Anton Khirnov <[email protected]> writes:

> On Mon, 02 Jan 2012 11:03:59 +0000, Måns Rullgård <[email protected]> wrote:
>> Anton Khirnov <[email protected]> writes:
>> 
>> > It's the same as av_vsrc_buffer_add_frame(), except it doesn't take pts
>> > or pixel_aspect parameters. Those are read from AVFrame.
>> >
>> > Deprecate av_vsrc_buffer_add_frame().
>> > ---
>> >  avconv.c                  |    5 ++---
>> >  doc/APIchanges            |    5 +++++
>> >  libavfilter/buffersrc.h   |    9 +++++++++
>> >  libavfilter/version.h     |   11 ++++++++++-
>> >  libavfilter/vsrc_buffer.c |   17 +++++++++++++++--
>> >  libavfilter/vsrc_buffer.h |    3 +++
>> >  6 files changed, 44 insertions(+), 6 deletions(-)
>> 
>> [...]
>> 
>> > +/**
>> > + * Add a frame to the filtergraph s.
>> > + *
>> > + * @param frame frame to be added.
>> > + * @warning frame data will be memcpy()ed, which may be a big performance
>> > + * issue. Use av_buffersrc_buffer() to avoid copying the data.
>> > + */
>> > +int av_buffersrc_add_frame(AVFilterContext *s, AVFrame *frame);
>> 
>> Why does this name not have an avfilter_ prefix?
>> 
>
> avfilter_buffersrc_* is
> a) a bit long

Can't argue with that, but it's not the only option.  For example,
dropping the buffersrc part (rendering avfilter_add_frame()) would make
sense.

The whole notion of exposing this as a filter to the caller is a bit
weird since it is the only way of feeding data into the system.

> b) the buffersrc part already specifies the namespace, no need to add
> avfilter.  We do the same in lavu and lavc (e.g. avfft.h).

Those are different.  Whereas the FFT functions have nothing to do with
codecs, this function is an integral part of libavfilter.

-- 
Måns Rullgård
[email protected]
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to