[FFmpeg-devel] compile issues under mingw-cross-env

[Måns Rullgård]

Reimar D?ffinger <Reimar.Doeffinger at gmx.de> writes:

> On Sat, Jul 10, 2010 at 10:22:01PM +0100, M?ns Rullg?rd wrote:
>> Ramiro Polla <ramiro.polla at gmail.com> writes:
>> 
>> > On Sat, Jul 10, 2010 at 6:14 PM, Reimar D?ffinger
>> > <Reimar.Doeffinger at gmx.de> wrote:
>> >> However, I am quite convinced that all uses (except in ffsever) of
>> >> strcasecmp in FFmpeg are just plain _wrong_ and will break things
>> >> if a calling application uses setlocale.
>> >> For example, with TR_tr locale,
>> >> if (!strcasecmp(m->elems[j].key, "provider_name"))
>> >> will not match provIder_name.
>> >> I seriously doubt this can be considered correct behaviour,
>> >
>> > So are we going to have av_strcasecmp()? =)
>> 
>> http://git.mansr.com/?p=ffmpeg.mru;a=commitdiff;h=0dc41ad1923
>
> That is quite funny in this context, by using tolower it still
> has exactly the same issues.

Indeed.  I just happened to have the patch sitting around.

> av_strcasecmp_ascii
> with
> while (1) {
>     int delta = *s - *d;
>     if (!*s || (delta &&
>         (delta != 'a' - 'A' || *s < 'a' || *s > 'z') &&
>         (delta != 'A' - 'a' || *s < 'A' || *s > 'Z')))
>         return delta;
>     s++; d++;
> }
> should avoid it and would be faster.

This looks simpler:

#define ascii_tolower(c) ((c) | ((unsigned)((c)-'A') < 'Z'-'A')<<5)

All assuming ascii is what we care about of course.

-- 
M?ns Rullg?rd
mans at mansr.com