Geth rakudo: ugexe++ created pull request #6592:
Derive typed hash storage keys the way Hash does
02:27
rakudo: ugexe++ created pull request #6593:
RakuAST: resolve &?ROUTINE in the scope the term stands in
03:43
06:32 finanalyst joined 08:48 finanalyst left
Geth rakudo/main: 0da4c7c454 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: resolve &?ROUTINE in the scope the term stands in

A routine's implicit &?ROUTINE declaration comes into force at that routine's own scope, and a method's comes into force before its body is parsed. A sub nested in a method therefore resolved its &?ROUTINE outward to the method's declaration, leaving the sub's own declaration with no use registered against it. The lowering analysis dropped it, and ... (7 more lines)
10:00
Blin/main: d1bf28ff42 | (Will Coleda)++ | resources/essential.json
This module is unmaintained...

  ... and so cannot be on the essentials list.
16:28
[Coke] Crap. I updated raku on my blin box from 2026.07 to 08. Have to reinstall deps. Whateverable is now giving me trouble again with the "Merging GLOBAL Symbols failed" 16:32
Tried installing from ecosystem and git HEAD.
I thought we fixed this years ago (and it worked fine when I setup this new blin box) 16:33
ah. uninstalled IRC::Client and reinstalled. went from 4.0.9 4.0.15, whateverable installed ok 16:37
I was in the middle of writing the "what depends on X" code for zef-deps - does zef itself already have that somewhere? (So I can see what in the ecosystem depends on a module called 'SSL', e.g.?) 16:38
(was) (like 2 years ago)
ugexe zef rdepends SSL 16:42
[Coke] ok. and I should not expect that to be fast, right? (ISTR speed was my biggest concern when doing it with zef-deps) 16:43
that generated no output here. 16:45
ugexe yeah its slow 16:50
[Coke] tried it with/without --verbose, no output for 'zef rdepends SSL'
.. which I guess is good news. :)
ah. I think I was going to do all deps, not just direct deps (which is kind of insane) on zef-deps 16:51
ugexe rdepends is also a bit weird in that it isn't reversing the dependency resolution, but rather taking the input and seeing if a module's listed dependency matches it. so distribution Foo:ver<1> depends on Bar, but zef rdepends Bar:ver<42> and zef rdpenends Bar:ver<69> would both return Foo:ver<1> (since both match the unconstrainted Bar dependency on Foo) 16:58
18:05 [Coke]_ joined 18:07 tbrowder__ joined 18:08 sjn_ joined, sjn left, [Coke] left, kurahaupo left, gfldex left, gfldex joined 18:09 tbrowder left, tbrowder__ is now known as tbrowder, kurahaupo joined 19:45 camelia left, camelia joined 20:03 [Coke]_ is now known as [Coke] 21:48 sjn_ is now known as sjn
Geth rakudo/main: d1a0e2fc72 | (Nick Logan)++ (committed using GitHub Web editor) | 4 files
RakuAST: don't run LAST in a loop that never runs its body (#6471)

Previously a pre-test loop (while/until, or a C-style loop with a condition) fired its LAST phaser even when the condition was false at once and the body never ran, though FIRST already did not. Gate LAST on the body having run at least once.
... (8 more lines)
22:27
rakudo: ugexe++ created pull request #6596:
RakuAST: type check the source of a bind
22:31
rakudo: ugexe++ created pull request #6597:
RakuAST: classify the node before offering it to the optimize ladder
22:51