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.
00:04 fake_space_whale joined
MasterDuke down to one failing test in my default-to-int-not-num nqp branch 00:41
01:00 Ven`` left
MasterDuke although stage optimize of rakudo dies with `Unhandled exception: This type (Scalar) does not support associative operations at SETTING::src/core/Rakudo/Internals.pm6:735` 01:06
01:13 Ven`` joined 01:53 Ven`` left 01:54 Ven`` joined 02:02 Ven`` left 03:08 robertle left 03:10 robertle joined 05:15 fake_space_whale left 05:26 lizmat left, robertle_ left 05:36 fake_space_whale joined 05:52 fake_space_whale left 06:41 ufobat___ joined 07:32 |Tux| joined 07:45 brrt joined
brrt \o 07:45
.tell AlexDaniel I've pushed a commit to MoarVM that should resolve the windows problem 07:52
yoleaux brrt: I'll pass your message to AlexDaniel.
brrt btw, i found this by the advanced technique of reading git log and reverting suspicious commits until I found the guilty commit 07:58
this technique is available to more people than just me
|Tux| Rakudo version 2018.06-271-g84d45f07f - MoarVM version 2018.06-327-gfbc61e491
csv-ip5xs0.974 - 0.995
csv-ip5xs-207.602 - 7.687
csv-parser25.006 - 25.385
csv-test-xs-200.446 - 0.449
test9.913 - 10.131
test-t2.210 - 2.249
test-t --race0.955 - 0.965
test-t-2039.604 - 39.647
test-t-20 --race12.677 - 12.707
 08:00
08:55 lizmat joined
|Tux| Rakudo version 2018.06-271-g84d45f07f - MoarVM version 2018.06-327-gfbc61e491 08:57
(because I added java12 to tux.nl/Talks/CSV6/speed5.html
09:05 brrt left
lizmat Files=1246, Tests=76391, 383 wallclock secs (16.04 usr 5.38 sys + 2687.37 cusr 259.24 csys = 2968.03 CPU) 09:05
Geth nqp: 2a3a3f3c83 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 13 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...7-gfbc61e4 fbc61e4 Revert "Implement JIT templates for 10 ops" 647fcb9 Fix typo that prevented MVM_GC_DEBUGging bindlex 7467836 Fix compiler warning from last commit f06a725 Save 3 instructions per 64bits in siphash24() function ... (9 more lines)		 09:12
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...7-gfbc61e4
rakudo: 729edb4a49 | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 56 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....8-g2a3a3f3 2a3a3f3 [MoarVM Bump] Brings 13 commits 042ffa3 Correct some test descriptions 7019401 Sort QAST::VM annotations when dumping them (#491) 3fc2378 Fix the backends for nqp::<eqatic eqatim eqaticim> ... (67 more lines)
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....8-g2a3a3f3
roast: 683296bf00 | (Zoffix Znet)++ | 2 files
Rename JSON excp. handler test file to more generic one
 09:17
rakudo: 1f27a19b48 | (Zoffix Znet)++ | t/spectest.data
Follow spectest file rename
rakudo: c978050399 | (Zoffix Znet)++ | t/05-messages/03-errors.t
Skip heredoc warning message test on windows

is-run test routine doesn't quite work on Windows because of RT#132258
  rt.perl.org/Ticket/Display.html?id=132258		 09:28
synopsebot RT#132258 [new]: rt.perl.org/Ticket/Display.html?id=132258 [SECURITY][WINDOWS] `run "perl6" ...` can be made to execute shell commands
09:38 stmuk_ joined
Geth star: f28a390364 | (Steve Mynott)++ | 3 files
bump submodules again, mainly to pull in zef's Build.pm6 to fix p6doc doc deployment
 09:38
rakudo: ef61ce32e2 | (Zoffix Znet)++ | src/core/Exception.pm6
Only run THE_END if custom ex. handler gave us False

