[libc-commits] [libc] [libc++] Fix the behavior of throwing `operator new` under -fno-exceptions (PR #69498)

[Louis Dionne via libc-commits]

ldionne wrote:

> Are there tests that we actually correctly call the overridden throwing version from the overridden non-throwing version?

Yes, see for example `libcxx/test/std/language.support/support.dynamic/new.delete/new.delete.single/new.size_nothrow.replace.indirect.pass.cpp`.

> Also, I wonder how much we really care about the library being built without exceptions vs allowing the library to be used without exceptions.

We currently support both ways of building the library (`-fexceptions` and `-fno-exceptions`). I don't know how widely used the `-fno-exceptions` variant is used and we could discuss dropping it (and forcing people to always build the library itself with `-fexceptions`), but IMO this should be a separate change. In practice, I suspect that a lot of folks in the embedded world might be building the library with `-fno-exceptions`.

https://github.com/llvm/llvm-project/pull/69498