[FFmpeg-devel] Unicode filenames support on Windows regression

Kirill Gavrilov gavr.mail at gmail.com
Wed Apr 18 07:42:48 CEST 2012


On Tue, Apr 17, 2012 at 22:26, Nicolas George <nicolas.george at normalesup.org
> wrote:

> I tried to prepare the patch that follows semantic of definitions' hell
>> adopted in FFmpeg.
>>
> Can you elaborate on that?
>
In "my world" macro definitions should be used carefully and rarely with
preference to standard C/C++ constructions when possible.
And application these macros should never conflict within existing system
functions and so on.
FFmpeg code on the contrary intensively used macros wherever it is possible
and also override printf/fprintf/open (I'm sure something other too).
But this patch is not a place for this discussion and being poor
contributor my opinion is unlikely interested main developers.

Why fstat?
>
this is from playing with headers order. Original 'os_support.h' doesn't
includes <sys/types.h> and <sys/stat.h>
(supposing that them will be included in place of usage) however without
them structure ff_win32_stat can not be declared.
I'll remove unnecessary changes at evening and prepare corrected patch.

Duplicating a system data structure definition does not look like a good
> idea at all.
>
> Of course, I see the nature of your problem: if #define stat overrides the
> stat function, it will override the structure name too.
>
> I do not have a really good idea on how to deal with that problem.
>
That is... But I have no idea how to declare structure synonym without
typedef too
(typedef will not work here due to exploitation of same name for structure
and function).
-----------------------------------------------
Kirill Gavrilov,
Software designer.


More information about the ffmpeg-devel mailing list