🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
Geth nqp/zhuomingliang/cleanup: 640856d117 | (Jimmy Zhuo)++ | src/vm/moar/QAST/QASTRegexCompilerMAST.nqp
cleanup some code in QASTRegexCompilerMAST
02:47
06:00 Xliff joined
Xliff \o 06:01
nine lizmat: there actually *is* a blog post :D niner.name/blog/rakuast_grant_report/index.html 06:23
lizmat weekly: niner.name/blog/rakuast_grant_report/index.html 06:58
notable6 lizmat, Noted! (weekly)
releasable6 Next release in ≈3 days and ≈7 hours. 2 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 11:00
Geth roast: 0d01cf9018 | (Will Coleda)++ (committed using GitHub Web editor) | S03-operators/shortcuts.t
Remove unused test file that doesn't compile. (#876)

Resolves #157
11:18
linkable6 ROAST#876 [closed]: github.com/Raku/roast/pull/876 Remove unused test file that doesn't compile.
ROAST#157 [closed]: github.com/Raku/roast/issues/157 Review S03-operators/shortcuts.t 11:19
Geth roast: 73eef396d7 | (Will Coleda)++ (committed using GitHub Web editor) | S32-io/IO-Socket-Async.t
Use dynamic port where appropriate (#875)
11:20
linkable6 ROAST#875 [closed]: github.com/Raku/roast/pull/875 Use dynamic port where appropriate
Geth roast/6.c-errata: c365a0c508 | (Will Coleda)++ (committed using GitHub Web editor) | 5 files
un-TODO several passing tests (#874)
11:21
[Coke] lizmat++ thanks 12:14
reminder - this week will be the third friday of the month, so news.perlfoundation.org/post/tprf-...roundtable will happen again. 12:19
notes from last month's: news.perlfoundation.org/post/perl-...march-2025 12:20
biggest news for me: Util is on the board now. 12:23
Geth nqp/zhuomingliang/cleanup: d8ba4b9009 | (Jimmy Zhuo)++ | 2 files
Reuse method name
12:28
ab5tract nine: just read your post 12:35
12:36 JimmyZhuo joined
ab5tract Congratulations on making it up that mountain 12:37
900 commits… incredible
lizmat indeed! 12:38
ab5tract I wish I could have helped more towards the end, but the comma revival had its claws in deep 12:39
JimmyZhuo comma is slow on windows, maybe my computer is too old. 12:47
I understand why the influx of community contributions to this project were a lot lower than I had hoped for. For me: compiling setting is so slow, when I tried to change something to debug the code, at least half time is waiting for compiling setting. that is, 8 hours work, 4+ hours is waiting :) 12:54
chaning one line code and wait for 120+ secs for confirmation 12:58
lizmat JimmyZhuo: wish I could get you a faster computer 13:04
JimmyZhuo not everyone has fast computer, it would be nice to improve rakudo comipling time 13:10
lizmat true... and in the short term it will only get worse with RakuAST in the bootstrap
but at least we should have a better idea on how to optimize that, as the code is definitely more accessible than "manually" messing around with QAST trees 13:11
JimmyZhuo I do like switching to new RakuAST grammar and drop the old one. 13:16
lizmat I guess we should do that after the 20245.04 release, right nine ? 13:17
JimmyZhuo not sure, the new one has no optimize stage, a lot of code will be slow. 13:18
Rakudo compiling setting time performance analysis pictures by Visual Studio, posted to github.com/rakudo/rakudo/issues/5851 13:26
Geth nqp: zhuomingliang++ created pull request #843:
Some code cleanup and re-order, reduces blib\Raku\Grammar.moarvm about 43KB
13:33
13:43 JimmyZhuo left
Geth ¦ rakudo: tbrowder self-assigned =begin para/=end para blocks do not preserve interior blank lines github.com/rakudo/rakudo/issues/5850 14:14
[Tux] Rakudo v2025.03-242-ge494d4d80 (v6.d) on MoarVM 2025.03-13-gb405e940e
csv-ip5xs0.268 - 0.280
csv-ip5xs-201.089 - 1.118
csv-parser1.093 - 1.097
csv-test-xs-200.115 - 0.116
test1.836 - 1.858
test-t0.412 - 0.427
test-t --race0.267 - 0.289
test-t-204.946 - 4.996
test-t-20 --race1.236 - 1.327
14:32
tux.nl/Talks/CSV6/speed4-20.html / tux.nl/Talks/CSV6/speed4.html tux.nl/Talks/CSV6/speed.log
