Post B9aBt89UoL4pa2MLnU by ariadne@social.treehouse.systems
 (DIR) More posts by ariadne@social.treehouse.systems
 (DIR) Post #B9a06EeUkxto3QLEg4 by ariadne@social.treehouse.systems
       0 likes, 0 repeats
       
       the #pkgconf maintainers would like to talk to humans, not bots which have gone on a wild ride.  this wastes our time as maintainers, and is why pkgconf does not allow LLM-generated contributions, especially agentic ones.https://github.com/pkgconf/pkgconf/pull/592(I'm tired, boss)
       
 (DIR) Post #B9a0zn0uhcCEYHJllg by ariadne@social.treehouse.systems
       0 likes, 0 repeats
       
       pkgconf has been written in such a way that anybody with basic C or C++ programming knowledge should be able to reasonably modify the software without LLM assistance.there really is no reason to throw an LLM at pkgconf.
       
 (DIR) Post #B9a3FUZRF11ltR58Wu by tknarr@mstdn.social
       0 likes, 1 repeats
       
       @ariadne Why did they even submit the PR in the first place? They even admit up front it doesn't fix a bug and that they didn't research whether or not pkgconf has any such ASCII-paths-only policy. The submitter seems to have developed LLM-induced brain damage. One more bit backing up your decision to ban LLM contributions.
       
 (DIR) Post #B9aBV3vOsPHVWGI1gW by imrehg@fosstodon.org
       0 likes, 0 repeats
       
       @ariadne that "bug" report is such a woosh-over-the-head anyways. "There's a non-ASCII character in test folder, added in a commit that says: 'Add a test where PKG_CONFIG_PATH contain non-ascii character' -- this must be a bug that I need to rectify!"Great choice of emoji, btw 😋
       
 (DIR) Post #B9aBt89UoL4pa2MLnU by ariadne@social.treehouse.systems
       0 likes, 0 repeats
       
       @imrehg at pkgconf we only use the finest emoji
       
 (DIR) Post #B9c2WvVkGMBBwdqzDM by lukeshu@social.coop
       0 likes, 0 repeats
       
       @ariadne fwiw, and this was a few years ago so I imagine the pkgconf codebase has changed quite a bit, I wanted to modify pkgconf so that it could tell me the value of a variable for each of a list of packages, the way that FDO pkg-config does, but in maybe 20 minutes of looking at the codebase I couldn't make enough sense of the program flow to figure out how.
       
 (DIR) Post #B9c7aX2dgIBKz1mILo by ariadne@social.treehouse.systems
       0 likes, 0 repeats
       
       @lukeshu pkgconf supports this already.
       
 (DIR) Post #B9c8DOCWU4TQrHvhi4 by lukeshu@social.coop
       0 likes, 0 repeats
       
       @ariadne `pkgconf --variable=girdir pkg1 pkg2 pkg3` prints a deduped-list of the girdirs set by those packages. FDO `pkg-config --variable=girdir pkg1 pkg2 pkg3` prints "{pkg1.girdir:-} {pkg2.girdir:-} {pkg3.girdir:-}\n"; I can tell what the variable is for each package separately. If I want to do what with pkgconf I need to do a separate invocation for each package.
       
 (DIR) Post #B9cD9JdG1oibIjJNA0 by ariadne@social.treehouse.systems
       0 likes, 0 repeats
       
       @lukeshu use --verbose
       
 (DIR) Post #B9cDUm5xOnwDesy4Xo by ariadne@social.treehouse.systems
       0 likes, 0 repeats
       
       @lukeshu actually I don't think --verbose will help here, I think --variable was restricted to a single item query.  we can change that though.
       
 (DIR) Post #B9do06SLXPQbIBdcnI by lukeshu@social.coop
       0 likes, 0 repeats
       
       @ariadne I just checked, and `--verbose` doesn't help here. But to be clear I'm not asking for help with pkgconf, I'm just sharing an anecdote of a time when a competent C programmer had trouble modifying pkgconf :)I imagine there were very few programs that relied on this behavior of pkg-config (it was undocumented, though the source code indicated to me that it was intentional), and that they have all been long updated to just make multiple calls to pkgconf.