404092 - Upgrade Cairo to 1.5.2-55-g39b8ddf

Status: RESOLVED FIXED

(Core :: Graphics, defect, P2)

Description

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

Upgrade; lots of cleanup and correctness stuff in there, just tracking cairo for a 1.6 release.

Comment 1

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

Attachment: gfx changes needed for new cairo

I got rid of the ignore_operators flag in cairo; we can get the same effect in gfx via something like this.  I plan on extending this hints mechanism to do smarter things with pixel snapping and the like in the future.

Comment 2

[Stuart Parmenter]

As your comments say, printing SVG would break pretty bad with this.  I'm not sure we want to ever actually do this do we?

Comment 3

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

(In reply to comment #2)
> As your comments say, printing SVG would break pretty bad with this.  I'm not
> sure we want to ever actually do this do we?

SVG is already forced down this path currently, with no way to disable it.  With the hints mechanism it would let the SVG frame actually disable the optimization while its rendering SVG.

Comment 4

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

Attachment: gfx changes, v2

How about this?

Comment 5

[Mats Palmgren (inactive)]

+    if (mFlags && FLAG_SIMPLIFY_OPERATORS) {

That should be bit-and operator?

Comment 6

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

Whoops, yes, fixed.

Comment 7

[Reed Loden [:reed]]

This regressed (at least) Ts on Linux.

http://build-graphs.mozilla.org/graph/query.cgi?tbox=bl-bldlnx03_fx-linux-tbox-head&testname=startup&autoscale=1&size=&units=ms&ltype=&points=&showpoint=2007%3A11%3A29%3A16%3A16%3A37%2C894&avg=1&days=2

Comment 8

[Reed Loden [:reed]]

(In reply to comment #7)
> This regressed (at least) Ts on Linux.

roc's patch for bug 320378 just covered this up by acting as a perf win on top of this, but the perf regression is still there.

Comment 9

[Evan Yan]

Solaris tree was broken. see
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Ports/1196366880.1196367518.25310.gz&fulltext=1

cairo-paginated-surface.c
"cairo-paginated-surface.c", line 52: only identifiers with external linkage may have linker scoping: cairo_paginated_surface_backend

cc: acomp failed for cairo-paginated-surface.c
gmake[6]: *** [cairo-paginated-surface.o] Error 2
gmake[6]: Leaving directory `/export/home/mrbld/tinderbox/SunOS_5.11_Depend/mozilla/gfx/cairo/cairo/src'

Comment 10

[Evan Yan]

Attachment: remove cairo_private since we already declared static

cairo_private was defined as __hidden, it conflicts with static under Sun CC

  76 #define cairo_private_no_warn   __hidden

  86 #define cairo_private       cairo_private_no_warn cairo_warn

Comment 11

[Reed Loden [:reed]]

Comment on attachment 290834 [details] [diff] [review]
remove cairo_private since we already declared static

This bug is blocking1.9+, so approval is not required.

Comment 12

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

Checked in the solaris followup patch.

I'm not totally sold on Ts regressions being from this, as there was a Ts change even before this landed; but will resolve this for now to get it off lists.