diff options
| author | Manuel Frischknecht <manuel.frischknecht@gmail.com> | 2025-05-25 17:45:28 +0200 |
|---|---|---|
| committer | Manuel Frischknecht <manuel.frischknecht@gmail.com> | 2025-05-25 17:45:28 +0200 |
| commit | 58e77c03a3a0201ad6f0e2d35ce5beae12dc08f6 (patch) | |
| tree | e2118b112e26341aa1c9d1e6ffd6fb40eb807b8d /pkgs/development/python-modules/python-mapnik/python-mapnik_std_optional.patch | |
| parent | f1c69b60ad81d9942207698bbf776273de6c3b4d (diff) | |
bfr: fix broken build because of stricter GCC default checks
GCC now complains if `main` has no explicitly declared return type.
Because `bfr`s `configure` script uses a very minimal C application
to check whether GCC works _at all_ (literally: `main(){return(0);}`),
this caused configure failures.
Another issue with the `configure` script is that it logs the compiler
errors for its checks to `configure.log` (and not to stdout or stderr),
which means that the root cause wasn't logged during the failed nix
build.
I added a patch that rewires the compiler error file descriptor so
that it will simply log to stderr and another patch that fixes the
missing return type in the C example.
Unfortunately, it looks like the server isn't serving the source
files anymore (and finding an official upstream repo which I could
send a PR to seems... non-trivial).
Diffstat (limited to 'pkgs/development/python-modules/python-mapnik/python-mapnik_std_optional.patch')
0 files changed, 0 insertions, 0 deletions