On Thu, Mar 01, 2012 at 10:41:19AM -0800, Ronald S. Bultje wrote:
> Hi,
> 
> On Thu, Mar 1, 2012 at 10:39 AM, Kostya Shishkov
> <[email protected]> wrote:
> > On Thu, Mar 01, 2012 at 10:31:57AM -0800, Ronald S. Bultje wrote:
> >> On Thu, Mar 1, 2012 at 10:26 AM, Kostya Shishkov
> >> <[email protected]> wrote:
> >> > On Thu, Mar 01, 2012 at 10:22:56AM -0800, Ronald S. Bultje wrote:
> >> >> 2012/3/1 Måns Rullgård <[email protected]>:
> >> >> > Kostya Shishkov <[email protected]> writes:
> >> >> >> On Thu, Mar 01, 2012 at 08:32:06AM -0800, Ronald S. Bultje wrote:
> >> >> >>> On Thu, Feb 16, 2012 at 9:19 AM, Ronald S. Bultje 
> >> >> >>> <[email protected]> wrote:
> >> >> >>> > On 64bit platforms with 32bit int, this means we won't have to 
> >> >> >>> > sign-
> >> >> >>> > extend the integer anymore.
> >> >> >>> > ---
> >> >> >>> >  libavcodec/arm/vp8dsp_init_arm.c |   32 ++++----
> >> >> >>> >  libavcodec/ppc/vp8dsp_altivec.c  |   16 ++--
> >> >> >>> >  libavcodec/vp8dsp.c              |   44 +++++-----
> >> >> >>> >  libavcodec/vp8dsp.h              |   38 +++++-----
> >> >> >>> >  libavcodec/x86/vp8dsp-init.c     |  158 
> >> >> >>> > +++++++++++++++++++-------------------
> >> >> >>> >  5 files changed, 144 insertions(+), 144 deletions(-)
> >> >> >>>
> >> >> >>> Ping.
> >> >> >>
> >> >> >> is it still intptr?
> >> >> >
> >> >> > I still think these should be ptrdiff_t.
> >> >>
> >> >> See old thread, some people object to intptr_t and others object to
> >> >> ptrdiff_t, for orthogonal reasons. Someone had to make a decision,
> >> >> that person was me, and it is intptr_t. This is consistent with what
> >> >> x264 does and thus probably a good long-term idea anyway. (Also please
> >> >> note that they're really always the same, so we're talking pure theory
> >> >> bs here.)
> >> >
> >> > And someone else made a decision naming a library libtidsp.
> >>
> >> This discussion was had, sorry, I could only please 50%, I decided to
> >> please the others. Reverting that will displease the other 50%. Please
> >> move on.
> >>
> >> Now can we review the patch?
> >
> > Using intptr_t type for something that does _not_ contain actual pointer is
> > misleading and that's why it's _not_ OK.
> 
> I understand, but again, we had this discussion, it's 50/50 and
> someone has to choose. We can't bikeshed over color forever, it was
> causing actual crashes.

Then always pick ptrdiff_t over intptr_t:
0) it does what's needed
1) it has right definition and purpose (addressing and pointer arithmetics)
2) it's POSIX instead of XSI
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to