421422 - Upgrade cairo to 1.5.12-56-ga33351f

Status: RESOLVED FIXED

(Core :: Graphics, defect, P1)

Description

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Tracking bug for bugs fixed by upgrading; ??? will be filled in when I do the upgrade, probably early next week.

Comment 1

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

pixman-0.9.6-43-gddfb69a (now with moar MMX!) is coming along for the ride.

Comment 2

[philippe (part-time)]

Does this patch touch text-rendering on Mac ?
Text looks more fuzzy and 'lighter' than it does on pre-patch builds, or with Safari.

With Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9b5pre) Gecko/2008031422 

Comment 3

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Some of the code changed, though I'm pretty sure that we don't actually set any of the flags that would cause us to hit the new code.  Can you attach some screenshots?  (Ideally with before-and-after.)

Comment 4

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Whoops, found it -- typo was accidentally forcing subpixel AA off.  Fix checked in.

Comment 5

[philippe (part-time)]

Attachment: screenshot, text on OS X

On the left a Camino build, around 17:00 PDT
right Minefield post patch

Look at the navigation bar at the top 'news' > 'world news'

From http://www.guardian.co.uk/world

Comment 6

[philippe (part-time)]

(In reply to comment #4)
> Whoops, found it -- typo was accidentally forcing subpixel AA off.  Fix checked
> in.
> 
Cool! That is much better now on a quickly made home-build Camino.

Comment 7

[[e-mail]]

Attachment: Fix unusual usage

For some reason the first typedef for ullong doesn't appear to be inside an #ifdef although the second one is. Also, there are a few cases that use unsigned long long where it looks as if it was meant to use ullong.

Sorry if this is the wrong bug but it was the one mentioned in your check in.

Comment 8

[Boris Zbarsky [:bzbarsky]]

So this drops support for gcc3, what with the *-growth CFLAGS.  Is there no way to not use those based on compiler?

Comment 9

[Peter Weilbacher]

Boris, I fixed that compiler thing for OS/2 in bug 423115 but if you know a way to make that depend on compiler version (using CC_VERSION?) instead of platform that would be much better.

Comment 10

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Comment on attachment 309783 [details] [diff] [review]
Fix unusual usage

Looks fine, I've pushed this upstream into pixman -- I'll pull it into our own tree early next week.