On 5/19/23 15:03, Xi Ruoyao wrote:
> On Fri, 2023-05-19 at 14:40 +0200, Alejandro Colomar wrote:
>> On 5/19/23 07:21, Xi Ruoyao via Libc-alpha wrote:
>>> On Fri, 2023-05-19 at 12:13 +0800, Xi Ruoyao wrote:
>>>> On Thu, 2023-05-18 at 20:29 +0200, Alex Colomar wrote:
>>
>> [...]
>>
>>>>
>>>> I guess when I took the note on the paper I misspelled "fflush" as
>>>> "setbuf" for some reason (maybe lack of coffee :( ).
>>>>
>>>> I'll retest all the changes in one hour or two hours.
>>>
>>> I've sent v2 with the mistakes in fflush, fflush_unlocked, and
>>> setbuf
>>> fixed.  Thanks for pointing out my mistake! >
>>> I guess I need to avoid creating a patch at 0100 AM in the future
>>> :).
>>
>> I'm glad to help :-)
>>
>> How about fclose(3)?  Was it left out on purpose, or was it a sleep(0)
>> issue too? :p
> 
> It was done at
> https://sourceware.org/git/?p=glibc.git;a=commit;h=71d9e0fe
> 
> Actually fclose is the start point of this __nonnull addition:
> 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109570
> 
> But unfortunately in Apr I didn't have enough time to check the entire
> stdio.h function by function.  Then in the following discussion for the
> GCC PR we found ferror should have __nonnull too and I decided to check
> all stdio.h functions this time.

Good to know!  Makes sense.  LGTM then.

Acked-by: Alejandro Colomar <alx@kernel.org>

Cheers,
Alex


-- 
<http://www.alejandro-colomar.es/>
GPG key fingerprint: A9348594CE31283A826FBDD8D57633D441E25BB5