If we got True, the default exception handler we run afterwards will run THE_END for us.		 09:40
09:40 stmuk left
Geth star: 7485def206 | (Steve Mynott)++ | Configure.pl
unattempt windows fix
 09:46
¦ rakudo: lizmat self-assigned Code review wanted: &THE_END with custom exception handlers github.com/rakudo/rakudo/issues/2150 10:01
stmuk_ lizmat: there should be a rakudo-star-2018.06-RC0 release in time for the weekly 10:15
lizmat stmuk_++
Geth ¦ star: stmuk self-unassigned Win installer does not set up file associations github.com/rakudo/star/issues/115 11:26
stmuk_ Geth: sssh 11:51
Geth nqp: e5975648a6 | (Elizabeth Mattijsen)++ | tools/build/MOAR_REVISION
Bump MoarVM to get extra JIT optimizations
 11:53
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...9-g21ea40f
11:54 travis-ci joined
travis-ci Rakudo build passed. Zoffix Znet 'Only run THE_END if custom ex. handler gave us False 11:54
travis-ci.org/rakudo/rakudo/builds/409770249 github.com/rakudo/rakudo/compare/c...61ce32e209
11:54 travis-ci left
stmuk_ pl6anet.org/drop/ has source and windows binary for rakudo-star-2018.06-RC0 (Mac binary to follow when ++virtualsue does it) 12:03
Geth ¦ nqp: tbrowder self-assigned need simple examples of using QAST inside nqp programs github.com/perl6/nqp/issues/492 12:25
12:25 Ven`` joined 12:26 Ven` joined 12:30 Ven`` left 12:33 diakopter joined 13:00 Ven` left 13:39 skids joined 14:06 fake_space_whale joined
AlexDaniel . 14:35
yoleaux 07:52Z <brrt> AlexDaniel: I've pushed a commit to MoarVM that should resolve the windows problem
AlexDaniel cool, thanks 14:36
14:42 diakopter left
Geth rakudo: bb6a6f80bc | (Elizabeth Mattijsen)++ | tools/build/NQP_REVISION
Bump NQP to get latest JIT optimizations
 14:48
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....9-ge597564
45080856b2 | (Zoffix Znet)++ | t/05-messages/03-errors.t

