Diego Biurrun <[email protected]> writes:

> On Tue, Oct 04, 2011 at 09:59:42AM +0100, Måns Rullgård wrote:
>> Diego Biurrun <[email protected]> writes:
>> > On Mon, Oct 03, 2011 at 02:26:46PM -0700, Ronald S. Bultje wrote:
>> >> On Mon, Oct 3, 2011 at 12:04 PM, Diego Biurrun <[email protected]> wrote:
>> >> > On Mon, Oct 03, 2011 at 11:37:36AM -0700, Ronald S. Bultje wrote:
>> >> >> From: "Ronald S. Bultje" <[email protected]>
>> >> >>
>> >> >> diff --git a/libavcodec/proresdsp.c b/libavcodec/proresdsp.c
>> >> >> index 7f20c83..d1506be 100644
>> >> >> --- a/libavcodec/proresdsp.c
>> >> >> +++ b/libavcodec/proresdsp.c
>> >> >> @@ -56,6 +56,8 @@ void ff_proresdsp_init(ProresDSPContext *dsp)
>> >> >>      dsp->idct_put = prores_idct_put_c;
>> >> >>      dsp->idct_permutation_type = FF_NO_IDCT_PERM;
>> >> >>
>> >> >> +    if (HAVE_MMX) ff_proresdsp_x86_init(dsp);
>> >> >
>> >> > Break the line please.
>> >> 
>> >> I think we do it like this everywhere in simd-init functions?
>> >
>> > We are all types of inconsistent in old files, but that is no reason
>> > to carry it forward to new files.
>> 
>> In this specific case, the one-line form actually does look better.
>> Compare:
>> 
>>     if (ARCH_FOO) ff_codec_init_foo();
>>     if (ARCH_BAR) ff_codec_init_bar();
>>     if (ARCH_BAZ) ff_codec_init_baz();
>> 
>> vs
>> 
>>     if (ARCH_FOO)
>>         ff_codec_init_foo();
>>     if (ARCH_BAR)
>>         ff_codec_init_bar();
>>     if (ARCH_BAZ)
>>         ff_codec_init_baz();
>
> Except for the fact that you have chosen three-letter arch names to
> make your example look extra good...

So add a little whitespace to make things line up.

> Creating arbitrary coding style exceptions is not a good idea, period.

It's not arbitrary.  It is using tabular formatting where the semantics
are tabular.

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

Reply via email to