd6...cf9a0489fb 00:02
japhb ugexe: Yeah, something is definitely not being cleaned up. 00:03
ugexe: Would you mind updating (or do you mind if I update) the bug report with your gist? 00:04
ugexe sure, go for it 00:06
patrickb I thought some more about handling command line arguments of subprocesses on Windows. Here is a short writeup of my findings and thoughts so far: gist.github.com/patzim/73d4e9c7f0c...cessing-md 00:07
japhb MasterDuke: shrinking MAX_THREADS gives ugexe's result faster, but not linearly so; it got maybe 2/3 as far with only 4 threads.
patrickb ugexe: ^ If you're interested. 00:08
And with that I'm off for today.
o/ 00:09
japhb OK, notes from our exploration added to the sync bug. 00:14
00:15 patrickb left 00:57 commavir_ joined, scovit_ joined 01:02 commavir left, scovit left, commavir_ is now known as commavir 01:07 MasterDuke left 01:09 MasterDuke joined, MasterDuke left, MasterDuke joined
MasterDuke vrurg: does --target=parse -e '' work with your branch to add .e.PREVIEW? 03:15
vrurg MasterDuke: dunno. Didn't try. Too late for today, will check tomorrow. 03:16
o/
MasterDuke k, later...
04:50 ggoebel joined
Geth nqp: 59d7a8869c | usev6++ | 2 files
[JVM] Set outer handler when calling unwind check

I'm not very sure about this patch but it helps with on failing test (cmp. RT #131167) and has no (additional) fallout for spectest. Please revert if it's the wrong thing to do.		 06:08
synopsebot RT#131167 [new]: rt.perl.org/Ticket/Display.html?id=131167 [JVM] Newly-added NQP test for catching exceptions fails on JVM
Geth rakudo: 8f1b1babd2 | usev6++ | t/spectest.data
[JVM] Don't run some newly added sprintf tests

Some of the tests from S32-str/sprintf-c.t and S32-str/sprintf-s.t are failing on the JVM backend due to Unicode/Grapheme handling (afaik). Since it's not easy to add fudges for the generated tests, it seems simpler to skip running those test files for now.		 06:22
06:51 ufobat__ joined 06:55 ufobat_ left
releasable6 Next release in ≈2 days and ≈11 hours. 13 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 07:00
lizmat Files=1262, Tests=102483, 398 wallclock secs (27.07 usr 7.13 sys + 2844.86 cusr 241.57 csys = 3120.63 CPU) 07:50
07:51 patrickb joined
|Tux| Rakudo version 2019.03.1-219-g8f1b1babd - MoarVM version 2019.03-82-g7b112da34
csv-ip5xs0.686 - 0.725
csv-ip5xs-205.878 - 5.931
csv-parser22.262 - 22.652
csv-test-xs-200.446 - 0.446
test6.959 - 7.054
test-t1.624 - 1.644
test-t --race0.852 - 0.866
test-t-2027.540 - 28.067
test-t-20 --race10.226 - 10.705
 07:55
20181015 1.659❙20181011 1.647❙20190418 1.644❙20190417 1.630❙20190418 1.624❙
:)
timotimo japhb: have you already tried getting stack traces from all threads when the perl6 interpreter is deadlocked? 08:40
08:53 vrurg left
Geth roast: ad2297b640 | (Elizabeth Mattijsen)++ | S32-str/sprintf-b.t
Streamline test generation / running of sprintf("b|B")

To allow for easier addition of additional tests that may not actually fit in the current construction.		 09:25
roast: 8f3fe2dad5 | (Elizabeth Mattijsen)++ | S32-str/sprintf-c.t
Streamline test generation / running of sprintf("c")

To allow for easier addition of additional tests that may not actually fit in the current construction. Also generate the "#" cases, as they are identical to the non-"#" cases.
patrickb lizmat: I fixed up r#2846 . Seemed like you intended to merge. ;-) 09:28
synopsebot R#2846 [open]: github.com/rakudo/rakudo/pull/2846 Remove relative elements from NQP_HOME and PERL6_HOME
Geth rakudo: 65217aaafc | (Patrick Böker)++ | 2 files
Remove relative elements from NQP_HOME and PERL6_HOME

This in turn results in the repository path specs to not contain any relative path elements.		 09:29
rakudo: e11d0d54e4 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 2 files
Merge pull request #2846 from patzim/no-rel-repo-paths

