On 11/08/14 5:31 PM, Michael Niedermayer wrote:
> On Mon, Aug 11, 2014 at 04:35:39PM -0300, James Almer wrote:
>> It was wrongly being exported and used by libavfilter.
>>
>> Signed-off-by: James Almer <jamr...@gmail.com>
>> ---
>>  libavfilter/deshake_opencl.c | 4 ++--
>>  libavfilter/unsharp_opencl.c | 6 +++---
>>  libavutil/opencl_internal.c  | 2 +-
>>  libavutil/opencl_internal.h  | 2 +-
>>  4 files changed, 7 insertions(+), 7 deletions(-)
> 
> note, the header says:
>  * This interface is considered still experimental and its API and ABI may
>  * change without prior notice.
> 
> not sure changing it to avpriv before its stable API/ABI is a good
> idea
> 
> also its in opencl_internal.h
> 
> maybe we should just disallow this with --enable-shared

The public header says that, not the internal one. This function should have 
never 
been called ff_ if it was meant to be used by libavfilter since its inception.
And at this point, any changes to it in the future would mean an ABI break 
regardless 
of name anyway. But we can change its name to a proper one right now without 
breaking 
ABI since we just bumped major.
And as far as linux distros go, disabling OpenCL with shared builds would be 
the same 
as not offering OpenCL at all. Probably for the best since it's supposedly 
experimental 
anyway.

In any case, removing the ff_* part from libavutil.v is the entire point behind 
this.
Worst case we can make it only export this one function instead of the entire 
namespace.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Reply via email to