On Fri Jun 6 10:20:20 2025 +0000, Rémi Bernon wrote:
It's probably fine to skip the generic interfaces entirely for now I don't think you can normally define new ones and I suspect their definition comes from a special handling in MIDL. Still, for later when typerefs to parameterized types will be generated (for parameters, requires, etc), the names here seem to be wrong. I don't think it should be C++-style, and I think parameters should properly be referenced through typerefs too.
I forgot to include the name mangling changes when splitting patches. Does this look better?