Remove relative elements from NQP_HOME and PERL6_HOME
patrickb lizmat: Thanks! 09:30
lizmat patrickb++ :-)
Geth roast: 8dda96a33b | (Elizabeth Mattijsen)++ | S32-str/sprintf-d.t
Streamline test generation / running of sprintf("d|i")

To allow for easier addition of additional tests that may not actually fit in the current construction.		 09:34
roast: 743bc0ef7e | (Elizabeth Mattijsen)++ | S32-str/sprintf-e.t
Streamline test generation / running of sprintf("e|E")

To allow for easier addition of additional tests that may not actually fit in the current construction.		 09:38
jnthn Ah, neat, seems like I can run `make test` again without having to first `make install` 10:10
timotimo hooray
jnthn patrickb++, I assume :)
Geth rakudo: 4702c245f0 | (Jonathan Worthington)++ | src/core/Metamodel/Primitives.pm6
Expose newmixintype via Metamodel::Primitives

Recently, types that were the target of a rebless operation started needing to be created explicitly as mixin target types, to assist optimization. Expose the ability to do this in Metamodel::Primitives by allowing a `:mixin` flag to be passed to `create_type`.		 10:14
roast: c78326fa14 | (Jonathan Worthington)++ | S12-meta/primitives.t
Cover Metamodel::Primitives.rebless

We'd never covered - and thus not specified - this method before now. This does so, and along the way covers the `:mixin` option to the
  `create_type` method. This means that there's now an official way
forward for github.com/rakudo/rakudo/issues/2839.		 10:16
rakudo: e352962284 | (Elizabeth Mattijsen)++ | src/core/Mixy.pm6
Suggest .roll when trying to do a Mixy.pick

Suggested by Ralph Mellor in:
   stackoverflow.com/questions/557432...ck-routine		 10:19
roast: 6a541f5e6b | (Elizabeth Mattijsen)++ | S32-str/sprintf-f.t
Streamline test generation / running of sprintf("f|F")

To allow for easier addition of additional tests that may not actually fit in the current construction.		 10:28
roast: c690803e82 | (Elizabeth Mattijsen)++ | S32-str/sprintf-o.t
Streamline test generation / running of sprintf("o")

To allow for easier addition of additional tests that may not actually fit in the current construction.		 10:31
roast: 2939371b55 | (Elizabeth Mattijsen)++ | S32-str/sprintf-s.t
Streamline test generation / running of sprintf("s")

To allow for easier addition of additional tests that may not actually fit in the current construction. Also generate the "#" cases, as they are identical to the non-"#" cases.		 10:36
roast: 4aadc2c09d | (Elizabeth Mattijsen)++ | S32-str/sprintf-u.t
Streamline test generation / running of sprintf("u")

To allow for easier addition of additional tests that may not actually fit in the current construction.		 10:39
jnthn m: use Test; class ParameterChild is Parameter { has $.foobar }; is ParameterChild.new(foobar => 'Baz').foobar, 'Bar', 'Subclassing of Parameter works'; 10:44
evalable6 (exit code 1) not ok 1 - Subclassing of Parameter works
# Failed test 'Subclassing of Parameter works'
# at /tmp/XHDQoe2GaQ line 1
# expected: 'Bar'
# got: (Any)
Geth roast: 7c142c6d5d | (Elizabeth Mattijsen)++ | S32-str/sprintf-x.t
Streamline test generation / running of sprintf("x|X")

To allow for easier addition of additional tests that may not actually fit in the current construction.		 10:45
patrickb jnthn: The respective PR r#2842 isn't merged yet...
synopsebot R#2842 [open]: github.com/rakudo/rakudo/pull/2842 Fix build runners
jnthn patrickb: Hm, curious... 10:46
Maybe the relative paths one helped a bit?
patrickb jnthn: Currently it's dependent on CWD
They do work with CWD = build dir
but should break when cwd is anything else
jnthn ah, I see 10:47
Geth roast: fae956e5a0 | (Elizabeth Mattijsen)++ | 2 files
Remove some superfluous .uc's
 10:48
rakudo: d31fb27064 | (Jonathan Worthington)++ | 2 files
Unbreak subclassing of Parameter

With this, the regression reported in #2829 that broke
  `Cro::OpenAPI::RoutesFromDefinition` is now resolved.
It's covered for now by a Rakudo test, though arguably we should consider the ability to do this spec and add spectest coverage also.		 10:51
synopsebot RAKUDO#2829 [open]: github.com/rakudo/rakudo/issues/2829 [BLOCKER][regression] Parameter subclasses don't initialize attributes
lizmat jnthn++ # good catch 10:55
Geth nqp: 3800095a42 | (Jonathan Worthington)++ | src/QRegex/Cursor.nqp
Protect against self-reference in match dump

