IRC channel logs

2021-01-08.log

back to list of logs

***Sleep_Walker[m] is now known as Sleep_Walker
<remby>who has guile 3? I usually see guile22 in distros
<Aurora_v_kosmose>remby: I'm surprised that this is the case even in Arch. The only Guile 3.0 is in AUR
<remby>yeah
<rlb>remby: 2.2 and 3.0 are in debian, though maybe to your point, 3.0's not in debian stable yet, only testing.
<Aurora_v_kosmose>Well, Guix has up-to-date Guile, but that's to be expected.
<rlb>Hah, didn't know a release was coming -- still have some (minor) fixes accumulated here for 3.0 (and maybe 2.2) that I haven't pushed yet. I'll try to do that in soonish.
<rlb>wingo: noticed your fector.scm doesn't have a copyright, and didn't know if that might be unintentional (assuming it's suitable for consideration/consumption).
<rlb>...I was wondering how it might compare with 3.0 to the C version I have (didn't know about yours when I started) -- much rather have scheme if it's competitive or better.
<rlb>(of course)
<Aurora_v_kosmose>For my own projects I'd probably use Guile or CL anyway unless I really needed the edge in speed/resources
<rlb>Oh, mine is also guile, it's just a C implemented
<rlb>C module/extension.
<remby>Aurora_v_kosmose: what is CL ?
<rlb>common lisp, I'd assume
<Aurora_v_kosmose>Yep
<remby>ah
<rlb>For anyone interested https://github.com/lokke-org/lokke/blob/main/lib/lokke-vector.c -- of course some bits you'd want to remove from that if it were just for guile (e.g. the metadata perhaps). And while hopefully not broken, it's not all that pretty (probably needs further clean up/reworking)...
<rlb>(And if the scheme implementation is competitive, then it can just go away instead.)
<Aurora_v_kosmose>Guile 3.0 has better performance characteristics than Cpython and Ruby, I find.
<Anonymous__>Gratz for Guile 3.0.5
<wingo>rlb: regarding fector, i think lack of copyright was probably unintentional
<wingo>manumanumanu: lmk how it works out in practice :)
<chrislck>curious about the 3.0.5 nested-if speedup -- which guile commit is responsible for it?
<chrislck>It's Friday!!!!
<wingo>chrislck: 03998db647546df017feb19c0d5912e7451a61fc, and surrounding commits
<tohoyn>sneek: botsnack
<sneek>:)
<chrislck>wingo: thx! wow, >5 months since 3.0.4
<wingo>not much happened tho :)
<chrislck>nowadays, that's a good thing! O_o
<dsmith-work>Happy Friday, Guilers!!
<rlb>wingo ok, when you get a chance, if you'd like to put a copyright in there, I might see about testing it against the C version as a possible replacement later.
<wingo>lgplv3+ work for you?
<rlb>sure, and/or just match fash (presumably I thought that was fine earlier :) ).
<wingo>other things possible too
<rlb>lokke is dual lgpl+epl (so it's available to both communities -- was the thinking)
<rlb>fwiw, but pieces of it aren't, like fash.
<rlb> https://github.com/lokke-org/lokke#license fwiw
<wingo>done
<rlb>oops, mispelled fash in that list...
<rlb>wingo: thanks much -- assuming I get time, very curious to see how they compare.
<wingo>aaah i need to catch up with lokke :)
<rlb>well, I'm perfectly happy for you to work on making guile better too :)
<wingo>hehe i don't think i have time to work on lokke but i want to know what is happening there :-)
<leoprikler>tbh fash could really have used a better name
<rlb>Fair enough -- and I'll probably have some questions at some point. Already a few things I had to hack around in a less than ideal way that might suggest enhancements to guile...
<wingo>no doubt! i trust your judgment fwiw as regards committing anythign you like to guile
<rlb>I'm also not sure yet, but if nothing else, I've contemplated creating a project somewhere for fash and fector to provide a place separate from lokke to get any fixes, and I do have a few bug fixes for fash, and the addition of a fash-update procedure.
<wingo>happy to give feedback too, irc pokes can work well
<wingo>yes please do, if you want to take those files and run with them, be my guest :)
*wingo steps out
<dongcarl>Hi all, I'm trying to fix the guile-git-lib package on arch linux right now, and running into this issue with the tests. Wondering if anyone has a clue? Here's the test-suite.log: https://termbin.com/ojl3h
<chrislck>how would I convert a char* buffer into a list of bytes? the buffer may include '\0'
<chrislck>(from C)
<chrislck>char *buf = ' \0\1'
<chrislck>output = (list 32 0 1)