1229853 - Root container layer on Fennec has a big low-res displayport for no reason

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

[Kartikaya Gupta (email:[e-mail])]

I was investigating some of the talos regression with Fennec switched over to the C++ APZ and noticed that the root container layer on Fennec has a bogus low-res displayport. This is what it used to be before APZ:

ContainerLayerComposite (0x832e8000) [shadow-visible=< (x=0, y=0, w=768, h=1038); >] [visible=< (x=0, y=0, w=768, h=1038); >] [metrics0={ [cb=(x=0.000000, y=0.000000, w=768.000000, h=1038.000000)] [sr=(x=0.000000, y=0.000000, w=384.000000, h=519.000000)] [s=(0,0)] [dp=(x=0.000000, y=0.000000, w=0.000000, h=0.000000)] [cdp=(x=0.000000, y=0.000000, w=0.000000, h=0.000000)] [color=rgba(0, 0, 0, 0.000000)] [scrollId=0] [z=2] }]

This is what it is with APZ:

ContainerLayerComposite (0x89d64c00) [shadow-visible=< (x=0, y=0, w=768, h=1038); >] [visible=< (x=0, y=0, w=768, h=1038); >] [metrics0={ [cb=(x=0.000000, y=0.000000, w=768.000000, h=1038.000000)] [sr=(x=0.000000, y=0.000000, w=384.000000, h=519.000000)] [s=(0,0)] [dp=(x=-576.000000, y=-778.500000, w=1536.000000, h=2076.000000)] [cdp=(x=0.000000, y=0.000000, w=384.000000, h=519.000000)] [color=rgba(0, 0, 0, 0.000000)] [scrollId=2] [z=2] }]

Note that the "dp" value is pretty large and exceeds the scrollable rect. I looked into why this was happening and it's because there is no root scrollable frame for the root document in Fennec. So we end up applying the low-res multiplier at [1] without the clamping that would normally happen at [2].

[1] https://dxr.mozilla.org/mozilla-central/rev/470f4f8c2b2d6f82e56e161a4b05262c85f55b59/layout/base/nsLayoutUtils.cpp#906
[2] https://dxr.mozilla.org/mozilla-central/rev/470f4f8c2b2d6f82e56e161a4b05262c85f55b59/layout/base/nsLayoutUtils.cpp#1031

Comment 1

[Kartikaya Gupta (email:[e-mail])]

Attachment: Patch

Not sure if this will actually make any difference but I don't think there is much point to applying the multiplier in these cases.

Comment 2

[Kartikaya Gupta (email:[e-mail])]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=dcc24d73ffd7

Comment 3

[Timothy Nikkel (:tnikkel)]

Comment on attachment 8694834 [details] [diff] [review]
Patch

>   if (!frame) {
>     // Turns out we can't really compute it. Oops. We still should return
>     // something sane. Note that although we can apply the multiplier on the
>     // base rect here, we can't tile-align or clamp the rect without a frame.
>     NS_WARNING("Attempting to get a displayport from a content with no primary frame!");
>-    return ApplyRectMultiplier(base, aMultiplier);
>+    return base;
>   }

Can you update the comment here too?

Comment 4

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/50a3b107d896

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/50a3b107d896