@lanodan @xerz (what also does not make sense is replacing it in a selectively picked amount of places with another impl of make where it seemingly works, because subtle behavior differences will actually introduce subtle breakages to the result that you won't notice until much later, such is make...)
Notices by q66 (q66@blahaj.social), page 2
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 28-Sep-2024 04:24:30 JST q66 -
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 28-Sep-2024 04:24:15 JST q66 @lanodan @xerz i mean it makes total sense to ditch gnu make where you realistically can (because gnumake is a pile of very poorly documented extensions that is impossible to replicate), but this should involve ditching it for something that is not make, and is not a distro thing to decide
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 28-Sep-2024 03:46:39 JST q66 @lanodan @xerz there is nothing to be gained from ditching gnu make, it's pretty much just a mildly interesting experiment for purists
it's a build tool, it should not be treated specially from other build tools
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 28-Sep-2024 03:42:57 JST q66 @lanodan @xerz none of those have that kind of scope though, these are either domain-specific or toys
as far as truly general-purpose ones go that do not target gnu userland by default, there is only chimera and to a degree alpine
(the closest you go otherwise is gentoo and apparently nix, when unofficial setups are considered)
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 28-Sep-2024 02:32:41 JST q66 @n0toose but that's, like, their whole thing
-
Embed this notice
q66 (q66@blahaj.social)'s status on Friday, 27-Sep-2024 08:45:54 JST q66 wow this supposed linux rce is even more of a fucking nothing than i thought it would be
-
Embed this notice
q66 (q66@blahaj.social)'s status on Friday, 27-Sep-2024 07:55:15 JST q66 34689 lines added, 83 lines deleted
https://github.com/chimera-linux/tangle/commit/2d735c648a4f353ba250906a70d3df6b83693f94
as it turns out my original plan to use my own port of sd-event with basu will not work out due to sd-bus itself integrating the event stuff and... basu just rips it out :)
so here is a fresh sd-bus port from systemd 256 (still need to reintegrate tests, and add busctl which will be "fun" due to more utility lib stuff, and ...)
-
Embed this notice
q66 (q66@blahaj.social)'s status on Friday, 27-Sep-2024 03:32:59 JST q66 @lanodan @xerz nobody drives 70s/80s american cars because they are perpetually busted
-
Embed this notice
q66 (q66@blahaj.social)'s status on Wednesday, 25-Sep-2024 02:17:54 JST q66 oh, you want to get the underlying pid of a pidfd? open `/proc/self/fdinfo/<fd>` (handle all possible error conditions along the way), read it out line by line, find a line that begins with `Pid:`, strip its whitespace (handle all possible error conditions along the way), and parse the number contained in the leftover string (handle all possible error conditions along the way)
-- statements dreamed up by the utterly Deranged
-
Embed this notice
q66 (q66@blahaj.social)'s status on Wednesday, 25-Sep-2024 02:00:08 JST q66 the linux kernel uapi is a bigger usability/security footgun than c/c++ alone could ever be
it's pretty impressive how much stuff is only accessible by parsing it out of randomly-formatted text files in procfs and similar
people keep bashing systemd for being an impenetrable monolith but it's like 70% having to work around this trash fire
-
Embed this notice
q66 (q66@blahaj.social)'s status on Tuesday, 24-Sep-2024 20:25:18 JST q66 @xerz @lanodan i feel like mine got slightly nerfed back when i had my laser eye surgery, so it was nice to gain some of it back
-
Embed this notice
q66 (q66@blahaj.social)'s status on Sunday, 22-Sep-2024 02:35:38 JST q66 @cafkafk honestly i got kinda tired of all this stuff over the years
in the end every system is such a combo of random stuff that putting *anything* in the name makes no sense; the only thing that unifies it all is linux, so for practical reasons it makes sense to call it linux
but for whatever reason just about everyone chooses to beat the dead horse
-
Embed this notice
q66 (q66@blahaj.social)'s status on Sunday, 22-Sep-2024 02:34:55 JST q66 @leftpaddotpy @wakame @cafkafk beat you to it
i mostly packaged gcc a while back to have two working compilers, but it fails to bootstrap on one of our tier 1 targets, and when i tried to get help upstream (because it crashes in code that's impossible to decipher for any outsider) all i got is a very unhelpful response due to my system not being gnu enough and subsequently ignored forever :(
so i guess we're not gonna have gcc...
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 21-Sep-2024 23:28:49 JST q66 @lanodan whatever state any implementation of gettext is in, it's better than using a scuffed ancient utility nobody really uses and that is often not even available, both for anybody working with your tools and on your tools
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 21-Sep-2024 23:17:13 JST q66 @lanodan realistically for your purpose the one from gettext-tiny is complete (what gettext-tiny does *not* implement is stuff like translation of .desktop files, but for cli programs it's not really missing anything)
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 21-Sep-2024 23:03:51 JST q66 @lanodan the runtime gettext api is in libc, so the program does not link to anything
-
Embed this notice
q66 (q66@blahaj.social)'s status on Saturday, 21-Sep-2024 23:00:09 JST q66 @lanodan it's available on chimera (but mostly useless)
-
Embed this notice
q66 (q66@blahaj.social)'s status on Friday, 20-Sep-2024 20:05:54 JST q66 @lanodan @domi apparently a dragonflybsd extension (backported to free in 10.0)
-
Embed this notice
q66 (q66@blahaj.social)'s status on Friday, 20-Sep-2024 19:58:14 JST q66 @lanodan @domi hm interesting, i was about to suggest `cut -w` (which treats all whitespace as the delimiter) but it seems that's a bsd extension
-
Embed this notice
q66 (q66@blahaj.social)'s status on Thursday, 19-Sep-2024 22:40:26 JST q66 @lanodan it depends on a lot of configuration things (e.g. your file descriptor limit...)
it's important mainly because bwrap can help you pretend you're a certain uid/gid, but as soon as you need to e.g. chown to other uids, assign suid bits, xattrs, and capabilities, bwrap will not help you
fakeroot does because it wraps these calls and records it all internally (and since apk mkpkg is invoked within the same fakeroot env as e.g. these chowns, it can access that)