MasterDuke what's a good way in nqp-land to know if something is an enum? 00:12
to be more precise, in World.nqp 00:14
timotimo you can .HOW things, can't you? 00:15
otherwise probably nqp::gethow
MasterDuke the .HOW.name is just the name of the enum
timotimo yeah, but what's the .HOW.HOW.name? :D 00:16
MasterDuke hm, i thought it did try that, but let me check
timotimo m: say True.HOW.HOW.name(True.HOW)
camelia Perl6::Metamodel::EnumHOW
MasterDuke well there you go. wonder what it was i tried... 00:17
AlexDaniel_ “Message body is not shown because it is too large.” *facepalm*
timotimo ;(
samcv should i just todo all jvm nqp failures? 00:26
y/n
timotimo modules.perl6.org now updated in about 55 minutes 00:27
so 20 minutes saved perhaps?
samcv AlexDaniel_, MasterDuke some of the waste that's in INTERPOLATE. i added some 'note's to nqp
gist.github.com/39c206ff7c45649e26...b9156b2323
'hi' ~~ /hi/ compared to my $var = 'hi'; 'hi' ~~ /$hi/ 00:28
3 vs 13 times called. the nqp !INTERPOLATE method
MasterDuke which is 13? 00:29
AlexDaniel_ 'hi' ~~ /$hi/
samcv ^
timotimo it looks like we can give Pithub an LWP::UserAgent object we create ourselves; if we initialize it with keep_alive => 10 or something we can save a bit of time spinning up and tearing down connections perhaps?
samcv where it outputs "literal␤" is in `method termish` btw 00:30
AlexDaniel_ samcv: by the way, why am I being pinged? What's so special about me in this particular case? :)
samcv idk
you are here
what, 20 seconds before i highlighted you :) 00:31
and you're often interested in text parsing/unicode things
AlexDaniel_ ah ok, I was just thinking that maybe something has to be bisected or whatever 00:32
samcv i've at least added to roast the issues between 'ignorecase' not working properly for INTERPOLATE. so that's decent. wasn't able to figure out the cause though. i wish i understood how nqp worked that way 00:33
may be issues with 'ignoremark' or 'ignorecase+ignoremark' but i have yet to find bigs with those when the /$thing/ is interpolated versus /not/. unless you can think up some (or remember) 00:34
MasterDuke i thought i fixed the ignore* in INTERPOLATE recently? 00:36
samcv well there are at least some bugs still
MasterDuke i just spotted a wild TimToady over in #perl6, he might have some good input 00:37
samcv github.com/perl6/roast/commit/946d2ff79 and also see what i wrote in rt.perl.org/Ticket/Display.html?id=132233
though i can't count on it. i remember it failing in all cases i tried *before* but now it only fails with interpolated haystack in cases i cannot determine (it fails in roast test) 00:38
so if you say you fixed an issue with interpolated haystack, then maybe i remember correctly and it used to *always* fail with interpolated haystack and only does sometimes. because that test succeeds when i do it in REPL but in roast, it fails
have not yet figured out WTF is going on 00:39
timotimo going to bed, seeya :)
MasterDuke c6aacafd1f6cbb069befd0236e9f246092e11640 should have fixed some things, but i guess not everything 00:40
samcv ah nice. so that was the thing that fixed some cases 00:42
though it is still broken on the nqp side
MasterDuke yeah, i don't think i looked at nqp at all when i was working on INTERPOLATE 00:43
samcv i'm guessing that may be the descrepency. since nqp seems totally broken. but on perl6 some cases it works some cases it does not with interpolate
though i mean with perl6 at least. it works and doesn't work with perl6. and i'm not sure when/why it works/doesn't just that somee cases work and some don't
but big ++ to making the situation better :) 00:44
Geth rakudo: MasterDuke17++ created pull request #1181:
Suggest enum values as types		 00:59
rakudo/nom: 12774237a6 | (Daniel Green)++ | src/Perl6/World.nqp
Suggest enum values as types

Implements RT #123926, e.g., `enum E <Foo Bar>; sub x(Floo) {}` now says
  `Invalid typename 'Floo' in parameter declaration. Did you mean 'Foo'?`		 01:21
rakudo/nom: f62ae60c48 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/Perl6/World.nqp
Merge pull request #1181 from MasterDuke17/RT123926

Suggest enum values as types
synopsebot RT#123926 [new]: rt.perl.org/Ticket/Display.html?id=123926 [BUG] LTA error message without Levenshtein distance suggestion when mistyping enum value in signature in Rakudo
rakudo: skids++ created pull request #1182:
Warn on typical precedence errors with infix:<..> (RT#127279)		 02:27
synopsebot RT#127279 [new]: rt.perl.org/Ticket/Display.html?id=127279 [BUG] Infinite loop with |"a".."z" 02:28
Geth rakudo/corresponding-starter: 279bae08a5 | (Aleks-Daniel Jakimenko-Aleksejev)++ | src/Perl6/Grammar.nqp
Mention corresponding starter in <, <<, « subscripts

Resolves part of RT #125641.
Wordy descriptions were substituted with actual symbols to make it more consistent with other error messages. Also, this way the error message is much shorter.		 02:35
synopsebot RT#125641 [new]: rt.perl.org/Ticket/Display.html?id=125641 [LTA] error messages when right bracket is missing: $<> , <>, () and so on
rakudo/corresponding-starter: 6542bb8032 | (Aleks-Daniel Jakimenko-Aleksejev)++ | 2 files
Mention corresponding starter for most of the things

