Tue Sep 13 09:57:42 CEST 2005
Rich Felker <dalias at aerifal.cx> writes:
> On Mon, Sep 12, 2005 at 11:03:06PM +0100, M?ns Rullg?rd wrote:
>> Reimar D?ffinger <Reimar.Doeffinger at stud.uni-karlsruhe.de> writes:
>> > Hi,
>> > On Mon, Sep 12, 2005 at 09:50:25PM +0100, M?ns Rullg?rd wrote:
>> >> Jan Knutar <jknutar at nic.fi> writes:
>> >> > On Monday 12 September 2005 23:38, Rich Felker wrote:
>> >> >> well this will require special comparison functions to ensure that
>> >> >> case mapping behaves according to ascii only -- otherwise we'll have
>> >> >> problems when people set it to turkish locale and such... :(
>> >> This isn't really a problem. We can simply decide never to name an
>> >> option in Turkish, and translating options would be very stupid.
>> > The point was that you can't use strcasecmp, because setlocale can
>> > break it.
>> It could, but AFAIK all commonly used encodings are ASCII compatible.
> No, they're not. tolower/toupper behave in very strange ways with
> regards to 'i' and 'I' when in Turkish locale (there's an uppercase
> dotted 'I' and a lowercase undotted 'i', and the characters 'i' and
> 'I' are not considered the same...)
Ugh. I guess it could be a problem after all.
>> Things like UTF-16 are rarely used.
> Huh? Things like UTF-16 are totally unsupported. They're not remotely
> compatible with the function in question.
Yes, they're totally incompatible, but since they are never used that
isn't a problem.
mru at inprovide.com
More information about the ffmpeg-devel