[cairo] Cairo 1.6 is imminent

[Carl Worth]

On Tue, 08 Apr 2008 20:36:58 +0100, Chris Wilson wrote:
> > From the roadmap: (http://cairographics.org/roadmap/), the only
> > changes we expect to make are:
> ...
> > 	* Fixes for any leaks found with "make check-valgrind" on the
> >           test suite.
>
> I've run check-valgrind on a couple of machines and whilst I may have
> missed one or two small items, there is one major issue - a reference
> leak in self-copy. As far as I can tell, there are at least three
> different problems. The first is with the ps backend, the second is with
> similar surfaces and the third is on the error paths. IMHO a non-trivial
> bug.

I let this run on my machine last night as well, (it does take a
fairly long time). And I also didn't see any leaks other than this
thing in self-copy. If Chris looked closely and decided that this is
non-trivial, then I'm taking his word on that, and deciding that we
won't try to fix this issue before 1.6.

Meanwhile, here are the other two things I mentioned possibly doing
before 1.6:

	* Any obvious fixes for legitimate compiler warnings.

I had recently cleaned up all warnings on one system, but I had also
noticed that another system still spewed a lot. I double-checked that
and almost all of the warnings are "unused parameter", (note that that
is unused *parameter*, not unused variable). This is a mostly
worthless warning, (it appears here because we have functions that
must implement a particular interface even if they don't care about a
particular parameter).

So I think we'll just want to turn that warning off, (either globally
or with some attributes or whatever). But that's not worth touching
now I don't think.

Another one was "cannot optimize possibly infinite loops" which is
again something I don't want to touch.

	* Potentially a tweak to the buggy_repeat decision. I'm
          currently exploring various X server builds to find out
          where the bug existed and where the fix occurred.

Here, I think we've done the best we can with the reports we have
received so far. So any further change will have to wait for later.

Meanwhile, I've heard back from some mozilla and inkscape developers
that have started using the 1.5.20 snapshot and haven't heard any
scary noises yet. So my inclination is to relabel the exact 1.5.20
code as the cairo 1.6.0 release. Please speak now if you have any good
reason why we shouldn't do that.

-Carl
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20080409/d7e0f5e7/attachment.pgp