This should handle quotes, parens, brackets, subscripts and many other things.
Resolves part of RT #125641.
Unfortunately I don't know how to print the actual starter, but as long as the line number is mentioned it's alright.
rakudo: AlexDaniel++ created pull request #1183:
“corresponding … was at” RT #125641		 02:36
AlexDaniel_ that's depressing 02:41
MasterDuke AlexDaniel_: did you happen to try with passing :cache(1) to lineof()? 02:47
AlexDaniel_ MasterDuke: no, I don't know what cache does 02:48
MasterDuke: fwiw this commit has no :cache also github.com/rakudo/rakudo/commit/38...eb2d50fef6 02:49
MasterDuke sometimes the line info has already been calculated and cached, it's a bunch faster
yeah, i added it more recently than that, when i got backtraces to print the actual filename+linenumber instead of the generated file 02:50
Geth roast: 3014999fcd | (Daniel Green)++ | S32-exceptions/misc.t
Add test for suggesting enum values as types

For RT #123926.		 02:52
synopsebot RT#123926 [new]: rt.perl.org/Ticket/Display.html?id=123926 [BUG] LTA error message without Levenshtein distance suggestion when mistyping enum value in signature in Rakudo
AlexDaniel_ MasterDuke: oh, it was you who did that 02:54
awesome
MasterDuke yeah, almost a year ago now 02:56
AlexDaniel_ MasterDuke: that seems to be working fine 02:58
MasterDuke cool
Geth rakudo/corresponding-starter: a462d0a26e | (Aleks-Daniel Jakimenko-Aleksejev)++ | src/Perl6/Grammar.nqp
Cache some line numbers

  MasterDuke++
  irclog.perlgeek.de/perl6-dev/2017-...i_15269029		 03:00
rakudo: MasterDuke17++ created pull request #1184:
Use cached line numbers for token label		 03:23
samcv going to improve the error when you try and use whitespace with a combiner as a separator 05:33
to try and close this LTA RT
[Tux] This is Rakudo version 2017.09-230-gf62ae60c4 built on MoarVM version 2017.09.1-576-g64b5dc03 07:57
csv-ip5xs 1.257 - 1.360
test 9.982 - 10.344
test-t 3.219 - 3.236
csv-parser 0.883 - 0.901
smls Does anyone know how variables interpolated into regexes like / @foo / are parsed by Rakudo? 08:47
P6Regex'es `metachar:sym<var>` looks like it only catches backreferences, not external variables: github.com/perl6/nqp/blob/576d78ee...r.nqp#L334 08:48
ah, there's github.com/rakudo/rakudo/blob/f62a....nqp#L5314 08:52
buggable New CPAN upload: IRC-Client-Plugin-Github-0.1.2.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.2.tar.gz
New CPAN upload: Config/Config-1.2.1.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...2.1.tar.gz 09:02
New CPAN upload: Config-Parser-toml/Config-Parser-toml-1.0.0.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...0.0.tar.gz 09:12
New CPAN upload: Config-Parser-yaml/Config-Parser-yaml-1.0.0.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...0.0.tar.gz
New CPAN upload: Hash-Merge/Hash-Merge-0.1.0.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.0.tar.gz
New CPAN upload: MPD-Client/MPD-Client-0.1.2.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.2.tar.gz
Geth roast: Tyil++ created pull request #342:
WIP: Expand file tests		 09:36
lizmat Files=1227, Tests=75600, 346 wallclock secs (14.70 usr 5.24 sys + 2419.35 cusr 218.79 csys = 2658.08 CPU) 09:44
tyil github.com/perl6/roast/pull/342/fi...d3a273fR54 this returns a Failure, not a Bool at this moment, should I adapt the test or is the Failure return incorrect? 09:46
Geth rakudo/nom: 26bdc95c78 | skids++ | 2 files
Warn on typical precedence errors with infix:<..> (RT#127279)

Alerts user to potential accidental use of prefix:<|> or
   prefix:<~> on a range start value, when they are usually intended
   to apply to the entire range, and encourages use of parens.		 10:41
rakudo/nom: cb9d2e040d | lizmat++ (committed using GitHub Web editor) | 2 files
Merge pull request #1182 from skids/rt127279

Warn on typical precedence errors with infix:<..> (RT#127279)
synopsebot RT#127279 [open]: rt.perl.org/Ticket/Display.html?id=127279 [BUG] Infinite loop with |"a".."z"
rakudo/nom: 4868e9271a | (Zoffix Znet)++ | 2 files
Remove dead code for `use v5`

Dies while complaining about missing `Perl5` module. A thought occured to make it load Inline::Perl5, but felt like useless magic, so I removed the code entirely and just make it complain there's no compiler for v5, like it'd do for `use v6.z`
Fixes RT#130834: rt.perl.org/Ticket/Display.html?id=130834		 11:50
synopsebot RT#130834 [open]: rt.perl.org/Ticket/Display.html?id=130834 Is `use v5` still a thing?
Geth rakudo/full-make-test: 50 commits pushed by (Zoffix Znet)++, skids++, (Elizabeth Mattijsen)++, (Jonathan Worthington)++, (Patrick Spek)++, (Samantha McVey)++, (Daniel Green)++
review: github.com/rakudo/rakudo/compare/c...54d5096732		 12:20
rakudo/full-make-test: 6757f93473 | (Zoffix Znet)++ | t/02-rakudo/01-is_approx.t
Fix up t/02-rakudo/01-is_approx.t tests

The author of planck tests wrote them without any regard to `e-34` part of the number and adjusted the numbers in the decimal portion alone, expecting changes to be of 1e-5 magnitude when they were actually of 1e-39 magnitude.
Fix by shifting the original value by expected differences instead.		 12:48
roast: 4f224ec562 | (Patrick Spek)++ (committed by Zoffix Znet) | S32-io/file-tests.t
Expand file tests (#342)

