🦋 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.
lizmat nine: is parsing of regex / rule / token already 100% implemented in RakuAST that you know? 12:22
working on deparsing RakuAST::RegexDeclaration, and I don't seem to be able to find a way to find the difference between those three 12:23
as they all codegen to RegexDeclaration
also: something like:
token { foo } 12:24
is codegenned as three Regex::Literal nodes, instead of one
which feels... wrong ?
oddly, It looks like the QAST generation is combining consecutive Regex::Literals again 12:27
Geth DBIish/main: a8b4676c3e | (Rod Taylor)++ | .github/workflows/test.yml
Swap testing 2020.12 for 2022.12

Keep up to date and 2020.12 has started regularly failing for unknown reasons.		 12:42
lizmat also: on the subject of codegenning $s ~~ s/foo/bar/ 13:06
why don't we create a dedicated HLL method for that, Rakudo::Internals.APPLY-SUBSTITUTIONS 13:07
and codegen it as:
$/ = Rakudo::Internals.APPLY-SUBSTITUTIONS($s, /foo/, "bar", flags) ? 13:08
Geth rakudo/main: da963241a8 | (Elizabeth Mattijsen)++ | src/Raku/ast/regex.rakumod
Make RakuAST::Regex::Sequence join Regex::Literals

When feeding ...Sequence.new() a bunch of consecutive Regex::Literals, they will be combined into a single RegexLiteral. This corrects deparsing in many cases, but not all yet.
Brings number of passing tests to 621		 14:33
rakudo/main: 8ee619270f | (Elizabeth Mattijsen)++ | 2 files
Fix issue in deparsing RakuAST::Regex::Literal

We should just use .raku semantics if there are non-word chars in the string. Adjusts tests accordingly. Introduce a new internal method for this functionality, as we're probably going to be needing that again.
patrickb What's an easy OP to put in raku code and break on in Moar? 14:49
(say / MVM_string_say doesn't work. I didn't investigate.) 14:50
lizmat so you want it to generate proper bytecode yet bomb on Moar? 14:51
patrickb I want gdb to break once it started executing by code (so skip over compilation) 14:53
break as in "be able to put a breakpoint in gdb"
lizmat nqp::die() ? 14:54
patrickb But then my program aborts. :-P
breaking on "write_bytes" works with "say" 15:02
Ah. say obviously codegens to write_fhb 15:10
I guess the "say" op is unused then... 15:11
Is there any writeup on how the regex machinery works? How do Cursors, NFAs, backtracking, ... play together? 15:53
lizmat nothing in nqp/docs jumps out :-( 15:56
patrickb I'm off for today. Have a nice weekend everyone! o/ 15:59
Geth rakudo/main: 9998d1338c | (Elizabeth Mattijsen)++ | src/Raku/ast/regex.rakumod
Join Regex::Literal with RakuAST::Regex::WithSigspace

Fixes another deparse issue.		 16:08
nine patrickb: I usually break on nqp::sin_n(1e0) 18:37
patrickb: Works well....unless of course you're debugging very mathy code :)
lizmat: no idea about any of the regexy stuff 18:38
Geth nqp: usev6++ created pull request #794:
[JVM] Add some constants DISP_*		 21:02

Powered by Cro and the Raku® Programming Language.

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

Thank you!