Test for the bug with windows.compile_resources

[bruchar1]

Add test for #15206

[bonzini]

This should fix the issue, if either the creator or someone else wants to turn this into a non-draft PR.

[bonzini]

I did put together something that should work, but I don't have time to finish it. I think it's best to revert 632115e for 1.10 and try again in 1.11

[bruchar1]

I'm confused. The initial problem seems to be that current directory is not added to include paths when compiling resources. I tried to add it, and now the "15 resource scripts with duplicate filenames" test is failing with a gcc internal compiler error. But if I don't add the current dir to include path for "windres" resource compiler, my test fail because it doesn't find an included file in the .rc file.

The gcc bug has nothing to do with the preprocessor command. It is caused by the added include dir. I think that the added preprocessor command just showed a bug that was already present.

[bonzini]

Removing from milestone (#15206 is still there) because I don't think this is the right fix. The problem stems from adding project and global arguments for C to the preprocessor invocation, the include path is a red herring.

[bruchar1]

I don't agree. The first commit shows a bug that is present even if 632115e is reverted. The second commit fixes that.

The remaining question is whether it fixes #15206 or not. @lb90 should confirm it.

[bonzini]

This might be considered a bug for sure, but it's not the root cause of #15206. And given the windres issue with spaces in file names, it's dangerous to add to the include directories unconditionally.