  * Use subtests
  * Add tests for .d
  * Add tests for IO.l
... (34 more lines)		 12:51
Skarsnik Hello :) 12:59
Geth rakudo/full-make-test: 0d5bb9097a | (Zoffix Znet)++ | t/02-rakudo/dump.t
Fudge failing tests

Not sure there's any point in fixing this, since it's not part of the language and we already got `dd` and language-wise we got `.perl`
Skarsnik How the squashaton going? 13:00
Geth rakudo/full-make-test: 93c4555db3 | (Zoffix Znet)++ | t/02-rakudo/dump.t
Unfudge now-passing test
Skarsnik squashable6, status 13:01
squashable6 Skarsnik, 🍕🍕 SQUASHathon is in progress! The end of the event in ≈22 hours. See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
Skarsnik, Log and stats: gist.github.com/276df459191e96cb5c...dcfdc60aef
Zoffix .tell bartolin I think Proc::Async is already implemented on JVM, so maybe this ticket can be closed and all the fudges unfudged? rt.perl.org/Ticket/Display.html?id=126524 13:06
yoleaux Zoffix: I'll pass your message to bartolin.
12:55Z <tyil> Zoffix: thanks for the tip :>
Zoffix Hm, looks like we're running with zero REPL tests whatsoever. 13:11
Because all the tests are run as non-TTY and it just evals the code instead of firing up REPL. 13:12
Do we have a way to make Proc pretend STDIN it as a tty? 13:13
Zoffix doesn't know anything about TTY stuff
Geth roast: c4d7af0c4d | (Zoffix Znet)++ | 2 files
Toss is_run_repl helper routine

It's not used anywhere and all but the basic sanity REPL tests have been moved to rakudo's test suite		 13:15
rakudo/full-make-test: 78eca2141d | (Zoffix Znet)++ | 2 files
Start making repl tests work
 13:25
Zoffix After a bit of googling don't see a sure-fire portable way to fake a TTY, so I'm gonna add a compiler command line flag to force interactive mode even when STDIN aint a TTY 13:27
dogbert17: doubt it. I think it was already fixed before. Bisectable can likely tell 13:32
"Stage parse : 94.820" man. And I used to feel sad about low 80s :( this box sucks
lizmat hmmm... --profile-compile on the setting compile does not appear to generate a profile :-( 13:39
timotimo does it finish? 13:44
dogbert2 Zoffix: can you do IO stuff with that bot? 13:46
bistect: my $f = open("test.txt", :rw); $f.write("hello".encode); $f.seek(0); say $f.read(5).decode 13:47
bisect: my $f = open("test.txt", :rw); $f.write("hello".encode); $f.seek(0); say $f.read(5).decode
bisectable6 dogbert2, Bisecting by output (old=2015.12 new=4868e92) because on both starting points the exit code is 1
dogbert2, bisect log: gist.github.com/3d98e8c1cd197ef3b3...a006db2cd7
dogbert2, (2015-12-25) github.com/rakudo/rakudo/commit/07...dc61f84053
dogbert2, The result looks a bit unrealistic, doesn't it? Most probably the output is different on every commit (e.g. ｢bisect: say rand｣)
Zoffix usually creates all the files in `/tmp` to not pollute dirs with test files
timotimo yeah, you have to do it in /tmp, otherwise no write access 13:48
Zoffix bisect: my $f = open("/tmp/dasdasdasdastest.txt", :rw); $f.write("hello".encode); $f.seek(0); say $f.read(5).decode
bisectable6 Zoffix, On both starting points (old=2015.12 new=4868e92) the exit code is 0 and the output is identical as well
Zoffix, Output on both points: «hello»
Zoffix releases: my $f = open("/tmp/dasdasdasdastest.txt", :rw); $f.write("hello".encode); $f.seek(0); say $f.read(5).decode
c: releases my $f = open("/tmp/dasdasdasdastest.txt", :rw); $f.write("hello".encode); $f.seek(0); say $f.read(5).decode
committable6 Zoffix, ¦releases (23 commits): «hello»
Zoffix c: releases my $f = open("/tmp/dasdasdasdastest.txt", :r, :w); $f.write("hello".encode); $f.seek(0); say $f.read(5).decode
committable6 Zoffix, ¦releases (23 commits): «hello» 13:49
dogbert2 so it was fixed ages ago then
ok, if I find some tests in roast I'll close it 13:51
lizmat timotimo: yes, but no file :-( 13:54
timotimo huh.
Geth roast: skids++ created pull request #343:
Add tests for new range precedence worries (RT#127279)		 14:02
synopsebot RT#127279 [open]: rt.perl.org/Ticket/Display.html?id=127279 [BUG] Infinite loop with |"a".."z"
Geth roast: 9d56ba1e58 | skids++ | S03-operators/range.t
Add tests for new range precedence worries (RT#127279)
roast: 6e9bab5de4 | skids++ (committed using GitHub Web editor) | S03-operators/range.t
Merge pull request #343 from skids/rt127279

Add tests for new range precedence worries (RT#127279)
Zoffix heh parse stage 96s now! :) It goes up by 2s every time I run it ^_^ 14:05
timotimo oops 14:06
Zoffix What's the standard practice to rebuild nqp when I make changes to it? 14:07
timotimo make install; cd ../rakudo; make clean; make install
Zoffix thanks
timotimo just what i use 14:08
timotimo is taking a heap snapshot of the heap snapshot analyzer analyzing a heap snapshot
Geth roast: e3e1f75cc5 | usev6++ | 6 files
[jvm] Remove fudges for RT #126524

