On Dec 6, 2023, Jan Hubicka <hubi...@ucw.cz> wrote: > There is also access attribute which speaks directly about individual > arugments, perhaps you want to drop this one too?
Ah, I've looked a little into it, and now I have a vague recollection of why I don't mess with them: they only apply to arguments of pointer or reference type, and those are not (supposed to be) affected by the changes, not by indirection (which is what would make some "fn spec" notes impossible to convey), not by insertion of synthetic parameters. Those are placed at the end in part to avoid messing with parameter-index attributes. So attribute access can safely be left alone. Thanks for raising the issue. Maybe there should be at least a comment there, and perhaps some asserts to check that pointer and reference types don't make to indirect_parms. -- Alexandre Oliva, happy hacker https://FSFLA.org/blogs/lxo/ Free Software Activist GNU Toolchain Engineer More tolerance and less prejudice are key for inclusion and diversity Excluding neuro-others for not behaving ""normal"" is *not* inclusive
