Meson fixes and cleanups #256
Conversation
To avoid warnings about string functions like strdup which are otherwise undefined, but succeed at linking anyway when the C standard is c99.
This would be triggered without the previous addition of -D_BSD_SOURCE for strdup, among others
autoconf uses either C99 or Gnu99. Meson does not provide a graceful way to select gnu99 if possible or c99 (though there are several proposals currently happening to get there), so I've selected c99 as the conservative default. Without this, the compiler uses whatever it's default happens to be, which may or may not work out correctly, and hides bugs from CI that are present with c99 as the default.
|
Looks fine, I will merge it into my tree today. (If I click the button here, it will mess up when my personal git instance mirrors the repo back to Github) |
dcbaker
force-pushed
the
submit/meson-cleanups
branch
from
f6f4874
to
36c2055
Compare
|
I updated this PR to add -Werror to the woodpecker CI as well, but otherwise this latest push is exactly the same as the last (since you mentioned it on the other PR) |
This will help catch any new warnings added in CI.
It's more terse, and we don't need the support of a fallback value.
Instead of writing `['HAVE_FOO_H', 'foo.h']`, use meson's string methods to just write `['foo.h']`, and let meson create `HAVE_FOO_H` for us.
The latter doesn't work correctly when being used as a subproject, as it returns the *absolute* root. So if pkgconf is being built as part of muon, then it will return muon's source root. current_source_dir, on the other hand returns the directory correctly whether being built as a subproject or superproject.
Instead of attempting to figure out what the paths will be, take the returned file object and pass that around, meson will then automatically figure out the correct paths.
Because we really should be checking that it succeeds, and because not setting it is deprecated.
dcbaker
force-pushed
the
submit/meson-cleanups
branch
from
36c2055
to
dd779ad
Compare
|
|
||
|
|
||
| configure_file(input : 'Kyuafile.in', output : 'Kyuafile', configuration : cdata) | ||
| kyuafile = configure_file(input : 'Kyuafile.in', output : 'Kyuafile', configuration : cdata) |
There was a problem hiding this comment.
May want to match the : style with the rest of the files. Seem to be attached to the key elsewhere.
There was a problem hiding this comment.
The style is all over the place here, and I'm inclined to leave it alone since all I did was assigned the output.
|
Migrated to gitea |
This Gitea migration is kinda confusing to be honest. The README says to report bugs on GitHub, while the mailing list in on sourcehut, where there's also another pkgconf repo. I guess that the bug reporting link in the README simply has to be updated to point to Treehouse's Gitea instance (what's Treehouse btw? It's hard to determine by searching the web), but is the mailing list still monitored? What should it be used for? |
|
https://gitea.treehouse.systems/ariadne/pkgconf is the url. We really should update the docs. |
Yep, already forked that. My main question is about the sourcehut mailing list really (and also who/what Treehouse is) |
Initially I started on this because I'm using this in a subproject that set's it's C version to c99, and libpkgconf is full of warnings because of undefined functions. I realized that meson isn't setting a c_std unlike autotools. So i fixed that, added some warnings, updated the CI to be more picky, and then cleaned up the meson a bit (I am a meson dev, I might as well do something useful with all that knowledge :)