Proc::Async has been implemented for the jvm backend; jnthn++.
Unfortunately, some of the fudged tests still fail. I updated the fudge reasons accordingly.		 14:11
synopsebot RT#126524 [open]: rt.perl.org/Ticket/Display.html?id=126524 [JVM] Proc::Async NYI
Zoffix \o/
bartolin Zoffix++ I'm gonna close the issue
yoleaux 13:06Z <Zoffix> bartolin: I think Proc::Async is already implemented on JVM, so maybe this ticket can be closed and all the fudges unfudged? rt.perl.org/Ticket/Display.html?id=126524
bartolin sadly it's not possible to run spectest on the jvm backend atm due to a severe memory leak (RT #132104), but i looked at the aboved tests in isolation 14:13
synopsebot RT#132104 [open]: rt.perl.org/Ticket/Display.html?id=132104 [JVM] EvalServer seems to leak memory
timotimo i've already filtered to only major collections, but it's already past 4 gigs of heap snapshot data
Geth rakudo/nom: f59b2887ad | (Zoffix Znet)++ (committed using GitHub Web editor) | lib/NativeCall.pm6
Fix precedence error; RabidGravy++
 14:14
lizmat afk for some 2049 &
timotimo cool
Zoffix bartolin: sucks about `doesn't-hang` errors. I'm trying to make repl tests in rakudo's test suite run and some of them use it :/ And there's no fudger for `make test` 14:15
bartolin Zoffix: I'll try to take a closer look at that error (haven't seen that specific error before). Please just go ahead and use it, anyway! 14:16
Zoffix OK
By the sounds of it. Proc::Async.kill is not implemented on JVM 14:17
(that's what doesn't-hang uses under the hood)
bartolin yeah, jnthn++ did map that op with nqp commit 76098315da, but it's NYI 14:20
timotimo oh, looks like i was running a moar without the filter functionality %) 14:23
Zoffix man, this is taking too long. Still have trouble rebuilding nqp :/ 14:35
"/home/zoffix/CPANPRC/rakudo/install/bin/nqp-m: 2: exec: ../install/bin/moar: not found"
$ type build-nqp
build-nqp is aliased to `cd nqp; perl Configure.pl --gen-moar --backend=moar --prefix=../install; make; make install; cd ..; make clean; make install'
timotimo perhaps the cd fails? :\ 14:38
Zoffix timotimo: do I need to manually build MoarVM too then? 14:41
timotimo shouldn't have to
as long as the moar binary is in that prefix it should be fine 14:42
Zoffix this is getting on my tits 14:43
Zoffix runs `rm -fr install; build-rakudo` 14:45
timotimo: how do you get it to the point that all you do is `make install; cd ../rakudo; make clean; make install` ? What's the setup with a fresh checkout of everythinng? 14:49
timotimo i git clone moar, nqp, and rakudo to be side-by-side rather than nested, and the install/ folder which is my prefix lives as another sibling 14:51
Geth nqp: c20643c97c | (Samantha McVey)++ | docs/ops.markdown
[docs] Document `ordbaseat` op in docs/ops.markdown
nqp: 0bed230bda | (Samantha McVey)++ | src/HLL/Grammar.nqp
Improve error message when space with combiners is encountered

New message: q ̌̌(hi) Whitespace character ‘SPACE’ (0x20) with 2 combining characters is not allowed as a delimiter at line 2, near " ̌̌(hi)\n"
Zoffix ok, I'll try that
Geth rakudo/nom: ba2d858fb2 | (Zoffix Znet)++ (committed using GitHub Web editor) | lib/NativeCall.pm6
Fix precedence error; RabidGravy++
Zoffix gah, my terminal's busted now due to that commit message 14:53
samcv oh?
Zoffix Yeah, crappy unicode support
samcv can't handle two combining carons or something. or maybe that it's on a space. it shows up in hexchat as on the parenthesis for me 14:54
Geth nqp: 108db21525 | (Zoffix Znet)++ | src/HLL/Compiler.nqp
Remove trailing whitespace
 14:58
Zoffix ZofBot: two combining carrots! 15:03
ZofBot Zoffix, In the absence of that information, the tool can choose either to download and use the compunit directly, or the tool can proceed in ignorance
travis-ci NQP build failed. Samantha McVey 'Improve error message when space with combiners is encountered 15:06
travis-ci.org/perl6/nqp/builds/284691199 github.com/perl6/nqp/compare/576d7...ed230bdaa0
NQP build failed. Zoffix Znet 'Remove trailing whitespace' 15:12
travis-ci.org/perl6/nqp/builds/284693561 github.com/perl6/nqp/compare/0bed2...8db2152508
Zoffix `make clean; cd nqp; make install; cd ..; make install` does the trick with the standard nested installation 15:14
Geth nqp: c38cfe87f5 | (Zoffix Znet)++ | src/HLL/Compiler.nqp
Implement --force-stdin-eval-mode command line arg

  - Takes 'interactive' and 'non-interactive' as valid values
  - Provides the means to switch between REPL and plain eval STDIN
   without depending on whether STDIN is a TTY or not. At the
   very least, this lets us easily write REPL tests.		 15:21
dogbert2 bisect: my @isprime = False,False; say (for 1 .. 10 -> $i { $i if @isprime[$i] //= so $i %% none 2 ...^ * > $i.sqrt.floor; }) 15:28
bisectable6 dogbert2, Bisecting by output (old=2015.12 new=f59b288) because on both starting points the exit code is 0
dogbert2, bisect log: gist.github.com/80f33c40afa832a376...f73a7626be
dogbert2, (2017-01-13) github.com/rakudo/rakudo/commit/d1...dcb4918120
Skarsnik hm 15:31
what that resolved at some point?
m: role A {method foo(){say "a"}}; class B does A { method foo(Str $a) {say "b"}}; my B $b .= new; $b.foo
camelia Too few positionals passed; expected 2 arguments but got 1
in method foo at <tmp> line 1
in block <unit> at <tmp> line 1
dogbert2 the bisect above is from RT #126415 15:35
synopsebot RT#126415 [new]: rt.perl.org/Ticket/Display.html?id=126415 [BUG] WhateverCode loses its mind inside //= 15:36
travis-ci NQP build failed. Zoffix Znet 'Implement --force-stdin-eval-mode command line arg 15:37
travis-ci.org/perl6/nqp/builds/284700960 github.com/perl6/nqp/compare/108db...8cfe87f54e
Geth rakudo/nom: 5f6896bde8 | (Timo Paulssen)++ | src/core/Str.pm
use nqp::codes op to speed up Int

