[cairo] sincos patch final(?) patch
behdad at cs.toronto.edu
Tue Sep 13 07:49:45 PDT 2005
Apparently with that problem, sincos is not worth using.
I'm monitoring cairo-bugs now.
On Tue, 13 Sep 2005, Carl Worth wrote:
> On Mon, 12 Sep 2005 23:50:26 -0400 (EDT), Behdad Esfahbod wrote:
> > I updated the sincos patch to reflect Owen's feedback. I think
> > the patch can be applied now.
> [Here's an introduction to one thing I don't like about
> bugzilla. It walls off communication into small subsets of
> people. [*]]
> Cairo previously had one call to sincos in its implementation. As it
> turned out, that call caused problems on Tru64 where sincos apparently
> doesn't work. See this bug report for details:
> Bug 4401 - cairo detects and uses buggy sincos() on Tru64,
> In response to that bug report I recently removed the call to
> sincos. So before we could add it back again anywhere, we would at
> least have to have a workaround to avoid repeating this bug. But that
> leaves me asking if any of this is worth it. Do we have any
> measurements that actually indicate that using sincos is a significant
> performance win?
> [*] In one attempt to avoid this problem, I've set up a cairo-bugs
> mailing list to allow interested parties to follow all cairo bugzilla
> I'm subscribed to that and it seems to help me stay on top of things
> fairly well. I sure miss threading for all bugzilla-generated mail
> The link above contains a funky subscription process, (through the
> bugzilla watch feature), which does prevent bugzilla from sending
> duplicate messages to users who are also directly involved with a bug
> in some way.
More information about the cairo