Closed Bug 1214878 Opened 9 years ago Closed 9 years ago

Plugin is not hidden during async scrolling, lags instead

Categories

(Core :: Graphics: Layers, defect)

defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla44
Tracking Status
e10s m8+ ---
firefox44 --- fixed

People

(Reporter: botond, Assigned: jimm)

References

Details

Attachments

(2 files)

STR:
  1. Load http://telewizjarepublika.pl/szumlewicz-korwin-chce-dogadac-sie-z-pis-tylko-zjednoczona-lewica-jest-antysystemowa,24842.html
  2. Scroll down until you see the Flash video just above the comments
  3. Make sure you're running recent Nightly with APZ enabled, or
     enable APZ manually (layers.async-pan-zoom.enabled=true).
  3. Scroll up and down and observe the video.

The video's position lags behind as you scroll. I would have expected bug 1137944 to fix this by hiding the plugin during an async scroll.
Jim, would you expect the fix in bug 1137944 to apply to this case, or is there something special about this page?
Depends on: 1137944
Flags: needinfo?(jmathies)
Hmm, yeah this should get hidden. Something must be going wrong with our scroll tracking in the content process in nsGfxScrollFrame I guess. I'll take a look.
tracking-e10s: --- → ?
No longer depends on: 1137944
Flags: needinfo?(jmathies)
Ah this plugin is embedded in an iframe, bet that has something to do with it.
Assignee: nobody → jmathies
Attached patch patch (deleted) — Splinter Review
Need to walk sub documents as well.
Attachment #8674425 - Flags: review?(roc)
Comment on attachment 8674425 [details] [diff] [review]
patch

Review of attachment 8674425 [details] [diff] [review]:
-----------------------------------------------------------------

Oops. needs a test :-)
Attachment #8674425 - Flags: review?(roc) → review+
Attached patch tests (deleted) — Splinter Review
good idea. carbon copies of the original scroll tests with a sub frame plugin.
Attachment #8674951 - Flags: review?(roc)
Is is possible that we might have an iframe within an iframe that has a plugin? If so, I think I need to add yet another layer to this where I walk subframes, and then try to walk subframes within each sub frame, and so on down the tree.
Hmm, we don't do this in nsPresShell, so maybe EnumerateSubDocuments picks up all sub documents. I can't tell from the nsIDocument header. I'll test once this work gets out to mc.
https://hg.mozilla.org/mozilla-central/rev/59588063362c
https://hg.mozilla.org/mozilla-central/rev/eddce960bab6
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla44
@ Carsten, should it look like this? I've sent you an e-mail with video of this bug.
Flags: needinfo?(cbook)
(In reply to to_du from comment #12)
> @ Carsten, should it look like this? I've sent you an e-mail with video of
> this bug.

Hi, i'm not the initial developer who wrote the fix so i can't say (i only checked in the patch to the central repo). Jim is the developer who worked on this fix, redirecting the needinfo to him
Flags: needinfo?(cbook) → needinfo?(jmathies)
Can you send this video to Jim?
maybe it would be more useful to attach it here ?
I'm not sure what's being discussed, is this bug 1212813? If so attach it there. New regressions related to the work here (hiding plugins in subframes while scrolling) should be filed individually.
Flags: needinfo?(jmathies)
(In reply to to_du from comment #17)
> https://youtu.be/dWhFiQOufnw

This is not just bug 1212813; the video is only partially redrawn when scrolling stops. Please file a new bug for this.
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: