854296 - SVG miterlimit rendered incorrectly with gfx.content.azure.enabled = true

Status: RESOLVED FIXED

(Core :: SVG, defect, P3)

Description

[Patrick Storz]

Attachment: Testcase showing the incorrectly rendered linejoins

When hardware-acceleration is turned on, the stroke property ‘stroke-miterlimit’ is rendered incorrectly:
Instead of switching from stroke-linejoin="miter" to stroke-linejoin="bevel" when the miter is getting too long, the miter is simply cut when reaching the miterlimit.

See the attachment for a testcase: The miter should only be visible for linejoins with 45° and 30°. For 15° and especially 0° the linejoin should be changed to bevel, therefore stopping at the vertical line (that is on the same x-coordinate than the right-hand nodes of the paths)


Furthermore even with hardware acceleration turned off the rendering seems to be not totally correct:
The linejoin for 15° isn't changed to bevel but one can clearly see two paths overlaid without real linejoin. The 0° linejoin reaches beyond the vertical line instead of stopping directly at the node.

Comment 1

[Alice0775 White]

setting gfx.content.azure.enabled = false helps


Regression window(m-c) with force gfx.content.azure.enabled = true
Good:
http://hg.mozilla.org/mozilla-central/rev/9c4f810196d2
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:12.0a1) Gecko/20120105 Firefox/12.0a1 ID:20120105060225
Bad:
http://hg.mozilla.org/mozilla-central/rev/595fba5b92d4
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:12.0a1) Gecko/20120105 Firefox/12.0a1 ID:20120105065423
Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=9c4f810196d2&tochange=595fba5b92d4

Regression window(m-i) with force gfx.content.azure.enabled = true
Good:
http://hg.mozilla.org/integration/mozilla-inbound/rev/c87c1aedefd4
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:12.0a1) Gecko/20120104 Firefox/12.0a1 ID:20120104212324
Bad:
http://hg.mozilla.org/integration/mozilla-inbound/rev/590333161e6f
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:12.0a1) Gecko/20120104 Firefox/12.0a1 ID:20120104231823
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=c87c1aedefd4&tochange=590333161e6f

Regressed by: Bug 711063

Comment 2

[Robert Longson [:longsonr]]

(In reply to Eduard Braun from comment #0)
> 
> Furthermore even with hardware acceleration turned off the rendering seems
> to be not totally correct:
> The linejoin for 15° isn't changed to bevel but one can clearly see two
> paths overlaid without real linejoin. The 0° linejoin reaches beyond the
> vertical line instead of stopping directly at the node.

You should raise other bugs for these cases.

Comment 4

[Jonathan Watt [:jwatt]]

Bas, can you take a look at this and see what's going on?

Comment 5

[Jonathan Watt [:jwatt]]

Attachment: testcase from bug 975958

Comment 7

[Jonathan Watt [:jwatt]]

(In reply to Alice0775 White from bug 975958 comment #0)
> Aaannndd... just to push me over the edge, same web page does not fail (all
> shapes display just fine) on Windows 7 with 27.0.1, while fails still on my
> Windows 8 also with 27.0.1.   Gaaaaa.....

Not sure why that would be the case. I doubt driver blacklisting affects this aspect of painting.

Comment 8

[Bas Schouten (:bas.schouten)]

Did your patches fix this problem jwatt?

Comment 9

[Jonathan Watt [:jwatt]]

Attachment: patch

No, but this will.

Comment 10

[Jonathan Watt [:jwatt]]

Attachment: document Moz2D's JoinStyle::MITER to note that its behavior is backend dependant

Comment 11

[Jonathan Watt [:jwatt]]

Comment on attachment 8537226 [details] [diff] [review]
patch

https://hg.mozilla.org/integration/mozilla-inbound/rev/91928ad7ef73

Comment 12

[Jonathan Watt [:jwatt]]

Comment on attachment 8537232 [details] [diff] [review]
document Moz2D's JoinStyle::MITER to note that its behavior is backend dependant

https://hg.mozilla.org/integration/mozilla-inbound/rev/83d940ea1491

Comment 13

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/91928ad7ef73
https://hg.mozilla.org/mozilla-central/rev/83d940ea1491