IRC channel logs

2025-08-29.log

back to list of logs

<daviid>this fails forme: git checkout https://codeberg.org/guile/guile.git main
<daviid>fatal: not a git repository (or any of the parent directories): .git
<rlb>Did you mean clone?
<daviid>ah, right - lost my mind ... :-)
<rlb>:)
<mwette>old: common-subexpression-elimination will merge duplicate expressions but multiple parameter references may produce different values, so they are not guarenteed to be duplicate, right?
<mwette>(I assume you mean parameters generated by make-parameter.)
<mwette>my bad: parameters are thread (dynamic context), so it seems maybe could be possible if the object could be identified as a parameter.
<mwette>If p is a parameter then `(p)' in `(p) (foo) (p)' may produce two different values.
<mgd>rlb: Thank you for the algorithm. I was looking at that earlier. Just wondering if there was a sort in a module I could use
<old>mwette: I guess it can but not in: (format (current-output-port) ...) (format (current-output-port) ...)
<old>and generally speaking, people tend to use `parameterize' and not just calling the parameters
<old>given that, load-fusing _could_ be done I guess
<old>but it could also break existing code. Option for O3?
<mwette>old: In your example the compiler would have to know that format is not mutating current-output-port, right?
<old>yes
<old>which ofc is not that easy to proof