Closes github.com/rakudo/rakudo/issues/1813 R#1813
AlexDaniel weekly: reportable: gist.github.com/ea7f276bbc09adff12...11ae71acca 14:51
notable6 AlexDaniel, Noted!
14:52 pmurias joined
lizmat AlexDaniel++ 14:52
AlexDaniel weekly: Rakudo 2018.07 cancelled
notable6 AlexDaniel, Noted!
timotimo this is the only time that happened in a very long time. i think we can still be somewhat proud 14:53
AlexDaniel weekly: Squashathon this weekend: twitter.com/zoffix/status/1022879562387087360
notable6 AlexDaniel, Noted!
15:00 diakopter joined 15:10 Zoffix joined
japhb AlexDaniel: And I think it may actually be good for our reputation that a large part of the cancelation was that we decided not to leave a major platform behind. 15:12
Zoffix "somewhat" proud? I don't see skipping the release as a negative at all. Quite the opposite: we're not taking risks by releasing something that we don't want to release just because we gotta stick to some made up schedule.
It's ready when it's ready™.
AlexDaniel japhb: that wasn't even an option, yes 15:13
japhb Zoffix: Generally I agree with you. But there is some reputational advantage to be had in displaying a super-regular high-quality release cadence. Occasionally showing delaying or skipping a release on quality grounds shows prudence. Doing it too often shows lack of baseline quality, or lack of release cycle maturity. 15:14
s/showing// 15:15
Personally I'm all for erring on the side of release quality, but making sure that we introspect if we delay or cancel too often. 15:16
15:16 ExtraCrispy joined
Zoffix Literally a few lines above it says this haven't happened in a very long time. 15:17
AlexDaniel well, “delays” are somewhat normal though. The release date doesn't have that meaning nowadays 15:18
.oO( release week? )
japhb Zoffix: I know that. I was stating my personal philosophy, not giving anyone a hard time. 15:19
15:19 Zoffix left
japhb AlexDaniel: Actually, I find I have mixed emotions about that. I like that we're erring on getting it right. I worry that constantly slipping off our published expectation indicates its time to think about why. 15:20
[Coke] If we really want to cut a release every month on a pre-determined date, we need to change how we're doing releases. 15:32
(mainly: don't commit directly to master) 15:33
I don't feel strongly enough about monthly super-on-time releases to push for that, though. 15:34
AlexDaniel well, I can start release branches a few days before the release and cut from them
[Coke] that only mitigates the issue.
could still be a surprise somewhere in master. 15:35
(assuming that your build process is only testing a subset of platforms, e.g.)
to be clear, I'm fine with missing the occasional release. Anyone who is going to complain about it and hold it up as an example of how we're screwing up was already not on the team. 15:36
15:36 benjikun left
[Coke] also, thanks again to everyone (AlexDaniel++ in particular) for automating the crap out of this process. 15:37
timotimo not to forget zoffix who kicked off that process :) 15:38
15:38 benjikun joined
[Coke] ofc. 15:39
japhb Yeah, I don't think we're screwing up. But I do think we can do even better. In some sense I see the variable release cycle not as a bad thing, but as "not yet as awesome as it could be". I'm not particular worried about addressing this in the short term, but it feels like something we could look at in the next year or two *usefully* (as in, makes things actually better for the Rakudo devs) 15:49
[Coke] +1 15:50
synopsebot R#1813 [open]: github.com/rakudo/rakudo/issues/1813 [testneeded][tests in rakudo not spec] X::OutOfRange stringyfies instead of gistifies
16:16 Ven`` joined
lizmat notable6: weekly 16:19
notable6 lizmat, 6 notes: gist.github.com/3c6bc593f37fc3c3d5...7d1561183d
16:22 robertle left 16:25 fake_space_whale left 16:38 Ven`` left 16:40 Ven`` joined 17:39 lizmat left 17:42 Ven`` left 17:44 lizmat joined 17:48 Ven`` joined 17:57 lucasb joined 18:04 Zoffix joined
Zoffix So we had that debate about whether `num` should default to 0e0... Turns out it already does in params: 18:04
m: sub foo (num $x?) { dd $x }()
camelia 0e0
18:04 Ven`` left
Zoffix and even attributes default to 0e0 :) 18:05
18:11 x[LGWs4x4i]uG2N0 left 18:14 x[LGWs4x4i]uG2N0 joined 18:16 dct joined
Geth rakudo: f045042d4f | (Zoffix Znet)++ | src/Perl6/World.nqp
[6.d] Default `num` to 0e0

  - Per 6.d-prep
  - Looks like everthing but `my` variables already defaulted to 0e0,
   so we could have saved some reels of discussion on this topic		 18:19
roast: 2abee09f88 | (Zoffix Znet)++ | MISC/misc-6.d.t
Spec `num` defaults to 0e0 in 6.d

POV impl: github.com/rakudo/rakudo/commit/f045042d4f		 18:20
lucasb yay Zoffix++ 18:21
Geth 6.d-prep: ce2ef9460f | (Zoffix Znet)++ | 3 files
Mark num defaults to 0e0 as completed

