IRC channel logs

2025-10-11.log

back to list of logs

<apteryx>has anyone used python-on-guile recently (past discussions include e.g. https://lists.gnu.org/r/guile-devel/2021-04/msg00026.html)? I was trying to fix the packaging in Guix (it fails to build), but it's nonobvious.
<apteryx>can (ice-9 match) be used to bind the multiple values returned by a call?
<characteristic>apteryx: (ice-9 match) does not implement ‘match-values’
<apteryx>that'd be handy
<apteryx>instead of this awkard combination of srfi-71 + match (list a b c)
<apteryx>ackward*
<characteristic>there is https://codeberg.org/dpk/extensible-match aka SRFI 262 does, there is a guix package for it
<characteristic>s/there is//
<apteryx>thanks
<euouae>Hello
<ekaitz>euouae: hi
<ekaitz>did you find the problem with the strings?
<euouae>ekaitz: not yet, I'm trying to learn the VM instruction set now
<euouae>I'm on it though, I'll figure out that bug
<ekaitz>:)
<ekaitz>that's the proper attitude
<euouae>It's a particularly annoying bug because if I can fix it I can make guile-ropes much more optimized
<ekaitz>:)
<dsmith>euouae, Without knowing much this stuff, I suspect your bug is in https://codeberg.org/guile/guile/src/commit/f54a7cf264b2ef69a453cd04151f11656661426c/module/language/tree-il/compile-cps.scm#L976-L999
<dsmith>s/much/much about/
<euouae>dsmith: Thank you! Will take a look there eventually.
<dsmith>euouae, Courtesy of ripgrep
<euouae>these are transformations from scheme to other intermediate representations?
<euouae>That's what it looks like, it's before bytecode compilation, right?
<euouae>scheme -> Tree-IL -> CPS language?
<mra>enouae: iirc CPS is the last IR before bytecode
<euouae>It's a good guess because it happens on -Ocps -Oresolve-primitives
<dsmith>euouae, Also has that u64-< and scm->u64 in it, which seems to appear in https://termbin.com/8ihs
<mwette>I think level 0 optimization skips cps.