Geth rakudo: 31bf3a5907 | (Aleks-Daniel Jakimenko-Aleksejev)++ | 2 files
Log all changes (+ announcement)

Deliberately not logged:
44ff0a2a 494ffb6a b93656d8 42af2a35 01ce5632 07f30e4b 1c649461 af4d6b46 671c3d60 85ae1d1e 7fe23136 7fa57657 2daad12f ae697080 e72eb01e d74dac81 c83202db f8c2b55e d74010ba ec23031d d769f0a6 ... (13 more lines)		 03:03
samcv AlexDaniel: MoarVM has been released :) 04:15
Geth rakudo: d75f4ae5de | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
Actual date, claim next release
 05:53
AlexDaniel gah… 07:25
AlexDaniel is waiting for networking on his server to go up again 07:26
[Tux] Rakudo version 2018.01-212-gd75f4ae5d - MoarVM version 2018.01-97-g22d2db5e0
csv-ip5xs0.833 - 0.978
csv-ip5xs-207.586 - 8.220
csv-parser11.953 - 11.978
csv-test-xs-200.453 - 0.481
test9.277 - 9.458
test-t2.597 - 2.679
test-t --race1.121 - 1.218
test-t-2047.085 - 50.346
test-t-20 --race16.443 - 19.444
 07:49
lizmat Files=1235, Tests=76221, 316 wallclock secs (14.85 usr 5.28 sys + 2168.25 cusr 215.07 csys = 2403.45 CPU) 08:36
AlexDaniel
.oO( I hope your data center is burning, I'm not taking any other excuse )
yaaaaaaaaaaaaaaay 08:40
Geth rakudo: c7d2f76567 | (Elizabeth Mattijsen)++ | docs/ChangeLog
It's $*PERL, not $*Perl
 08:44
rakudo: 21f6fb64c9 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/announce/2018.02.md
Fix announcement also
 08:46
AlexDaniel heh, edit wars over that line :) 08:47
ok the release will happen in ≈50 minutes when releasable finishes everything 08:48
lizmat AlexDaniel++ 08:52
AlexDaniel quotable6: ZOFFLOP 09:19
quotable6 AlexDaniel, OK, working on it! This may take up to three minutes (4506639 messages to process)
AlexDaniel, gist.github.com/0d693ba0a6323f42bb...ad7df55e92 09:20
AlexDaniel hang on S17-supply/syntax.t :S 09:21
timotimo quotable6: source
quotable6 timotimo, github.com/perl6/whateverable
Geth nqp: e860aa1a91 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
[release] Bump MoarVM revision to 2018.02
 09:53
nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2......2018.02
febb04d59b | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
rakudo: b69bc8b9a9 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[release] Bump NQP revision to 2018.02
 09:54
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018.......2018.02
1d11fd26d0 | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
AlexDaniel \o/ 10:01
samcv \o/ 11:06
timotimo cool 11:07
jnthn :D 11:08
Back to breaking things! ;)
timotimo maybe i'll merge the profiler multithreading enhancements right away, though tbh i haven't checked if it impacts single-threaded profiles negatively in any way 11:10
jnthn Well, at least check it still gives useful results :) 11:11
timotimo i'll do that before any merge 11:23
Zoffix AlexDaniel++ samcv++ # releases 11:25
Geth rakudo: zoffixznet++ created pull request #1547:
Post release		 11:29
rakudo/master: 5 commits pushed by (Zoffix Znet)++
roast: zoffixznet++ created pull request #389:
Post release
roast: d43b90d70c | (Zoffix Znet)++ | S32-str/ords.t
Cover Uni.list optimization

Covers github.com/rakudo/rakudo/commit/8b7385d810		 11:30
roast: e2b662a4aa | (Zoffix Znet)++ | S32-str/ords.t
Fix up counter in test description
roast: f364110614 | (Zoffix Znet)++ (committed using GitHub Web editor) | S32-str/ords.t
Merge pull request #389 from perl6/post-release

