IRC channel logs

2021-04-10.log

back to list of logs

<civodul>janneke: interesting
<civodul>i don't think Guix has troubles with it
<civodul>but it means there's an allocation going on? how can it be?
<leoprikler>There are only two build systems worth using in 2021: Autotools and GWL :)
<rekado>even I don’t use GWL as a build system … yet ;)
<iv-so>gwl?
<morganw>This, I think: https://www.guixwl.org/manual/gwl.pdf
<iv-so>oh, so it's a build system
<morganw>probably more a "process orchestrator"
<leoprikler>if GWL is a process orchestrator, then make is not even a scheduler
***terpri_ is now known as terpri
<everstone>How do I split a string on newlines? (string-split "foo:bar baz" #\:) works for colon, but guile says #\\n is wrong
<daviid>everstone: try #\newline
<everstone>daviid: Thank you, that worked!
<janneke>sneek: if you think this https://paste.debian.net/1193016/ is interesting then let me ping wingo ;) should i mail to bug-guile?
<sneek>Okay.
<janneke>sneek: botsnack
<sneek>:)
<lampilelo>i thought it was supposed to do it the new way because afaik inner defines are implemented in terms of letrec inside a function block, i didn't realize it was different before
<lampilelo>the info manual explains it that way, so i guess it's intentional (?)
***rekado_ is now known as rekado
<rekado>iv-so: the GWL is a workflow language extension for Guix. The intended audience is scientists who want to run reproducible computer-aided experiments in HPC environments.
<iv-so>too abstract explanation
<rekado>iv-so: if you prefer a video: https://fosdem.org/2021/schedule/event/guix_workflow/
<rekado>there’s also a tutorial here: https://guixwl.org/tutorial
<iv-so>nah, I am fine with build system
<rekado>yeah, it’s not a build system
<iv-so>saw tutorial but understood nothing
<rekado>if you are not a scientist the GWL probably isn’t for you.
<rekado>not that it’s too complicated to use, just that you don’t have the kinds of problems that the GWL attempts to solve.
<iv-so>I played with this "reproducible research" thing a bit, but performance of emacs, org and giant postgres together killed all the fun and usefulness
<rekado>postgres?
<gagbo[m]>I wonder why pg as well
<tohoyn>sneek, botsnack
<sneek>:)
<iv-so>I was to investigate the reason why some rows in our DB were corrupted and fix them
<leoprikler>sounds not very reproducible to me
<janneke>lampilelo: ah yes, i forgot about that change
*janneke wasn't so thrilled of allowing duplicate identifiers there, but hey
<roptat>how can I prompt the user from a guile script?
<roptat>ha simply (read)
<mwette>roptat: or (use-modules (ice-9 readline)) (readline "$ ")
<mwette>[maybe requires (activate-readline)] -- see "Readline Support" section in the manual
<rlb>Anyone happen to know if we still need the lower(upper(x))? https://lists.nongnu.org/archive/html/guile-devel/2011-03/msg00111.html
<rlb>oh, wait, there's more in the thread.
<rlb>Right, looks like I can drop the extra call.
<rlb>Or maybe not - i.e. I think maybe that was saying that we could drop the extra call if we lowercased entire strings via libunistring and they were both the same encoding (here utf-8), in a call like compare_strings(). i.e. if we're still going char-by-char (for now), we might need the lower/upper.
***bonz060_ is now known as bonz060
***sneek_ is now known as sneek
<rlb>how do we pull in another gnulib function? i.e. I think I might want memcmp2
<wingo>o/
<stis>\o
***theruran_ is now known as theruran