not a vast improvement, but having one allocation less is a nice bonus.		 16:09
rakudo: jonathanstowe++ created pull request #1185:
Add tests to cover a common NC usage		 16:24
rakudo/nom: d8a20d1134 | (Jonathan Stowe)++ | 2 files
Add tests to cover a common NC usage

C functions that consistently take some struct or pointer as the first argument are declared as methods of that CStruct or CPointer		 16:25
rakudo/nom: 64d41ffa11 | (Zoffix Znet)++ (committed using GitHub Web editor) | 2 files
Merge pull request #1185 from jonathanstowe/nom

Add tests to cover a common NC usage
roast: 5f64bc0c6e | (Zoffix Znet)++ | S03-operators/range.t
Fix missing `use lib`
 16:36
rakudo/nom: 4 commits pushed by (Aleks-Daniel Jakimenko-Aleksejev)++ 16:39
roast: 85b9380acd | (Zoffix Znet)++ | S02-lexical-conventions/comments.t
Remove incorrect test

Per samcv++ from github.com/perl6/nqp/commit/02a426e0e6
Remove support for ornate parenthesis from quoting constructs Reasons:
  * These are not recognized by Unicode as being brackets
... (6 more lines)		 16:43
roast/6.c-errata: d78ab87a65 | (Zoffix Znet)++ | S02-lexical-conventions/comments.t
Remove incorrect test

Per samcv++ from github.com/perl6/nqp/commit/02a426e0e6
Remove support for ornate parenthesis from quoting constructs Reasons:
  * These are not recognized by Unicode as being brackets
... (6 more lines)
rakudo/nom: 9ce896d8bf | (Zoffix Znet)++ | tools/build/NQP_REVISION
Bump NQP
 16:46
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....5-gc38cfe8
20518454ef | (Zoffix Znet)++ | src/Perl6/Compiler.nqp

Since recently-ish, if STDIN is not a TTY, we've treated STDIN as simply a file with code to eval, without starting REPL.
rakudo/nom: 5c7bbea0a9 | (Zoffix Znet)++ | src/Perl6/Compiler.nqp
Add --force-stdin-eval-mode opt to usage msg
 17:04
RabidGravy Ah, that very thing there is probably why the fdreopen fd 0 to a listening socket isn't working 17:06
Zoffix doubts it 17:07
RabidGravy why?
Zoffix If you're talking about --force-stdin-eval-mode. That's only used when deciding whether to start the REPL or just eval code from STDIN without REPL machinery
(without the option, this is decided on whether STDIN is a TTY; starting REPL when it is) 17:08
RabidGravy no I'm talking about " if STDIN is not a TTY, we've treated STDIN as simply a file with code to eval"
Zoffix *based on
RabidGravy: that's only when otherwise we'd start REPL
RabidGravy: if you passed -e '' or gave some file to read code form, this codepath isn't used
RabidGravy Oh okay 17:09
misunderstanding
moritz --force-interactive? 17:12
sorry, I wasn't there for the bikeshedding, so I should shut up 17:13
Zoffix moritz: I thought of that originally, but figured if we're adding an option to force REPL on non-TTY displays, may as well add one to force non-interactive mode on TTY displyas
moritz: there wasn't any bikeshedding. I just added it, 'cause I need to fix REPL tests :) If there's a better name we can change it 17:14
Zoffix leaves for a quick jam session 17:15
RabidGravy Mmm jam 17:19
Zoffix "--repl-mode" ? The only thing is that suggests the non-interactive stuff still loads REPL.pm but it doesn't 17:30
Don't like `--force-stdin-eval-mode` no more 17:38
Gonna swap to `--repl-mode` later, unless someone suggests better name 17:39
Skarsnik hm thoughs on gist.github.com/Skarsnik/3fc513cbd...1a64286f51 18:15
AlexDaniel_ q.viva64.com/ # that's kinda a product ad, but fun anyway 18:18
travis-ci Rakudo build passed. Zoffix Znet 'Add --force-stdin-eval-mode opt to usage msg' 18:19
travis-ci.org/rakudo/rakudo/builds/284733899 github.com/rakudo/rakudo/compare/2...7bbea0a920
[Coke] waves. 18:26
Skarsnik the 5th one is dubious
A buffer containing private data should be erased before function returns. The memset() function is used for this task, but it is incorrect. As compiler is aware that after calling memset() the 'final' buffer will not be utilized anymore, it is possible that compiler will remove memset() call altogether. 18:27
what can lead a compilater to remove a function call, that make no sense
AlexDaniel_ shrugs 18:34
Skarsnik that the kind of reason I don't like coding in C++ 18:39
You need to much of compiler knowledge
timotimo C and C++ compilers have heavy optimizations regarding memset and friends 18:52
they will even recognize a for loop that is equivalent to memset and turn it into memset instead
Skarsnik I wonder what this tool could give on MoarVM Code 18:54
timotimo i think i've done this quiz before 18:55
AlexDaniel_ fwiw I gonna kill anybody who merges my last 5 tickets that I've intentionally submitted separately 19:04
Zoffix :o 19:06
Skarsnik it's fail.6lang.org now?
Zoffix Skarsnik: no, fail.rakudo.org 19:08
err
Skarsnik: no, fail.rakudo.party
ZofBot: too many domain names
ZofBot Zoffix, However, if it is completely empty or consists of a single list, the first element of which is either a hash or a pair, it is executed immediately to compose a "Hash" object
Skarsnik ZofBot, fail 19:11
ZofBot Skarsnik, It has been argued that "
Skarsnik ZofBot, RT 19:12
ZofBot Skarsnik, 2 # false, equivalent to 1 <= 2
Skarsnik what 19:13
AlexDaniel_ Skarsnik: it is retarded
Skarsnik how the gumbo crash perform with the nc-jit call now btw? x) 19:16
AlexDaniel_ I haven't tried it, should I? 19:17
Skarsnik probably change nothing 19:18
AlexDaniel_ Skarsnik: ok, that's bad 19:23
Skarsnik: it's no longer crashing
Skarsnik hm 19:24
AlexDaniel_ but I really don't think that the issue is resolved, so we just masqueraded it
Skarsnik: 50000 iterations and nothing 19:27
ok I can bisect it for fun 19:28
Geth roast: 34577134e0 | TimToady++ | S32-exceptions/misc.t
overspecific test disallows eject in correct spot
dogbert2 would anyone be terribly upset if I enabled the test mentioned in RT #122951 and then closed the issue with an appropriate comment? 19:29
synopsebot RT#122951 [open]: rt.perl.org/Ticket/Display.html?id=122951 [BUG] negative lookahead doesn't LTM properly
AlexDaniel_ dogbert2: does it pass on HEAD? 19:30
dogbert2 m: say "abcde" ~~ / ab <![e]> cde | ab.. /
camelia ｢abcd｣
dogbert2 wait a sec, it doesn't work does it 19:32
dogbert2 must learn the art of reading
Geth rakudo/nom: 5747bc7121 | TimToady++ | src/Perl6/Grammar.nqp
deconfuse statement modifier errors

