On Wed, 27 Apr 2016 at 14:04 Anton Khirnov <an...@khirnov.net> wrote:
> Quoting Ronald S. Bultje (2016-04-27 14:37:20) > > Hi, > > > > On Wed, Apr 27, 2016 at 7:37 AM, Alexandra Hájková < > > alexandra.khirn...@gmail.com> wrote: > > > > > --- > > > libavcodec/bitstream.h | 475 > > > +++++++++++++++++++++++++++++++++++++++++++++++++ > > > 1 file changed, 475 insertions(+) > > > create mode 100644 libavcodec/bitstream.h > > > > > > diff --git a/libavcodec/bitstream.h b/libavcodec/bitstream.h > > > new file mode 100644 > > > index 0000000..8793556 > > > --- /dev/null > > > +++ b/libavcodec/bitstream.h > > > > > > > So, have you considered to just change the implementation in get_bits.h > > instead of this new API? I mean, the API looks largely identical - other > > than the prefix. It might even be 100% identical, but I can't be bothered > > to check. Large parts of the implementation look like they're copied > also. > > > > IMO the current names are highly confusing and generally horrible, so if > this is done at all, we might as well replace them. > I disagree, the old names are relatively clear. Whilst I think the speed improvements in this patch are great, the function names like bitstream_read_32 are really confusing. IMO adding a number suffix should be the exception rather than the norm (i.e when reading large numbers of bits). Kieran _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel