Notices by anna (navi@social.vlhl.dev), page 5
-
Embed this notice
anna (navi@social.vlhl.dev)'s status on Sunday, 20-Oct-2024 03:56:04 JST anna @lanodan it's either some super small square screen smaller than my phone in a device that is too big for it
or a weird super-ultra wide aspect ratio screen that is too short to be usable
ugh... -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Sunday, 20-Oct-2024 03:11:44 JST anna why so many cyberdeck projects have such awful screens -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Saturday, 19-Oct-2024 05:35:47 JST anna gotta love how the c++ impl of json-ld just... ships a whole json impl as a single .hpp file
which is 24k lines long -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Friday, 18-Oct-2024 09:07:50 JST anna @chozu https://open-std.org/JTC1/SC22/WG14/www/docs/n3322.pdf
probably in c2y aka the next standard -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Friday, 18-Oct-2024 08:52:21 JST anna i found another test the official implementation does not pass... -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Friday, 18-Oct-2024 08:35:34 JST anna @julia meanwhile me writing c and worrying about every copy and/or allocation i make -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 09:18:11 JST anna *okay*
both tests i'm failing, pyld also fails -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 09:18:09 JST anna @rose i'm running the official test suite, so
honestly? both -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 09:18:05 JST anna @ada @rose i hate this so much
did they not run the whole test suite against their impls??? -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 09:18:02 JST anna @chozu @rose @ada tbh
how can i even trust those tests are right
since to make them all pass, i needed not only multiple spec changes, but also to implement something not on the spec and neither on the reference implementation
how do did they come up with a test the reference impl doesn't pass, did they make the test by hand and go "yep, this looks right"??? -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 09:14:14 JST anna Ok: 276 Expected Fail: 61 Fail: 0 Unexpected Pass: 48 Skipped: 0 Timeout: 0 this is it
i am more spec compliant than pyld (the unexpected pass are things i marked TODO)
-
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 08:33:34 JST anna fixing two bugs causes three others
and the fix isn't on the algorithm spec, but i see no way this would work without it -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 04:39:59 JST anna > For instance, if =mail-client/mutt-2.2.9 is masked (poor doggy!),
poor doggy -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 03:25:14 JST anna @julia take rust's safety, c's simplicity, meson-like ecosystem of distro first libraries but with automatic wraps/"vendoring" as an option, and ensure both dynamic and static linking work
and i would be happy -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Thursday, 17-Oct-2024 02:12:38 JST anna black friday, and it's decline, purely show how awful capitalistic consumerism is -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Wednesday, 16-Oct-2024 11:46:07 JST anna @lanodan i would like having depend be a declarative-ish function like in openrc, would be neat
tho we'd need some better syntax than what openrc does bc ebuilds dependency graphs get a lot more complex than openrc's -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Wednesday, 16-Oct-2024 11:44:18 JST anna @lanodan @cybertailor i find it funny ebuilds that just define an array... to right after do `foo_src_configure $ARRAY[@]`
so like, why the array, you're not really doing anything array-ish at all, just write arguments directly... -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Wednesday, 16-Oct-2024 11:43:44 JST anna @lanodan that shebang does basically two things
dependency resolution, so well, depend() works and starts stuff it depends on
then call the actual script through /lib/rc/sh/openrc-run.sh (soon to be /usr/libexec/rc) which defines the helper functions (like default `start` and `stop`, loading the supervisor=...) and such
and portage would be similar, since it already has /usr/lib/portage/python-*/*.sh scripts, so if we do that with bash already, similar would be posix sh defining helper functions there -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Wednesday, 16-Oct-2024 11:35:37 JST anna @cybertailor `set --` and `IFS` /j -
Embed this notice
anna (navi@social.vlhl.dev)'s status on Wednesday, 16-Oct-2024 11:32:15 JST anna @lanodan openrc init scripts are supposed to be posix though (we do use things that are on every posix shell but not really on the spec, like `local`, but still, supposed to be parseable by any /bin/sh)