Catch missing expressions on statement modifiers earlier so that the tests eat_terminator are only for statement controls. Also, move the cursor to the front of the cursor controls to indicate where the semicolon is actually perceived to be missing if we report that. Additionally, we change the kok check to use .panic instead of .sorry to avoid multiple errors resulting from repeated terminator checks that use statement modifier keywords for terminators.
Fixes #125596 and #125674.		 19:34
synopsebot RT#125596 [new]: rt.perl.org/Ticket/Display.html?id=125596 [LTA] error message with if; and unless; without conditions
RT#125674 [new]: rt.perl.org/Ticket/Display.html?id=125674 [LTA] error message prints wrong eject position (if True if { };)
AlexDaniel_ dogbert2: once you learn how to read, please teach me too
dogbert2 :)
anyway, this is not something we can close atm 19:36
another one then RT #126415, according to bisectable it was fixed with github.com/rakudo/rakudo/commit/d1...dcb4918120 19:37
synopsebot RT#126415 [new]: rt.perl.org/Ticket/Display.html?id=126415 [BUG] WhateverCode loses its mind inside //= 19:38
dogbert2 jnthn: added tests for that commit here, github.com/perl6/roast/commit/e797...1a05fe8a46
but do they really cover the RT
Skarsnik btw last week I try to look at rt.perl.org/Ticket/Display.html?id=127345 but I was not sure where/how to add the code that annotate the instruction with the line (since MoarVM does not get it) 19:39
timotimo dogbert2: fwiw, the output abcd is the correct one for that alteration 19:40
dogbert2: the negative lookahead isn't declarative, so the longest declarative prefix is 2 letters on the LHS and 4 on the RHS 19:41
Geth roast: a0b678199c | (Moritz Lenz)++ | S32-num/rat.t
RT #126101: Another test for 0.9999...
synopsebot RT#126101 [open]: rt.perl.org/Ticket/Display.html?id=126101 [BUG] 0.lotsof9s badly approximated 19:42
dogbert2 timotimo: so the report and test is bogus then ?
or did things change between Parrot and Rakudo?
timotimo whoa 19:43
moritz there is no parrot.
and chromatic is its prophet.
timotimo there was a quote from the synopses
i did not know about that detail
.o( the regex optimizer could throw that match out )
(i mean the negative lookahead) 19:44
dogbert2 moritz: what do you think about RT #122951, do you cover this in yout new book?
synopsebot RT#122951 [open]: rt.perl.org/Ticket/Display.html?id=122951 [BUG] negative lookahead doesn't LTM properly
moritz dogbert2: it doesn't go into that much detail on LTM 19:45
dogbert2 it was worth a shot :)
moritz it's suprising how little fits into ~150 pages :/
but I trust pmichaud's judgement on this one
dogbert2 which means that the problem still exist even though Parrot does not 19:46
moritz it's not as if parrot dictated Rakudo regex semantics at the time this bug was submitted 19:47
dogbert2 interesting, I know to little about the history of these things 19:49
timotimo the regex impl was probably pretty much 1:1 ported to jvm and moar 19:50
AlexDaniel_ c: HEAD gist.githubusercontent.com/AlexDan...bocrash.p6
committable6 AlexDaniel_, Successfully fetched the code from the provided URL.
AlexDaniel_, ¦HEAD(5747bc7): «0␤MoarVM panic: Internal error: zeroed target thread ID in work pass «exit code = 17»»
AlexDaniel_ Skarsnik: ummmm
timotimo that's not nice 19:51
AlexDaniel_ Skarsnik: I don't know? How come it happens on the first iteration?
dogbert2 is that Gumbo thing still a problem
AlexDaniel_ dogbert2: not entirely sure. locally it works fine but blows up immediately on committable
Skarsnik does not crash on my vm 19:52
AlexDaniel_ actually, maybe the output is cached so we only see 0 but there was more stuff?
dogbert2 which P6 version does c: use?
Skarsnik but crash on my dedicated 64bits server
AlexDaniel_ dogbert2: HEAD(5747bc7)
Skarsnik AlexDaniel, did it fail on fetching the website? 19:53
AlexDaniel_ Skarsnik: no because it printed 0
hm or… I don't know
timotimo: RT #131003 19:54
synopsebot RT#131003 [open]: rt.perl.org/Ticket/Display.html?id=131003 [ANNOYING][SEGV] Heap corruption when using Gumbo
timotimo i know about this bug :(
hm, i wonder if nativecast is involved there
Skarsnik could be 19:55
timotimo i had some code where i nativecast one kind of CArray or something to another and it proceeded to double-free the backing array
Skarsnik I had issue trying to replace the nativecast with a typed pointed object and doing unref on it but it corrupted the addre 19:56
timotimo m: use NativeCall; sub doit { my $foo = CArray[int64].new(1, 2, 3, 4, 5, 6, 7, 8, 9, 10); my $bar = nativecast(CArray[int8], $foo); 1 }; for ^100 { doit }
camelia ( no output )
timotimo well, that's not how
m: use NativeCall; sub malloc(int32 $size) is native(Str) returns CArray[int8] {*}; sub doit { my $foo = malloc(128); my $bar = nativecast(CArray[int64], $foo); 1 }; for ^100 { doit } 19:57
camelia ( no output )
timotimo well, that's not how i guess?
m: use nqp; use NativeCall; sub malloc(int32 $size) is native(Str) returns CArray[int8] {*}; sub doit { my $foo = malloc(128); my $bar = nativecast(CArray[int64], $foo); 1 }; for ^100 { doit; nqp::force_gc }
camelia ( no output )
timotimo m: use nqp; use NativeCall; sub doit { my $foo = CArray[int64].new(1, 2, 3, 4, 5, 6, 7, 8, 9, 10); my $bar = nativecast(CArray[int8], $foo); 1 }; for ^100 { doit; nqp::force_gc }
camelia ( no output )
AlexDaniel_ hm but it doesn't crash here so I really don't know 19:58
AlexDaniel_ wonders if [ANNOYING] should be capitalized as [aNNOYING] 19:59
timotimo don't make it even harder to work on this kind of bug :) 20:01
Skarsnik I wonder if my typered branch of gptrixie still make MoarVM segfault at compile/init time xD 20:05
moritz ‘789’.comb(/ . {say $/} <!> /)'
m: ‘789’.comb(/ . {say $/} <!> /)'
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3‘789’.comb(/ . {say $/} <!> /)7⏏5'
expecting any of:
infix
infix stopper
statement end
statement modifier…
AlexDaniel_ trailing '
moritz ‘789’.comb(/ . {say $/} <!> /)
m: ‘789’.comb(/ . {say $/} <!> /)
camelia ｢7｣
｢8｣
｢9｣
AlexDaniel_ oh, I remember that
moritz AlexDaniel_: thanks
yes, fun one
AlexDaniel_ wasn't it closed?
or is it TESTNEEDED? 20:06
moritz the latter
Geth roast: afa1f74636 | (Moritz Lenz)++ | S03-operators/repeat.t
RT #121327: tests for scoping bug in xx
 20:08
synopsebot RT#121327 [open]: rt.perl.org/Ticket/Display.html?id=121327 [TESTNEEDED] Scoping bug with `xx`
AlexDaniel_ errr that's not the ticket I was thinking about 20:09
right, because it isn't. 20:11
nvm
Geth roast: 005b4f76f7 | (Moritz Lenz)++ | S32-str/comb.t
RT #130186: $/ inside comb
synopsebot RT#130186 [new]: rt.perl.org/Ticket/Display.html?id=130186 Match object inside .comb is weird (‘789’.comb(/ . {say $/} <!> /)) 20:12
Skarsnik releasable6, status
releasable6 Skarsnik, Next release in 13 days and ≈22 hours. 1 blocker. Changelog for this release was not started yet
Skarsnik, Details: gist.github.com/a2b5146055cf744685...4f181faf04
Geth rakudo/nom: 5144216f6e | (Elizabeth Mattijsen)++ | 2 files
Cache types we always need in the CompilerServices

