IRC channel logs

2025-02-25.log

back to list of logs

<ArneBab>rlb: I hope it’s OK that I chimed in on your behalf in the mailing list. I think the snarky reply you got was uncalled for so I tried to explain that what you did is a service to getting Lilypond and Guile closer together and does not deserve defensiveness but rather a big thank you.
<ArneBab>rlb: since you did not get that there as much as would have been warranted: thank you very much! Getting Windows fixes merged is quite important for many use-cases of Guile (all that need to be usable everywhele) and your review does help a lot!
<rlb>ArneBab: no worries -- unless I get the time to (and then randomly decide I feel like) doing a more thorough review and some testing, I'll probably leave it alone. In any case, the main thing I'd want to try to be sure of of course, for anyone pursuing it, is that we don't risk breaking the 3.0 ABI.
<rlb>But the series may well be fine. If nothing else, ideally, it'd be good to have a solid second opinion about changing the hash time.
<rlb>s/time/type/
<rlb>(which also might be fine)
<rlb>wingo ^
<rlb>(wingo wrt changing the hash type, and *if* we're going to change it, is (u)intptr_t the one we want)
<cow_2001>no sha256 support on guile-git! if anyone wants to help: https://gitlab.com/guile-git/guile-git/-/issues/32
<cow_2001>someone said that the reason my package definition cannot used is that guile-git cannot use sha256 git repositories
<cow_2001>civodul: know how one (me? maybe?) could add that feature in?
<cow_2001>(guix package definition of https://codeberg.org/kakafarm/guile-bytevector-peg/)
<cow_2001>oh, right. RTFM. there is a big guile-git info manual!
<cow_2001>looking for "dynamic-func" i see that there is no mention of it in the current guile manual, but it does appear in guile 2 versions
<cow_2001>(@ (system foreign) dynamic-func)
<old>if you are writing new code, I strongly suggest to use the new API instead
<cow_2001>should be documented :|
<cow_2001>also, the guile 2 manual is spicy!
<old>agree
<cow_2001>"The more one hacks in Scheme, the more one realizes that there are actually two computational worlds: one which is warm and alive, that land of parentheses, and one cold and dead, the land of C and its ilk."
<cow_2001>maybe it is used for backward compatibility?
<cow_2001>the manual authors should add again the olde API, mark it deprecated, and link to the new API
<dariqq>Hello, does someone have a TAP driver for srfi-64 tests?
<civodul>dariqq: there’s the Automake driver used in Guix and other packages, not sure if that’s quite what you’re looking for
<dariqq>civodul: I know of the automake driver but I find autotools a bit unintuitive and overly complicated. Currently trying to make a new project with meson but having a hard time convincing it to accept the test suite
<dariqq>meson supports TAP format so I currently trying to make a custom runner that outputs that but hoped someelse had a similar issue before me
<ray1729>dariqq: this exists, it doesn't do exactly what you're asking but it you might be able to build on it: https://github.com/ft/guile-tap
<civodul>dariqq: ok, that makes sense to me; i just happen to know of TAP via Automake :-)
<dariqq>ray1729: Thanks, a bit more complex than I need but found some useful things in there
<rlb>dariqq: also https://codeberg.org/lokke/lokke/src/branch/main/mod/lokke/scm/test-anything.scm
<rlb>That's driven by https://codeberg.org/lokke/lokke/src/branch/main/mod/lokke/scm/test.scm
<rlb>fwiw
<rlb>(and see test/lokke-* for examples of use -- it's not beautiful, but works)
<dariqq>rlb: the test-runner looks similar to what i currently have. Now i just need to hook it up to meson and probably discover the 1000 edge cases
<rlb>:)
<dariqq>yaym the basic stuff works (even with nice color output of the different test status)
<dariqq>Really liking meson so far apart from not being able to build/install into subdirectories directly which means youd need to recurse into all subdirectories to compile the scheme files there
<fantazo>I played a little with spritely goblins. interesting to finally compile it.
<fantazo>but I'm not sure how to code with it yet. it feels similar to Erlang, but somehow the edges show.
<dthompson>fantazo: feel free to join #spritely if you want to talk goblins
<dthompson>what do you mean by "somehow the edges show"?