Post release
Zoffix ZOFVM: Files=1289, Tests=153231, 149 wallclock secs (20.05 usr 3.23 sys + 3192.34 cusr 157.45 csys = 3373.07 CPU) 11:53
dogbert17 samcv++, AlexDaniel++ release 12:17
lizmat++ weekly
ZOFFLOP: t/spec/S32-str/ords.t 12:18
Zoffix dogbert17: what are the failures in ords? And are you sure it's a flopper rather than a solid failure? 12:32
I think I wrote the test badly: it assumes .couny-only opt is definitely implemented (and you're seeing failures 'cause you're not on latest rakudo) 12:33
|4d fix to not assume .count-only impl github.com/perl6/roast/commit/d43b90d70c 12:34
ZofBot Zoffix, Will remind you on 2018-02-24T07:34:59.166033-05:00 about fix to not assume .count-only impl github.com/perl6/roast/commit/d43b90d70c
dogbert17 Zoffix. solid failure 12:35
1..17 12:36
No such method 'count-only' for invocant of type '<anon|254307192>'
in block at t/spec/S32-str/ords.t line 39
in block <unit> at t/spec/S32-str/ords.t line 35
Zoffix ZofBot: called it
ZofBot Zoffix, We walked on and on, but it was all so new to me that I forgot how tired I was
dogbert17 Zoffix: hold on it seems I'm on rakudo 2018.02 which is not the latest 12:37
Zoffix right
dogbert17 Zoffix: and the error magically disappeared, sorry for the noise 12:41
Zoffix :) 12:42
dogbert17: well it wasn't noise. You did find a bug in the test 12:43
dogbert17 is building 2018.02 on hir RPi 2 atm, a slow process 12:46
Geth rakudo: 16b57af549 | (Elizabeth Mattijsen)++ | 4 files
Make all explicit only subs a multi

Inspired by stackoverflow.com/questions/488190...ific-multi
  See also github.com/perl6/doc/issues/1783		 12:59
rakudo: 55bc053c56 | (Elizabeth Mattijsen)++ | src/core/Any-iterable-methods.pm
Make only sub min/min/minmax multi's
 13:34
nqp: c7f7762bb0 | (Jonathan Worthington)++ | 2 files
Map new nqp::tryfindmethod op for MoarVM backend
 13:40
nqp: c0e7e75b7d | (Jonathan Worthington)++ | 11 files
Update MoarVM bootstrap

So that we will be able to use nqp::tryfindmethod in Cursor.
nqp: ce831d0d53 | (Jonathan Worthington)++ | src/vm/jvm/QAST/Compiler.nqp
Add nqp::tryfindmethod on JVM backend

Which maps to an existing implementation.		 13:51
nqp: 1f75964b86 | (Jonathan Worthington)++ | 10 files
Update JVM backend bootstrap

So we can use the nqp::tryfindmethod op in NQP code.
nqp: 96c24117ec | (Jonathan Worthington)++ | src/QRegex/Cursor.nqp
Use nqp::tryfindmethod in Cursor!reduce

