<roelj>When compiling something with Guile 3.0.7 I get: "Wrong type to apply: #<syntax-transformer foreign-library?>". While compiling with 3.0.2 seems fine. What has changed? <roelj>But adding --r6rs to the guild command doesn't seem to make a difference. <Zelphir>Has it become more or less conforming from 3.0.2 -> 3.0.7 then? <roelj>That is an interesting question. <roelj>I'm sure I'm doing something on the fringes of how I should use syntax transformers. So I'd like to avoid doing that in the first palce. <tohoyn>daviid: what is the minimum version of Guile 3.0 required for G-Golf to work? is it 3.0.7? *janneke pushes an updated wip-mingw branch <janneke>arrays.test, bytevectors.test, numbers.test now also pass \o/ <sneek>dthompson, you have 1 message! <sneek>dthompson, daviid says: same here, since switched to guile 3, and the latest geiser, not sure which causes the problem- if you find a solution, let meknow ... ***dsmith-w` is now known as dsmith-work
<janneke>dthompson: no, i'm using --disable-jit <janneke>wine throws a null pointer reference/usage immediately, so it seems <dthompson>I've given up on ever having guile work on windows *chrislck is sad to hear about guile-on-windows <leoprikler>Well, in actuality it is an error to say that some software runs on Windows. *dthompson hands out Grand Pedant award *janneke just tries to keep some users happy... *chrislck sad about windows-under-guile then <janneke>it's unbelievable that crap didn't already die <str1ngs>tohoyn: 3.0.7 I had asked him about that a couple of days ago. Also thanks for the deb packages, I do use them now and again. <apteryx>is [0-9A-f] not a valid regexp range on Guile 3? <apteryx>on guile 3.0.2 it isn't, but on guile 3.0.7 it seems to work fine. weird <leoprikler>to be on the safe side A-Fa-f is the correct one <dsmith-work>apteryx: Remember, Guile just uses whatever the C lib provides for regexp. <dsmith-work>As Guile is moving more and more from C to Scheme, it might make sense to replace that with irregex or something. <dsmith-work>scheme@(guile-user)> (map (lambda (s) (string-match "a.c" s)) '("abc" "aXc" "a\0c")) <ArneBab_>Can I do a conditional import that does not fail when the module is not available, but uses a fallback? <ArneBab_>concrete: If guile-websocket is not available, I just want to disable the option to use it. <form_feed>Yeah, was just reading that, thanks. Now I need a oneliner for serving static files locally. <ArneBab_>leoprikler: in scheme, so I can have a script that I can start with --testserver to use the websocket as stdout and stdin <ArneBab_>but still work without additional dependencies to install on desktiop <leoprikler>should code already be excluded on the expand side or is it enough if it's missing from eval? <leoprikler>if you just care about eval, doing a (false-if-exception (begin ...)) might just work <ArneBab_>the hard problem is that the import must not cause a failure <ArneBab_>if I find no other solution, I’ll have to create a separate script <leoprikler>hang on a sec, (import ) in the sense of load, yes? <leoprikler>in that case, as long as you don't put it in the top-level (define-module ) clause, you can absolutely use (false-if-exception) to wrap your import <leoprikler>problem being that you might end up with a lot of "possibly unbound" variables from that module <leoprikler>so ideally, you'd like to check that in a cond-expand-esque manner <civodul>ArneBab_: false-if-exception is in boot-9.scm <ArneBab_>(I don’t find it in the reference manual) <ArneBab_>warning: possibly unbound variable `fail-if-exception' <leoprikler>Alternatively, you can use autoloads, but then you'll get an exception when you actually load that thing