C ASSEMBLER_NAME patch

[Zack Weinberg]

Richard Henderson <rth@redhat.com> writes:

> On Thu, Jul 24, 2003 at 03:48:03PM -0700, Geoffrey Keating wrote:
>> It turns out that the 'alias' attribute wants the *assembler name* of 
>> the function, so the code generated was "correct".  IMHO, this is less 
>> than useful functionality in this context...
>
> Indeed.  I've been thinking that the alias attribute ought to take
> an identifier as well, and use the DECL_ASSEMBLER_NAME for that
> identifier.  Would make some uses in C++ nicer, since you wouldn't
> need to know all the mangling.

Heh.  I remember doing half a patch to that effect, oh, four or five
years ago now.

zw