[FFmpeg-cvslog] r18351 - in trunk/libavcodec: 4xm.c 8bps.c 8svx.c aac.c aasc.c ac3dec.c adpcm.c adxdec.c alac.c apedec.c asv1.c atrac3.c avcodec.h avs.c bethsoftvideo.c bfi.c bmp.c c93.c cavsdec.c cinepak.c cljr.c...

Thilo Borgmann thilo.borgmann
Wed Jul 14 14:18:49 CEST 2010


Am 14.07.10 13:50, schrieb Diego Biurrun:
> On Tue, Jul 13, 2010 at 10:07:49AM +0200, Thilo Borgmann wrote:
>> Am 12.07.10 23:25, schrieb Diego Biurrun:
>>> On Sun, Jul 11, 2010 at 05:14:02PM +0200, Thilo Borgmann wrote:
>>>> Am 08.07.10 20:39, schrieb Thilo Borgmann:
>>>>> Am 08.07.10 20:23, schrieb Diego Biurrun:
>>>>>> On Tue, Apr 07, 2009 at 05:59:51PM +0200, rbultje wrote:
>>>>>>>
>>>>>>> Log:
>>>>>>> Implement avcodec_decode_video2(), _audio3() and _subtitle2() which takes an
>>>>>>> AVPacket argument rather than a const uint8_t *buf + int buf_size. This allows
>>>>>>> passing of packet-specific flags from demuxer to decoder, such as the keyframe
>>>>>>> flag, which appears necessary to playback corePNG P-frames.
>>>>>>>
>>>>>>> --- trunk/libavcodec/rl2.c	Tue Apr  7 17:37:26 2009	(r18350)
>>>>>>> +++ trunk/libavcodec/rl2.c	Tue Apr  7 17:59:50 2009	(r18351)
>>>>>>> @@ -181,8 +181,10 @@ static av_cold int rl2_decode_init(AVCod
>>>>>>>   */
>>>>>>>  static int rl2_decode_frame(AVCodecContext *avctx,
>>>>>>>                                void *data, int *data_size,
>>>>>>> -                              const uint8_t *buf, int buf_size)
>>>>>>> +                              AVPacket *avpkt)
>>>>>>>  {
>>>>>>> +    const uint8_t *buf = avpkt->data;
>>>>>>> +    int buf_size = avpkt->size;
>>>>>>>      Rl2Context *s = avctx->priv_data;
>>>>>>
>>>>>> Here (and probably in many other places) the function signature is
>>>>>> changed without updating the doxygen documentation.  Please update
>>>>>> the doxygen comments to match.
>>>>>
>>>>> IIRC this was one of the patches committed during my GSoC qualification
>>>>> task (which should also be part of the commit message btw.).
>>>>> I will have a look at all the files again and change the doxy's
>>>>> accordingly at the weekend.
>>>>
>>>> I cycled through all of them. Most of them don't include doxygen
>>>> comments, only three (escape124,fraps,rl2) feature doxygen comments
>>>> describing function parameters at all - none of them said anything about
>>>> "buf" and "buf_size" before the patch so I assume that was the reason I
>>>> did not add something new to the comments.
>>>>
>>>> The API description in avcodec.h looks fine.
>>>>
>>>>
>>>> Thus I don't see anything to do here?
>>>
>>> Either the documentation has to be completed or removed, but the
>>> current situation is buggy.  So patches welcome :)
>>
>> Why do I feel like wasting my time?
>> https://lists.mplayerhq.hu/pipermail/ffmpeg-cvslog/2010-July/031074.html
> 
> When the evil overlord demands it, you sometimes have to sacrifice a
> goat or worse.  What can I do?
> 
> Patches still welcome, no time is being wasted.

Thus I understand it wrong and by "documentation has to be completed",
you mean to add doxy's for all decoders. I assumed you meant to complete
the three existing ones, what you already did by the linked patch.

Then, here's a patch for removing these doxy's. Although, it negates
your work of the linked patch, of course. But if there is a consensus
about removing them, then, in the name of the evil overlord, let's
remove them. :)

-Thilo
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: undoc_decode_frame.patch
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-cvslog/attachments/20100714/c974e4a7/attachment.txt>



More information about the ffmpeg-cvslog mailing list