Seems to give a few percent speedup on JSON::Tiny, by avoiding the need for double lookups.		 13:52
rakudo: 182b7ea573 | (Elizabeth Mattijsen)++ | src/core/Any.pm
Make sub classify / categorize multi's
 14:15
rakudo: 63775474db | (Elizabeth Mattijsen)++ | src/core/array_operators.pm
Make sub push/append/unshift/prepend/splice multi's
 14:49
timotimo i had an interesting observation over in #whateverable: 14:54
these two commands
time perl6 -npe 'use nqp; INIT $*ARGFILES.nl-in = qq{\0\0}; next unless $_ ~~ /"hello there"/; last if $++ > 10_000' /tmp/\#perl6.cache
time perl6 -e '$*ARGFILES.nl-in = qq{\0\0}; for lines() { next unless $_ ~~ /"hello there"/; last if $++ > 10_000; .say }' /tmp/\#perl6.cache
on an about 300 megs big file 14:55
the latter takes about 36 seconds, the former takes about 46 seconds 14:57
dogbert17 gah, running the spectest suite on an RPi2 is slooow 14:58
timotimo m: say 48898900 / 73340315 15:00
camelia 0.666739705
timotimo that's the difference in frames entered from the overview page 15:02
and indeed the -npe one does 8147306 calls to CatHandle's nl-in proxy fetch program 15:03
the other doesn't seem to 15:04
Zoffix Maybe left-over from when we didn't have a CatHandle?
ilmari first read nl-in as a locale name and was wondering who spoke dutch in india 15:05
timotimo the --target=ast just looks like it calls map on &lines, though
dogbert17 isn't Proxy very underoptimized ?
RT #126520
synopsebot RT#126520 [new]: rt.perl.org/Ticket/Display.html?id=126520 [PERF] Proxy objects - frequent FETCHs
timotimo in total, the proxy parts take about 2.4 seconds out of the whole run 15:09
(62 seconds because of profiler overhead)
so i don't think it's that in particular 15:10
dogbert17 so there's something else lurking in the profiler output then 15:11
timotimo the call graph looks kind of drastically different and i don't really understand why 15:17
my mind is rather foggy right now for some reason
dogbert17 ENOCOFFEE ? 15:19
timotimo since i don't drink coffee ... yeah? 15:28
Geth rakudo: c2d0d3aced | (Elizabeth Mattijsen)++ | src/core/Buf.pm
Make sub subbuf-rw a multi
 15:40
[Coke] timotimo: no? coffee? how? 15:53
tea, maybe?
[Coke] has had a few coffees this morning, and is on his first coke zero
Zoffix
.oO( cannibalism... )
 15:57
timotimo autocannibalism even? 16:16
Zoffix m: print <a b c>.all.perl; print <a b c>.all; say <a b c>.all 16:55
camelia all("a", "b", "c")abca
b
c
Zoffix Wonder if Junction.gist should be more like .perl (Mu.gist is just Mu.perl) and only for print/put to actually junct over it 16:56
m: say ().all # especially due to this case not even printing a newline; You'd think a `.gist` would tell you a gist of a thing rather than nothing at all about it 16:57
camelia ( no output )
Zoffix m: print <a b c>.all.gist 16:58
camelia all(a, b, c)
Zoffix huh... 16:59
I see; .say threads through signature on self and and &say threads explicitly by junction.THREAD 17:01
Zoffix gonna change it later then
New blog post: "Perl 6: On Specs, Versioning, Changes, and… Breakage": rakudo.party/post/Perl6-On-Specs-V...d-Breakage 17:12
timotimo a propos blog: i'm looking for inspiration for how to get my TPF grant reports onto the 'net 17:21
i.e. what blogging software and maybe hosting platform
tyil timotimo: I know samcv posted her reports on a jekyll-based blog 17:30
dogbert17 RPi 2 spectests looks reasonably good, two files SEGV's, t/spec/S02-types/native.t and t/spec/S12-attributes/native.t 18:14
according to gdb the tests fail here: github.com/MoarVM/MoarVM/blob/mast...6num.c#L62 18:36
and it crashes there when running this line: github.com/perl6/roast/blob/master...ive.t#L135 18:43
gfldex Zoffix++ # for cheesy abstractions 19:00
samcv timotimo: yeah i use jekyll based blog. and setup with pygments as the highlighter so I get perl 6 syntax highlighting. tutorial I wrote is here cry.nu/tutorials/syntax-highlighting-in-jekyll/ 19:30
MasterDuke m: my $N = 2; my @x = FatRat.new(1, 1), -> $x { $x - ($x ** 2 - $N) / (2 * $x) } ... *; my $a = @x[12]; say $a.num; say now - INIT now 19:56
camelia No such method 'num' for invocant of type 'FatRat'. Did you mean any of these?
Num
sum

in block <unit> at <tmp> line 1
MasterDuke m: my $N = 2; my @x = FatRat.new(1, 1), -> $x { $x - ($x ** 2 - $N) / (2 * $x) } ... *; my $a = @x[12]; say $a.Num; say now - INIT now
camelia 1.4142135623731
0.01020226
MasterDuke m: my $N = 2; my @x = FatRat.new(1, 1), -> $x { $x - ($x ** 2 - $N) / (2 * $x) } ... *; my $a = @x[12]; say $a.chars; say now - INIT now
camelia 1571
2.44118459
MasterDuke generating that sequence is pretty fast, but printing large FatRats is very slow 19:57
github.com/rakudo/rakudo/blob/mast...#L102-L107 is the slow part 19:58
and it's the `$fract *= 10;` and `$fract -= $_;` in the loop that cause the slowness 19:59
is there another algorithm that doesn't require that repeated multiplication and subtraction? 20:00
Int.gcd is really where all the time is spent
called by github.com/rakudo/rakudo/blob/mast...at.pm#L42, which is called by github.com/rakudo/rakudo/blob/mast...at.pm#L144 20:01
fwiw i re-wrote that loop in all nqp and it wasn't any faster 20:02
Geth nqp: 218702caec | pmurias++ | 2 files
[js] Implement nqp::tryfindmethod
 20:07
nqp: d83b1c947c | pmurias++ | t/nqp/090-findmethod.t
Test nqp::tryfindmethod
nqp/master: 4 commits pushed by pmurias++ 20:10
MasterDuke Zoffix++ great post 20:53

Powered by Cro and the Raku® Programming Language.

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

Thank you!