  - this appears to make the setting compilation a bit faster
  - but getting timings on that is always difficult :-(		 20:14
dogbert2 interesting, the Gumbo example seems to work on my system as well 20:22
Skarsnik test with a previous release maybe? 20:23
dogbert2 ok have 2017.07 available 20:24
Skarsnik I use linuxfr.org to test, it crash faster ^^ 20:25
AlexDaniel_ Skarsnik: still does not crash 20:27
Geth nqp/jvm_killprocasync: 11ef6ea37d | usev6++ | 10 files
[jvm] Update stage0 for modified op killprocasync
 20:28
nqp/jvm_killprocasync: 7ca352fdf5 | usev6++ | 3 files
[jvm] Primitive implementation of killprocasync
Skarsnik it does stuff thou?
AlexDaniel_ facepalm 20:31
curl: command not found
bartolin ^^ it would be nice to get a feedback for that branch (jvm_killprocasync). it's rudimentary but it makes "doesn't-hang" usable -- though not with the default wait of 1.5 sec *g*
AlexDaniel_ Skarsnik: OK nevermind, no change, it still crashes 20:32
committable6 is right, I'm wrong
Skarsnik yay? 20:33
dogbert2 works with 2017.07 as well and also with linuxfr.org
AlexDaniel_ … yeeeeeeeeeeey… :'(
dogbert2: nope. Can't be. Do you have curl installed?
dogbert2 yes
AlexDaniel_ dogbert2: or do you have 32-bit system?
dogbert2 and yes :)
have you tried disabling the jit? 20:34
AlexDaniel_ ok, that we knew already… yeeeey… :'(
Geth rakudo/nom: a09f5f2194 | (Elizabeth Mattijsen)++ | 4 files
Set up Builtin encodings at setting compile time

