On 07/09/16 10:00, Anton Khirnov wrote: > Quoting Diego Biurrun (2016-09-07 09:51:57) >> --- >> tests/checkasm/bswapdsp.c | 7 +++++-- >> tests/checkasm/checkasm.c | 14 +++++++++----- >> tests/checkasm/checkasm.h | 4 +++- >> tests/checkasm/dcadsp.c | 7 ++++--- >> tests/checkasm/fmtconvert.c | 1 + >> tests/checkasm/h264dsp.c | 9 ++++++--- >> tests/checkasm/h264pred.c | 9 ++++++--- >> tests/checkasm/h264qpel.c | 7 +++++-- >> tests/checkasm/hevc_mc.c | 6 +++--- >> tests/checkasm/synth_filter.c | 3 ++- >> tests/checkasm/v210enc.c | 7 +++++-- >> tests/checkasm/vp8dsp.c | 6 +++--- >> 12 files changed, 52 insertions(+), 28 deletions(-) > > IMO this just unnecessarily pollutes the git history and wastes people's > time without adding anything of any value. > And as I said in the other thread, I disagree that we have (or should > have) any agreed-upon "canonical" order.
Being consistent helps readability, but I do not consider it _that_ important to have long discussions over it. Even more since we are talking about checkasm and not part of the libraries. IIRC Diego wasn't even the person that proposed the strict ordering and the wiki history points me and Vittorio as the perpetrators. If you want it relaxed, I do not have problems with it anyway, the only care should be over not mixing system headers and local headers to avoid unexpected behaviors. lu _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
