[cairo] Recommendations for 1.4.0

[Carl Worth]

On Sat, 03 Mar 2007 16:13:20 -0500, Behdad Esfahbod wrote:
> While updating docs for the 1.3.16 release yesterday I gathered the
> following that I suggest we fix before 1.4.0:

Thanks for the many documentation fixes.

>   - cairo_text_path() currently does not set the current point.  It will
> be implicitly set to the last point that an outline drew.  I recommend
> making it set the current point explicitly like cairo_show_text()
> does.

Yes, that should be done.

>   - Since _reference() and _destroy() functions just return on NULL,
> make _get_reference_count() return 0 on NULL.

Yes, that sounds good.

>   - Cosmetic: Rename cairo-font.c to cairo-font-face.c.

No, let's do this after 1.4, (I smell chances to break our currently
un-synced Makefiles for various platforms).

>   - Personal: In _cairo_path_append_to_context(), replace "!=" checks
> with "<".  That is, to allow more cairo_path_data_t elements in the path
> than needed.  I have requested this before:

Sure, easy enough.

>   - Make some of the test images smaller.  The 1.3.16 tarball is bigger
> than 4MB, while 1.2.0 was just 2.6MB and 1.0 only 1.4MB.  The main
> culprits seem to be extend-reflect and random-intersections reference
> images.  Maybe even replace romedalen.png with something smaller.

Ouch, that growth is pretty bad.

>   - Fix mutex initialization in win32, os2, and beos backends.  They are
> out of synch with the actually mutexes declared, and that causes link
> errors if not using pthread on win32.  I'm wondering if we should go for
> the backup plan:

I'm glad to see the new plan coming together now. But I'm hesitant to
do anything to major right now.

A problem that we have is that we don't have a lot of people doing
regular build-testing on non-Linux platforms between snapshots. So I
don't want to risk breaking the build again between 1.3.16 and 1.4.0.

We got reports of what was broken on win32 and we fixed that. I don't
think we've gotten any reports from os2 and beos, so we can just make
our best attempt, (without a compiler to verify).

The changes I accepted above look safe enough to implement and test on
one platform without a big risk of breaking other platforms.

I'll do what I can on these, (or Behdad, you can work on any you'd
like to as well).

Meanwhile, what I'll do today is roll up the various 1.3.x release
notes into a set of release notes giving an overview of what's changed
from 1.2 to 1.4 . I'll also do some cairo-perf testing from 1.2 to 1.4
so we can include a nice chart. I can easily do that on my x86 laptop,
but I don't think I'll have time to do any arm-based performance
runs. Anyone else want to take a whack at that now? Or maybe we'll
just have to let that appear after the release.

Anyway, things look good for 1.4 to happen today.

Thanks to everyone for the tremendous efforts over the last several
months.

-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.freedesktop.org/archives/cairo/attachments/20070305/0d22e66b/attachment-0001.pgp