IRC channel logs

2020-04-04.log

back to list of logs

<daviid>str1ngs: wrt this /dev/tty -> /dev/null np, feel free to substitute
<guix-vits>Hi there.
<guix-vits>daviid: test-g-value-boxed-semi-opaque failing in `guix build` environment. Can the "/dev/tty" be changed to "/dev/null", for example (at least as a 'quick way to disable this test')?
<daviid>guix-vits: yes, substitute to /dev/null is ok and better then removing the test
<guix-vits>daviid: thanks
***apteryx_ is now known as apteryx
<str1ngs>alextee[m]: looks good. GtkSourceView is pretty handy
<alextee[m]>yeah
<rekado>module/language/elisp/boot.el defines a bunch of macros in terms of things that don’t exist
<rekado>for example “defmacro catch” uses “call-with-catch”, which isn’t defined anywhere else.
<rekado>same with save-excursion and a bunch of others
<RhodiumToad>what version?
*civodul had an idea
<civodul>rekado: what about writing a blog post for Guile about Guile-Studio, with animated GIFs and all?
<civodul>we could even offer Guile-Studio packs for download somewhere
*RhodiumToad tries compiling guile3 on freebsd
<RhodiumToad>+ test '=' no
***dsmith-work is now known as dsmith
<spk121>what about adding animated gifs as aliases to r6rs keywords, to make code more fun
<RhodiumToad>m4/timegm.m4 tests $gl_cv_func_working_mktime without regard for the fact that it's not set.
<RhodiumToad>sed -i is nonstandard, and while bsd sed accepts it, the option value is required, so -i -e doesn't work (should be -i '' -e)
<rekado>RhodiumToad: I’m on the wip-elisp branch
<rekado>trying to fix tests first before trying to rebase onto some newer version.
<rekado>civodul: this sounds like a good idea, but I don’t think I have the bandwidth at this time :-/
<holomorph>yay wip-elisp
<rekado>I wonder why boot.el contains so many definitions written in Elisp that could probably be part of the language spec (written in Guile) instead.
*rekado reads more
<str1ngs>rekado: I don't think there is anyway to call scheme from elisp
<str1ngs>maybe language spec is different though.
*rekado is confused to see that there are changes to elisp in stable-2.2 that are not in wip-elisp
*rekado rebases
<chrislck>question about (ice-9 match) -- I'd believed (wrongly) that (match something ((_ x . _) x)) would simply capture something's 2nd element, and ignore the rest. but it seems the two _ are considered regular patterns.
<RhodiumToad>_ matches anything
<chrislck>i.e. I know; does the above (_ x . _) mean that the two _ must be equal?
<RhodiumToad>no
<chrislck>there's some weird stuff happening in my code then
<RhodiumToad>(match '(a b c d) ((_ x . _) x)) gives a result of b
<chrislck>I know you&I are right but the code behaves better when the pattern becomes (_ x . y)
<chrislck>where y is ignored
<RhodiumToad>I disbelieve
<chrislck>I'd like to disbelieve too
<RhodiumToad>so let's see a more complete example
<RhodiumToad>note that if _ is bound to something, things get weird
<chrislck>ahhh
<chrislck>that's it, _ was bound to gettext
<RhodiumToad>hm. wonder if that is a bug somewhere
<ATuin>hi, is there any example on how to use the pre-post-order in sxml?
<mwette> https://wingolog.org/pub/fold-and-xml-transformation.pdf
<mwette>ATuin: see above reference
<ATuin>ahhh that looks nice
<ATuin>thanks
<ATuin>real examples to read
<alextee[m]>i get a buunch of dupicated declarations when i include the x file: https://paste.debian.net/1138445/
***jonsger1 is now known as jonsger
<mwette>are you #include-ing on line 27 and then again on line 74 perchance?
<alextee[m]>i was including <libguile.h> on line 27
<alextee[m]>and then the snarfed .x file on line 74
<alextee[m]>are you supposed to wrap #include <libguile.h> inside an "#ifndef SCM_MAGIC_SNARFER" ?
<alextee[m]> http://git.savannah.nongnu.org/cgit/nomad.git/tree/guile/util.c this doesn't seem to do that
<alextee[m]> https://paste.debian.net/1138446/
<alextee[m]>here is what im doing
<alextee[m]>it compiles and links when i put the #include <libguile.h> inside the #ifndef above, but i get warnings that the functions are unused and in scheme i get unbound variable, so that's probably wrong
<alextee[m]>here are the first 1k or so lines from the .x file https://paste.debian.net/1138447/ i can't paste the whole thing on debianpaste
<alextee[m]>not sure if that's snarfed correctly
<mwette>I have not done any snarfing lately. I can't help w/ that, sorry.
<alextee[m]>mwette: thanks anyway :)
<rekado>just rebased wip-elisp on top of stable-2.2; now I’m getting two more test failures in vm.test
<rekado>load-thunk-from-memory expects a pair, but I guess wip-elisp broke that and produces a bytestring.
<rekado>I’m very grateful to cwebber for the best-effort changelog commit messages in wip-elisp; sadly the commits themselves are still kinda hard to understand because in some of them there are a lot of changes.
<rekado>will take me some time to get enough intuition to dare changing anything
<rekado>found the offending commit: d808027782a7ac978a26bf3bd904c58a133316e1 (intern arbitrary constants)