This is sufficient to deal with the hang that was reported in
  github.com/rakudo/rakudo/issues/2827.		 11:09
jnthn patrickb: Should github.com/MoarVM/MoarVM/pull/1089 still be merged? 11:14
patrickb jnthn: It should. 11:15
Kaiepi is kind of waiting on this actually.
jnthn OK, wasn't sure if work was underway to not need it :) Merged now 11:16
Lunch time o/
lizmat jnthn: shall I do some bumps ? 11:17
Geth roast: d74f77549e | (Elizabeth Mattijsen)++ | 9 files
Use more specific is-deeply for testing
 11:36
lizmat m: dd sprintf("%08f",Inf) # intriguing numbers 11:41
evalable6 "00000Inf"
lizmat m: dd sprintf("%08f",-Inf)
evalable6 "-0000Inf"
lizmat m: dd sprintf("%.3f",NaN) # I was hoping for NaN.00 11:43
evalable6 "NaN"
11:48 ggoebel left
Geth roast: abcef56825 | (Elizabeth Mattijsen)++ | S32-str/sprintf-d.t
Add variable precision tests + some big number tests for "%d|i"
 12:40
lizmat TIL about variable precision in sprintf 12:41
m: dd sprintf("%8.*d",[4,666]) 12:42
evalable6 " 0666"
lizmat I don't think that's documented, actually 12:44
m: dd sprintf("%8.*d",4,666)
evalable6 " 0666"
lizmat interesting
ah, no, it *is* documented 12:45
Geth roast: 31ecaedd47 | (Elizabeth Mattijsen)++ | S32-str/sprintf-d.t
Make sure we flatten the value to be checked
 13:04
13:22 vrurg joined
lizmat commute to NR.PM& 13:42
13:42 lizmat left
vrurg .tell MasterDuke it's not so many changes in the branch, so --target=parse fails too. 13:43
yoleaux vrurg: I'll pass your message to MasterDuke.
13:49 MasterDuke left 14:04 patrickb left 14:48 lucasb joined
Geth rakudo: patzim self-assigned Invocation used to run Perl 6 code under remote debugger broken github.com/rakudo/rakudo/issues/2850
vrurg++ created pull request #2852: Add revision 6.e and improve multi-revision support
 15:05
japhb .tell timotimo No, I hadn't gotten a stack trace from all threads at the point of deadlock. What's the easiest way to do that? 15:15
yoleaux japhb: I'll pass your message to timotimo.
timotimo . 15:17
yoleaux 15:15Z <japhb> timotimo: No, I hadn't gotten a stack trace from all threads at the point of deadlock. What's the easiest way to do that?
timotimo :P
japhb timotimo: Why the ':P' ? 15:20
timotimo i just arrived here the moment you sent that .tell 15:27
japhb Ah 15:39
Question remains open though. ;-)
timotimo oh 15:45
right
when you have gdb attached, for example through a core dump or because you started your script with perl6-gdb-m, then you'd ctrl-c and "threads apply all bt" 15:46
japhb timotimo: Is there a way to write that thread bt output to a file? 15:52
timotimo would it be enough if i told you to "set pager off"? 15:53
or maybe it's "set pagination off" 15:54
(gdb) set logging file my_god_object.log 16:00
this ought to write stuff to a file
though probably first have to "set logging on"
japhb OK, will try. 16:03
16:26 lizmat joined 16:50 vrurg left 16:51 vrurg joined 17:15 lizmat left 17:19 go|dfish joined 17:48 lizmat joined 19:05 lizmat left 19:38 lucasb left 19:44 lizmat joined
roast: vrurg++ created pull request #530:
Removed all .REVIEW modificators from v6.d		 19:49
19:54 lizmat left 20:05 brrt joined 20:16 AlexDaniel left 20:24 squashable6 left 20:27 squashable6 joined 21:00 brrt left 22:12 lizmat joined
vrurg releasable6: status 22:27
releasable6 vrurg, Next release in ≈1 day and ≈20 hours. 11 blockers. 59 out of 224 commits logged (⚠ 51 warnings)
vrurg, Details: gist.github.com/febfd495c76f979e73...afbeefac43
samcv woo release 22:43
23:44 vrurg left, vrurg joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!