nine lizmat: since not even make install runs without errors, I'd say we're still a bit away from switching defaults 14:37
ab5tract: so don't worry. There's still quite a few things left to do to chime in :)
lizmat nine: I assume the "make install" issue is another manifestation of the .assuming / BEGIN "foo".EVAL issue ? 14:39
nine No. BEGIN "foo".EVAL itself is fine after all. Without this not even module loading would work. 14:44
The problem with .assuming is just that it *doesn't* do the necessary things that BEGIN EVAL handling does
lizmat and what is the "make install" issue the, apart from it doesn't work? :-) 14:45
nine It's far more simple. It just runs into no outer $_ being available. The EVALed comp unit contains something like QAST::Var.new(:scope<local>, :name<$_>, :decl<static>, :value(QAST::Op(:op<getlexouter>, :name<$_>)) 14:46
There's also some RakuDoc issue with $=finish not being found in a CHECK phaser in RakuAST::L18N 14:47
And last but most worrying the whole module MoarVM::Profiler { class MoarVM::Profiler { }} WAT
m: module MoarVM::Profiler { my class MoarVM::Profiler { } }.HOW.^name.say 14:48
camelia Perl6::Metamodel::ModuleHOW
nine m: module MoarVM::Profiler { my class MoarVM::Profiler { } }; say MoarVM::Profiler.HOW.^name 14:49
camelia Perl6::Metamodel::ClassHOW
nine ^^^
[Coke] t/12-rakuast/xx-fixed-in-rakuast.rakutest (Wstat: 256 Tests: 105 Failed: 1) 16:09
Geth rakudo/main: 01a66c1288 | (Will Coleda)++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Add todo for legacy grammar

  (pass make test)
16:13
18:17 finanalyst joined
jdv we're thinkin more july for rakuast, right? 19:57
[Coke] I would guess at least that far. 19:58
need to fix make install, be able to run zef, make it the default, do a blin run without much fallout... 19:59
I'm not sure if there's a decision that that also triggers 6.d...
er, e 20:00
patrickb Does not trigger 6.e 20:24
We can, but no need.
tbrowder note rakudo issue 5851 was a mistake and was closed 20:26
jdv i feel triggered 20:27
jk
Geth ¦ rakudo: tbrowder self-unassigned =begin para/=end para blocks do not preserve interior blank lines github.com/rakudo/rakudo/issues/5850 20:31
21:59 MasterDuke joined
MasterDuke nine: i was thinking about that the other day (finding places where we re-do work) and it's tough. those are sort of algorithmic changes, which require both knowing what the current algorithm is and how to improve it 22:02
tellable6 hey MasterDuke, you have a message: gist.github.com/bbc99badfc2e2d940d...319a26db33
MasterDuke and finding them 22:05
japhb has an interesting idea. i was just wondering about counting all `method <...>` in the rakudo source and seeing how closely the number of calls to `<method-def>` match 22:06
well, that one seems to match up. `git grep -P '\bmethod [^ ]+\(' src/core.* | wc -l` give 7,691 and there are 7,896 calls in nine's profile 22:22
same with subs, grepping gives 2,295 and the profile has 2,470 calls to `<routine-def>` 22:25
`term:sym<nqp::op>` roughly matches also 22:46
which makes me think there probably isn't a problem of duplicating grammar? since nqp ops would show up in most constructs, except maybe signatures? 22:48
but i guess something like `method to-begin-time` from `class RakuAST::ParseTime` is a random example of something that could potentially be called more than necessary (3,235,503 times in this profile)? 23:00
though i could imagine 3.2m RakuAST nodes being created while parsing 23:01
timo we could log the byte offset and type of every node that to-begin-time is called on and side-by-side it with the source code to see what's up 23:14
MasterDuke `to-begin-time of: 2160 - 2199` 41 times 23:39
`to-begin-time of: 3226773 - 3226776` 15 times 23:40
those are the .from and the .to of the node's .origin
is that around the `#- StatementPrefix -------------------------------------------------------------` section of gen/moar/CORE.c.setting? 23:45
but afk for a bit
23:48 MasterDuke left