On Mon, Apr 18, 2011 at 12:53:50 (CEST), Martin Storsjö wrote:

> On Mon, 18 Apr 2011, Ronald S. Bultje wrote:
>
>> Hi,
>> 
>> On Mon, Apr 18, 2011 at 1:04 AM, Reinhard Tartler <[email protected]> 
>> wrote:
>> > On Fri, Apr 15, 2011 at 18:05:48 (CEST), Luca Barbato wrote:
>> >
>> >> diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
>> >> index 109d2a4..df4e617 100644
>> >> --- a/libavcodec/avcodec.h
>> >> +++ b/libavcodec/avcodec.h
>> >> @@ -1035,6 +1035,10 @@ typedef struct AVPanScan{
>> >>  #define FF_BUFFER_HINTS_PRESERVE 0x04 // User must not alter buffer 
>> >> content.
>> >>  #define FF_BUFFER_HINTS_REUSABLE 0x08 // Codec will reuse the buffer 
>> >> (update).
>> >>
>> >> +enum AVPacketSideDataType {
>> >> +    AV_PKT_DATA_PALETTE,
>> >> +};
>> >> +
>> >>  typedef struct AVPacket {
>> >>      /**
>> >>       * Presentation timestamp in AVStream->time_base units; the time at 
>> >> which
>> >> @@ -1057,6 +1061,17 @@ typedef struct AVPacket {
>> >>      int   stream_index;
>> >>      int   flags;
>> >>      /**
>> >> +     * Additional packet data that can be provided by the container.
>> >> +     * Packet can contain several types of side information.
>> >> +     */
>> >> +    struct {
>> >> +        uint8_t *data;
>> >> +        int      size;
>> >> +        enum AVPacketSideDataType type;
>> >> +    } *side_data;
>> >> +    int side_data_elems;
>> >> +
>> >> +    /**
>> >>       * Duration of this packet in AVStream->time_base units, 0 if 
>> >> unknown.
>> >>       * Equals next_pts - this_pts in presentation order.
>> >>       */
>> >
>> > To my reading AVPacket is a public struct and therefore cannot be
>> > extended without bumping major. Did we break ABI with this change?
>> 
>> We can add stuff, but only at the end. We probably broke ABI and this
>> should have been moved to the end. We can still do that now.
>
> Not in this struct. AVPackets are allocated statically (e.g. on the stack) 
> within calling user applications, and thus adding members even at the end 
> requires a major bump.

In this case, I seek for approval of the proposed patch to delay the
sidechannel introduction via #ifdeffery.

-- 
Gruesse/greetings,
Reinhard Tartler, KeyID 945348A4
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to