  - so we don't need to do this each time on startup
  - only saves maybe 1 msec of startup
  - but it makes future optimizations opening STDIN/OUT/ERR easier
Skarsnik it's an old crash
bartolin .tell jnthn if you have a few minutes to look at my try to implement killprocasync on jvm: github.com/perl6/nqp/tree/jvm_killprocasync -- when you stubbed the op, you planned to return $RT_OBJ. do you remember what you expected the op to return?
yoleaux bartolin: I'll pass your message to jnthn.
AlexDaniel_ dogbert2: yea, doesn't help
dogbert2 annoying 20:35
Skarsnik I should have a closer look at the bt of other threads
for this
dogbert2 I can get it to fail, if run under valgrind with MVM_SPESH_BLOCKING=1 20:44
Skarsnik interesting
Geth roast: 31506095f1 | usev6++ | 2 files
[jvm] Update fudge messages: killprocasync NYI
 20:45
dogbert2 but it won't fail outside valgrind with that same flag set, odd 20:46
Skarsnik for me it does not crash with efence before filling the whole memory x) 20:47
same with valgrind
AlexDaniel_ … and people argue whether [ANNOYING] tag is needed… 20:48
Skarsnik [Alexdanielable] 20:49
Geth rakudo/nom: c7a82d451d | usev6++ (committed by Aleks-Daniel Jakimenko-Aleksejev) | src/Perl6/Grammar.nqp
Try harder to diagnose problem

instead of reporting 'Malformed my'
Fixes RT #125902		 20:53
synopsebot RT#125902 [open]: rt.perl.org/Ticket/Display.html?id=125902 [LTA] error message “Type 'Str' is not declared” (my Str where 'foo' $test;)
Geth rakudo/nom: ac97a40161 | (Aleks-Daniel Jakimenko-Aleksejev)++ | src/Perl6/Grammar.nqp
Fix complaints about existing types

Call typo_typename only if <typename> or <value> failed, not if something else went wrong.
Resolves ½ of RT #127100.		 21:58
synopsebot RT#127100 [new]: rt.perl.org/Ticket/Display.html?id=127100 [LTA] error message when specifying return type before parameters ( --> Bool, Int $x, Int $y )
Geth rakudo/nom: bb5583aeb7 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm
Streamline initialization of $*IN/$*OUT/$*ERR

  - do as much as possible during setting compilation
  - only plug in the handle/encoder/decoder at runtime
  - shaves off about 1.5% from bare startup time		 22:12
rakudo: dwarring++ created pull request #1186:
implement typed pointer increment and array dereference salortiz++		 22:36
travis-ci Rakudo build errored. Zoffix Znet 'Merge pull request #1185 from jonathanstowe/nom 22:41
travis-ci.org/rakudo/rakudo/builds/284721668 github.com/rakudo/rakudo/compare/5...d41ffa1104
buggable [travis build above] ✓ All failures are due to: missing build log (1 failure).
Geth rakudo/nom: 64b001a146 | (Aleks-Daniel Jakimenko-Aleksejev)++ | src/Perl6/Grammar.nqp
Awesome error message for RT #127100

If you can do this:
   sub foo($x, --> 42) {}
Then why not this? ... (6 more lines)		 22:49
synopsebot RT#127100 [new]: rt.perl.org/Ticket/Display.html?id=127100 [LTA] error message when specifying return type before parameters ( --> Bool, Int $x, Int $y )
buggable New CPAN upload: Config/Perl6/Config-1.2.1.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...2.1.tar.gz 23:02
New CPAN upload: Config-Parser-toml/Perl6/Config-Parser-toml-1.0.0.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...0.0.tar.gz
New CPAN upload: Config-Parser-yaml/Perl6/Config-Parser-yaml-1.0.0.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...0.0.tar.gz
New CPAN upload: Hash-Merge/Perl6/Hash-Merge-0.1.0.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.0.tar.gz
New CPAN upload: MDP-Client/Perl6/MPD-Client-0.1.2.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.2.tar.gz
Geth nqp: 943f7f757c | (Zoffix Znet)++ | src/HLL/Compiler.nqp
Rename `--force-stdin-eval-mode` command line opt

To `--repl-mode`		 23:43
rakudo/nom: 93e599dbf2 | (Zoffix Znet)++ | tools/build/NQP_REVISION
Bump NQP
 23:44
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....6-g943f7f7
de0533c4d4 | (Zoffix Znet)++ | src/Perl6/Compiler.nqp

To incorporate renaming of `--force-stdin-eval-mode` command line opt to `--repl-mode`
rakudo/full-make-test: 19 commits pushed by (Aleks-Daniel Jakimenko-Aleksejev)++, (Zoffix Znet)++, (Timo Paulssen)++, (Jonathan Stowe)++, TimToady++, (Elizabeth Mattijsen)++, usev6++
review: github.com/rakudo/rakudo/compare/7...87dc51f514		 23:45
travis-ci NQP build failed. Zoffix Znet 'Rename `--force-stdin-eval-mode` command line opt 23:57
travis-ci.org/perl6/nqp/builds/284860386 github.com/perl6/nqp/compare/c38cf...3f7f757c9b

Powered by Cro and the Raku® Programming Language.

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

Thank you!