Re: git: 63ff982b17ee - main - simd.7: add scalar strrchr() for RISC-V to manpage

On 1 Nov 2025, at 18:00, Konstantin Belousov <kostikbel@gmail.com> wrote:
> 
> On Sat, Nov 01, 2025 at 11:48:19AM +0100, Robert Clausecker wrote:
>> Hi kib,
>> 
>> Am Sat, Nov 01, 2025 at 05:39:49AM +0200 schrieb Konstantin Belousov:
>>> On Fri, Oct 31, 2025 at 12:48:28PM +0000, Robert Clausecker wrote:
>>>> #include <machine/asm.h>
>>>> 
>>>> +        .weak   rindex
>>>> +        .set    rindex, strrchr
>>>> +
>>> Was this alias addition intended?
>> 
>> Yes, this is intentional.  See also lib/lib/{aarch64,amd64}/string/strrchr.S.
> 
> Ok, let me reformulate my question.
> Was the 'rindex' symbol exported from libc on riscv before your commit?
> 
> I remember that there was some effort to not provide 'obsoleted' exports
> from libc on the new arches like aarch64 and riscv.

It was exported prior to the series; see the last line in
lib/libc/string/strrchr.c and the unconditional entry in the
corresponding Symbol.map.

The mistake was making it a part of this commit rather than the one
that introduced a RISC-V strrchr.S, as the history is now a bit messed
up and confusing.

Jessica