On Thu, Oct 15, 2015 at 2:14 PM, Luca Barbato <[email protected]> wrote:
> On 15/10/15 13:38, Vittorio Giovara wrote:
>> On Thu, Oct 15, 2015 at 2:26 AM, Luca Barbato <[email protected]> wrote:
>>> `av_packet_unref` replaces it.
>>
>> Might be a tad bit more verbose here, deprecating is never fun for
>> users so each step has to be very well explained.
>
> I'll think something.
>
>>> ---
>>> diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
>>> index 5f1755a..3d5af09 100644
>>> --- a/libavcodec/avcodec.h
>>> +++ b/libavcodec/avcodec.h
>>> @@ -1152,13 +1152,13 @@ typedef struct AVPacketSideData {
>>> *
>>> * The semantics of data ownership depends on the buf or destruct
>>> (deprecated)
>>> * fields. If either is set, the packet data is dynamically allocated and
>>> is
>>> - * valid indefinitely until av_free_packet() is called (which in turn calls
>>> + * valid indefinitely until av_packet_unref() is called (which in turn
>>> calls
>>> * av_buffer_unref()/the destruct callback to free the data). If neither
>>> is set,
>>> * the packet data is typically backed by some static buffer somewhere and
>>> is
>>> * only valid for a limited time (e.g. until the next read call when
>>> demuxing).
>>> *
>>> * The side data is always allocated with av_malloc() and is freed in
>>> - * av_free_packet().
>>> + * av_packet_unref().
>>
>> not sure we want to do this: at least this does not happen in the
>> av_frame_unref version and probably would make more sense to free it
>> in free_packet
>
> This part needs rework indeed. What do you suggest exactly?
I think it would be simpler if APIs between frames and packets behaved
similarly, so packet_unref should not free side data.
>> y u no FF_API?
>
> Because I want to do that once at the end of the set, If you prefer I
> can do little by little.
yeah it'd be better to add FF_API every time we deprecate something
also make sure the code and fate still work with the FF_API disabled
>>> +
>>> +void av_packet_free_side_data(AVPacket *pkt)
>>> +{
>>> + int i;
>>> + for (i = 0; i < pkt->side_data_elems; i++)
>>> + av_free(pkt->side_data[i].data);
>>> + av_freep(&pkt->side_data);
>>> + pkt->side_data_elems = 0;
>>> +}
>>> +
>>
>> this is worth an APIchange entry, and while at it, mention the unref
>> function use instead of the deprecated one
>
> The function got moved aside but no API got changed yet.
right i missed that, however i wouldn't move the function at all, just
add the necessary macros and FF_API around the functions that got
deprecated
--
Vittorio
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
