CI:

• Linux
• Linux-aarch64
• macOS
• Windows

Looks like I missed one in time_win32.c; that should now be fixed. Otherwise, initial CI looks pretty good. I'm going to run a full CI just to make sure that everything still works for all downstreams as well. I expect that to still have one warning on Windows, which needs a separate patch for RViz.

CI:

• Linux
• Linux-aarch64
• macOS
• Windows

I'm surprised I couldn't find any issue saying this can happen without using /experimental:preprocessor, but maybe Windows introduced a bug in MSVC yesterday (?).

So we don't have /experimental:preprocessor on, as far as I can tell. We do turn on /Zc:preprocessor (the successor to /experimental:preprocessor) in one place: https://github.com/ros2/rosidl/blob/e22083c970f6235ef4266bab80947f9dbc316c37/rosidl_generator_cpp/CMakeLists.txt#L108 . But as far as I can tell, that should only be for that particular test, so it wouldn't be a global thing.

I'm going to go ahead and merge this as-is to reduce our warnings for now. If Windows/MSVC gets fixed, or we figure out that we are accidentally turning on flags, we can always revert this.

(oh, and all of the failures on Windows are there on the nightlies as well)

/Zc:preprocessor (the successor to /experimental:preprocessor) in one place: https://github.com/ros2/rosidl/blob/e22083c970f6235ef4266bab80947f9dbc316c37/rosidl_generator_cpp/CMakeLists.txt#L108 .

Yeah, I saw that one.

But as far as I can tell, that should only be for that particular test, so it wouldn't be a global thing.

Agreed, I don't think that will affect other places.