POV Impl: github.com/rakudo/rakudo/commit/f045042d4f Propspec: github.com/perl6/roast/commit/2abee09f88 Docs: github.com/perl6/doc/commit/33603b1f18		 18:23
18:24 Zoffix left 18:37 buggable left, buggable joined, ChanServ sets mode: +v buggable 18:38 buggable left, buggable joined, ChanServ sets mode: +v buggable 18:39 dct left
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/07/30/...t-garbage/ 19:08
[Coke] .ask curtispoe if there's a video for the future of 5&6. 19:17
yoleaux [Coke]: I'll pass your message to curtispoe.
19:29 travis-ci joined
travis-ci Rakudo build failed. Zoffix Znet '[6.d] Default `num` to 0e0 19:29
travis-ci.org/rakudo/rakudo/builds/409979526 github.com/rakudo/rakudo/compare/4...45042d4fa2
19:29 travis-ci left
buggable [travis build above] ☠ All failures are due to: failed make test (1 failure). Across all jobs, only t/02-rakudo/repl.t test file failed. 19:30
Geth rakudo: f923670210 | (Zoffix Znet)++ (committed using GitHub Web editor) | t/packages/Test/Helpers.pm6
Make `doesn't-hang` wait longer
 19:34
roast: db9b8bd189 | (Zoffix Znet)++ (committed using GitHub Web editor) | packages/Test/Util.pm
Make `doesn't-hang` wait longer
19:41 Zoffix joined
Zoffix So 6.d subs are totally additive, just as if you were loading a module? 19:42
$ ./perl6 -e 'use v6.d.PREVIEW; say 0e0/0e0'
Ambiguous call to 'infix:</>(Num, Num)'; these signatures all match
I thought it'd replace the 6.c stuff...
lizmat++ # weeekly 19:43
And I'm guessing if I define new versions of infix:</> in 6.d, any place that uses that op in 6.c setting would still get 6.c versions? 19:44
The 6.d version switching stuff sure looked a lot more complex in my head :)
yup, 6.c core still uses 6.c versions of the op. 19:48
and it's used all over the shop in trig routines. 19:49
Geth 6.d-prep: 12b28ede2f | (Zoffix Znet)++ | 2 files
Defer IEEE-fication of / to later versions

Tried implementing it, but hitting a problem on the same scope as R#1289 github.com/rakudo/rakudo/issues/1289
Basically the 6.d verions of the ops are currently loaded as if they were just comming from a module and aren't visible by the rest of the 6.c setting. There's a ton of trig and other math ops that aren't affected by the change, so I'm deferring it, until we figure out R#1289 and get a clearer picture on caller's lang ver.		 19:54
synopsebot R#1289 [open]: github.com/rakudo/rakudo/issues/1289 [6.e] Implement a Way to Know Caller's Language
Zoffix ZofBot: just push everything to 6.e and we can release 6.d tomorrow! >:} 19:58
19:59 Zoffix left
Geth rakudo: f43f118ab8 | (Elizabeth Mattijsen)++ | src/Perl6/Optimizer.nqp
Do the range optimization a bit differently

