Closed GoogleCodeExporter closed 9 years ago
The effect is not proportional. If the text size is made 10px, then the stroke
appears correctly even with the rectangle.
Original comment by classi...@floodgap.com
on 23 Dec 2012 at 1:00
If we take the shadow off the canvas context, then the outline appears
correctly also:
http://www.floodgap.com/software/tenfourfox/bleh/gracs.html
Original comment by classi...@floodgap.com
on 23 Dec 2012 at 5:01
Rendering is correct in 19.0a2 on 10.6, so this is our bug.
Original comment by classi...@floodgap.com
on 23 Dec 2012 at 5:26
Happens in 10.5 with Aurora 18 and 19, but NOT in the latest build (2012-12-22).
I suspected it had to do with M818004 (you did apply that patch?) but the first
build of Aurora that had that fix applied (2012-12-15) shows the same error.
On the other hand 2012-12-22 was my first build that had the Altivec blurring
enabled and it's also the first one that doesn't show this error.
(There's a different bug in AuroraFox 19 I noticed because of the samples,
though: while the stroke is correct and has the correct character spacing, the
coloured letters and the shadow have a different character spacing.)
Original comment by Tobias.N...@gmail.com
on 23 Dec 2012 at 2:17
The Altivec accelerated blurring is NOT the reason this does work in Aurora 19
build 2012-12-22. So this has been fixed in the mozilla sources between
2012-12-15 and 2012-12-22.
Original comment by Tobias.N...@gmail.com
on 23 Dec 2012 at 2:24
While this stroke bug happens (or rather happened) independently of the azure
backend used (in AuroraFox I can choose between Cairo, Skia and CoreGraphics)
the character spacing bug is different across all those backends here. Skia is
the only one that gets it right here. Cairo draws all the coloured letter in
the place of the first one, but the shadows are placed correctly. CoreGraphics,
like already described, shows larger character spacing between the coloured
letters and theirs shadows compared to the spacing between the strokes.
Original comment by Tobias.N...@gmail.com
on 23 Dec 2012 at 2:52
Looks like https://bugzilla.mozilla.org/show_bug.cgi?id=815648 -- the testcase
is nearly identical and as you say, Skia does get it right, but Skia doesn't
work on 10.4 and I have no plans to fix that since CG is good enough.
So we'll recheck this after the beta uplift.
Original comment by classi...@floodgap.com
on 23 Dec 2012 at 11:44
Test case now works.
Original comment by classi...@floodgap.com
on 12 Jan 2013 at 8:20
Original issue reported on code.google.com by
classi...@floodgap.com
on 23 Dec 2012 at 12:52