| 17 Oct 2024 |
 Alyssa Ross | WHAT THE FUCK | 10:11:41 |
| Alyssa Ross | https://github.com/NixOS/nixpkgs/blob/master/pkgs/development/tools/misc/pkg-config/requires-private.patch | 10:11:56 |
 emily | …lol? | 10:12:08 |
 K900 | WHOMST | 10:12:14 |
| emily | 
Download image.png | 10:12:23 |
| K900 | https://github.com/NixOS/nixpkgs/commit/1e6622f4d5d500d6e701bd81dd4a22977d10637d | 10:12:30 |
| Alyssa Ross | Redacted or Malformed Event | 10:12:35 |
| emily | cursed prehistory discoverd | 10:12:34 |
| K900 |
Thus libxcb has to declare the pthread-stubs package as a
propagatedBuildInput, otherwise packages that check for libxcb using
pkg-config (such as libX11) will fail. But then we get a totally
unnecessary additional runtime dependency, which is clearly a bad
thing. | 10:12:40 |
| Alyssa Ross |
Whether this is correct is debatable
| 10:12:46 |
| K900 | Oh god is this from before dev outputs | 10:12:47 |
| emily | are we early enough to eat a pkg-config rebuild? | 10:12:47 |
| K900 | THIS IS FROM BEFORE DEV OUTPUTS | 10:12:51 |
| K900 | In reply to@emilazy:matrix.org
are we early enough to eat a pkg-config rebuild? Yes | 10:12:54 |
| emily | because it seems like a good idea to just rip this out for now | 10:12:59 |
| K900 | Any time is early enough to eat WHATEVER THE FUCK THIS IS | 10:13:05 |
| emily | I'm so embarrassed… we have to tell wlroots that the issue was because we patched pkg-config to be broken in 2008… | 10:13:22 |
| K900 | Arguably Alyssa's patch is still more correct | 10:13:42 |
| emily | I hope other distros have awful stuff like this lying around too | 10:13:44 |
| K900 | But also WHAT | 10:13:49 |
| emily | the worst part is that someone redid the patch in 2014 | 10:14:24 |
| emily | to update it to break a newer pkg-config | 10:14:44 |
| emily | https://github.com/NixOS/nixpkgs/pull/2798 wow, we didn't even have staging qua staging back then? | 10:15:06 |
| emily | I guess this PR was the birth of the modern staging cycle, looking at the comments | 10:15:56 |
| emily | well, I guess modern is whenever -next was introduced | 10:16:12 |
| emily | anyway I guess the patch is meant to work around https://gitlab.freedesktop.org/pkg-config/pkg-config/issues/28 | 10:17:01 |
| emily | I get the impression that this behaviour is now ossified and maybe CFLAGS.private is intended as a workaround in pkgconf | 10:17:48 |
| Alyssa Ross | In reply to @k900:0upti.me
Arguably Alyssa's patch is still more correct Disagree. It would produce unnecessary direct linkage. | 10:17:57 |
| Alyssa Ross | So do I just run some test builds with that patch removed | 10:18:16 |
| K900 | Yes | 10:18:21 |
