IRC channel logs
2024-07-12.log
back to list of logs
<haugh>bjoli, very interesting project; i'm particularly interested in the inlining/DCE you're using to get a functional system up to speed with a mutative one. That said, goof's top level syntax is complex relative to egner's generators/qualifiers, please correct me but I can't see any ability to adapt to different numbers of return values, and I don't have a problem with mutation as long as there's <haugh>If I continue to butt up against 42's limitations, yours will be the first lib I dig into <haugh>I currently depend on an escape continuation for user-provided break; I'm interested in how you/Shinn achieved that. <ArneBab>(it’s completely based on call-with-current-continuation) <valorzard>Hey weird question: can guile work on platforms like iOS or game consoles? <ArneBab>sneek later tell valorzard: the LGPL is not a problem for proprietary platforms per-se, but there are challenges due to app-store policies. These are shared by Qt, too, though, and there are practical solutions: www.objc.io/issues/17-security/inside-code-signing/ <graywolf>It is not an end of the world, but would like to understand why it is necessary... <graywolf>I could just use (module-ref (current-module) 'open-pipe) which works, but I feel like I should understand the behavior here before I commit the workaround... <civodul>graywolf: use ‘use-modules’ only at the top level <civodul>if you want to manipulate a module at run time, there’s an (undocumented) API for that <civodul>in this case, you’d do: (let ((m (make-fresh-user-module))) (module-use! m (resolve-interface '(ice-9 popen))) (eval exp m)) <civodul>there are examples along these lines in Guix <graywolf>In "real" code I have module-use-interfaces! <civodul>oh, i don’t remember using that one :-) <graywolf>I wrote mine own re-export-all procedure, and writing tests turned to be harder then expected. <graywolf>Thanks a lot, was stuck on this for long time :) <graywolf>And indeed (eval '(begin e e* ...) (current-module)) works! Great ^_^ <sneek>I've been serving for 3 months <sneek>This system has been up 17 weeks, 3 days, 2 hours, 37 minutes <rlb>fwiw that libunistring bug never showed up in the bug tracker, but I did get a response. The "?" is expected behavior because the "escape" handler escapes unconvertable *characters* (I'd just assumed from elsewhere that it meant to escape undecodable *bytes*.). And in the example, it was the Latin-1 byte that couldn't be decoded as UTF-8. <rlb>My oversight -- the docs do say *characters*. <fnat>Ha, interesting, I see there's a 'alist->record' in Guix? It sounds promising, looking into that now. <fnat>Oh, ok, perhaps Guix's 'define-record-type*' then. <fnat>Hm, with the limitation that I can't do '(apply employee my-alist)' though... <fnat>Since I actually need to use 'define-record-type*', I'll go and ask in #guix.