IRC channel logs

<adanska>is there a way to extend what the initrd does beyond just copy-pasting base-initrd and adding on your own guile code? I feel like, in the spirit of guix, this should be possible.

<graywolf>Hi :) When using (rnrs records syntactic), is it a bug that (define-record-type foo (parent does-not-exist)) produces no error?

<graywolf>Right, so thanks to #scheme it seems our implementation is just buggy, will bug report

<old>graywolf: check the issue tracker first. I think this is already documented

<graywolf>Hm, there are other issues with the records, but I do not see this particular one

<graywolf>Are modules in guile loaded in serialized manner, or in parallel?

<jcowan>graywolf: That's deliberate, so that you can do things like create a syntactic record type whose parent is known at runtime. R6RS records are very much not what you expect.

<jcowan>Another example is that field names don't have to be unique.

<graywolf>jcowan: maybe I am missing something, but the specification seems to say "The parent argument must be either #f or an rtd.". I do not understand how (define-record-type foo (parent arinairostniarosntiarenst)) fits the bill of "either #f or an rtd"...

<graywolf>How is this behavior useful?

<ieure>Is there a simple way to profile my code? I found statprof, but all it prints is "No samples recorded."

<ieure>Do I have to give --statprof CLI arguments for that to work? All it does is print "No samples recorded." even with the tunables turned way up to #:hz 1000000 #:loop 50

<ieure>Losing my mind. Using (current-time) before/after the expression I want to profile returns the exact same value for the timestamp. But it takes 25 seconds from the time I hit enter until I get a prompt back.

<ieure>Maybe Guile is doing something like GC'ing after evaluating the expression, but before I get the prompt back?

IRC channel logs

2026-03-21.log