736439 - Large parts of the page not painted on initial load

Status: RESOLVED DUPLICATE of bug 732016

(Firefox for Android Graveyard :: General, defect)

Description

[Martijn Wargers (dead)]

Attachment: Screenshot of bug (made on HTC Desire HD)

Steps to reproduce:
- Go to http://www.yahoo.com
- If the mobile version of Yahoo appears, choose the desktop version (on the bottom of the page you can tap on the Desktop version lin)

Expected result:
- Page painted correctly

Actual result:
- Large parts at the left and bottom not painted initially.

Comment 1

[Martijn Wargers (dead)]

I noticed this on the HTC Desire HD, Samsung Galaxy SII and the Samsung Galaxy Nexus. Sometimes the whole page is not painted at all.
This was tested on today's build (2012-03-16) of Fennec14.0a1.

Comment 2

[Brad Lassey [:blassey] (use needinfo?)]

possible dupe of bug 732119, I'll let Chris determine if it is or not though

Comment 3

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

I believe this is caused by bug 732016; when yahoo.com loads it puts focus into the search field at the top, so we call scrollIntoView on it. The scroll coordinates go negative, our display port goes haywire, and we end up painting only part of the visible area.

With some logging I have in my build I see this output:

03-17 02:41:47.582 D/GeckoLayerClient( 3167): VIEWPORT beginDrawing {"width":720,"height":514,"pageWidth":720,"pageHeight":1884.4897959183675,"zoom":0.7346938775510204,"x":22,"y":-330}
03-17 02:41:47.582 I/Gecko   ( 3167): VIEWPORT painting document in pres shell 0x5c093680
03-17 02:41:47.707 I/Gecko   ( 3167): Compositor: Layers update took 37 ms (blocking gecko).
03-17 02:41:47.707 D/GeckoLayerClient( 3167): VIEWPORT endDrawing
03-17 02:41:47.715 D/GeckoLayerClient( 3167): VIEWPORT setFirstPaintViewport: v=RectF(22.0, -330.0, 742.0, 184.0) p=(720.0,1884.0) z=0.734694

which shows a y=-330 coming from Gecko.

Comment 5

[Aaron Train [:aaronmt]]

Renaming summary as this affects other top-sites such as http://www.youtube.com

Comment 6

[Chris Lord [:cwiiis]]

For the record, I think this and bug 732119 are caused by different bugs. This, by the lack of scroll clamping, the other by possibly incorrect code in CompositorParent.

Comment 7

[Aaron Train [:aaronmt]]

Another blank page today: wordpress.com (03/29 Nightly)

Comment 8

[JP Rosevear [:jpr]]

Waiting to see how the fix in bug 732016 impacts this.

Comment 9

[Brad Lassey [:blassey] (use needinfo?)]

please comment out this line [https://mxr.mozilla.org/mozilla-central/source/mobile/android/chrome/content/browser.js#1560] and dupe to 732016 if that fixes this bug

Comment 10

[Aaron Train [:aaronmt]]

Yep.

Comment 11

[Martijn Wargers (dead)]

Now that bug 732016 is fixed I see on various pages content being white/not rendered, but not on the Yahoo page.

Comment 12

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

Which sites do you see it on? Could you attach a log?

Comment 13

[Martijn Wargers (dead)]

I guess this is bug 747842 now.