Instead of returning the value, return the QAST that is needed. This should allow for some more flexibility and other optimizations.		 20:09
20:19 ufobat___ left
Geth nqp: 16745aee8a | (Paweł Murias)++ | src/vm/js/Operations.nqp
[js] Impelement _nd ops
 20:56
nqp: 9ec419223f | (Paweł Murias)++ | t/nqp/067-container.t
Test _nd ops
20:59 Ven`` joined 21:15 skids left
lucasb I tried to golf R#2151 21:31
synopsebot R#2151 [open]: github.com/rakudo/rakudo/issues/2151 Grammar flickering between working and not working on `:sym` rules
lucasb arrived at this: gist.github.com/lucasbuchala/7d51f...fc74f76338
someone should golf even more 21:32
sometimes it matches <end1>, other times it fails to match <end2> 21:33
(if someone confirm the behavior, please tell :)
AlexDaniel e: gist.github.com/AlexDaniel/8d1a373...b57db9dff1 21:34
evalable6 AlexDaniel, Using file “main.p6” as a main file, other files are placed in “sandbox/”
(exit code 1) 04===SORRY!04=== Error while compiling /home/bisectable/git/wh…
AlexDaniel, Full output: gist.github.com/607b0949c88359f715...8b91a92b3c
AlexDaniel ah I know 21:36
e: gist.github.com/AlexDaniel/8d1a373...b57db9dff1
evalable6 AlexDaniel, Using file “runner.p6” as a main file, other files are placed in “sandbox/”
(exit code 1) 04===SORRY!04=== Error while compiling /home/bisectable/git/wh…
AlexDaniel, Full output: gist.github.com/0512ae33d7b58402a9...6b30d208ba
AlexDaniel e: gist.github.com/AlexDaniel/8d1a373...b57db9dff1
evalable6 AlexDaniel, Using file “runner.p6” as a main file, other files are placed in “sandbox/”
end1
AlexDaniel e: gist.github.com/AlexDaniel/8d1a373...b57db9dff1
evalable6 AlexDaniel, Using file “runner.p6” as a main file, other files are placed in “sandbox/”
end1
AlexDaniel e: gist.github.com/AlexDaniel/8d1a373...b57db9dff1
evalable6 AlexDaniel, Using file “runner.p6” as a main file, other files are placed in “sandbox/”
(exit code 1) 04===SORRY!04=== Error while compiling /home/bisectable/git/wh…
AlexDaniel, Full output: gist.github.com/9bef07b1c6141ce312...7d3bdbf34c
AlexDaniel lucasb: so that's the issue? 21:37
yeah
ok it feels like I can't reproduce it on anything earlier than 2018.05 21:39
let's add a loop there… :)
bisect: old=2018.03 gist.github.com/AlexDaniel/8d1a373...b57db9dff1 21:42
bisectable6 AlexDaniel, Using file “runner.p6” as a main file, other files are placed in “sandbox/”
AlexDaniel, Bisecting by exit code (old=2018.03 new=f43f118). Old exit code: 0
AlexDaniel, bisect log: gist.github.com/4cf4acdc17f97993ea...6cc0138e38 21:43
AlexDaniel, (2018-04-25) github.com/rakudo/rakudo/commit/6c...f0a07b9361
AlexDaniel b2eb9e1e7 Randomly seed a hash secret to mitigate ease of DOS attack
that sound related
c: 6c1983a97c^,6c1983a97c gist.github.com/AlexDaniel/8d1a373...b57db9dff1 21:44
committable6 AlexDaniel, Using file “runner.p6” as a main file, other files are placed in “sandbox/”
AlexDaniel, gist.github.com/36acfe13be2ce080db...9abe780b07
lucasb the commit "For NFAs with many codepoints, use binary search"?
ah, hash order you mean
AlexDaniel maybe, but the fluctuating nature suggests it's hash related 21:45
21:54 pmurias left, pmurias joined
Geth rakudo: 7043559bbb | (Elizabeth Mattijsen)++ | src/Perl6/Optimizer.nqp
Make the for ^$max { } optimization also work

This codegens the optimized loop also when either end of the range specification is a native int *variable*. This makes for loops like this about 1.35x as fast (with only 10% of the call frames):
   my int $max = 1_000_000;
   for ^$max { }		 21:59
lizmat timotimo: ^^^ 22:00
:-)
timotimo phew, i hope that's very, very robust :) 22:01
lizmat I think it is 22:05
timotimo: it basically is the same structure, but instead of initialiizing the variables from constants, it can also initialize them from native integers 22:06
I think I can even make it 2x as fast still if it is possible to tweak the signature of the block to -> int $_ 22:09
if it didn't have any signature specified yet
22:09 lucasb left
lizmat but that's for another day 22:09
good night, #perl6-dev!
22:17 pmurias left
timotimo lizmat: that's not safe if $_ gets assigned to within the block, isn't it raw by default? 22:23
m: for ^100 { $_ = "hi" }
camelia Cannot assign to an immutable value
in block <unit> at <tmp> line 1
timotimo oh, ok
that's good, then
22:30 Ven`` left, go|dfish joined 22:36 Ven`` joined 22:55 Ven`` left, Ven` joined 23:02 Ven` left
MasterDuke github.com/rakudo/rakudo/issues/1955 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!