Closed
Bug 960859
Opened 12 years ago
Closed 12 years ago
[e10s] Can't scroll content when text in chrome UI is focused
Categories
(Core :: DOM: Events, defect)
Tracking
()
VERIFIED
FIXED
mozilla29
Tracking | Status | |
---|---|---|
firefox31 | --- | verified |
People
(Reporter: reuben, Assigned: billm)
References
Details
Attachments
(1 file, 1 obsolete file)
1.21 KB,
patch
|
smaug
:
review+
|
Details | Diff | Splinter Review |
STR:
1) Click on awesome bar or search bar
2) Hover content with the mouse and try to scroll it
Reporter | ||
Comment 1•12 years ago
|
||
Untriaged because I don't know what component to put this under.
Assignee | ||
Comment 2•12 years ago
|
||
I'm just looking for feedback here since this patch is probably wrong. Here's what I'm seeing when I try this STR.
1. We enter nsEventStateManager::PreHandleEvent for an NS_WHEEL_WHEEL event. It sets mCurrentTargetContent to the be the URL bar (assuming it was in focus).
2. (In non-e10s builds) We enter nsEventStateManager::PostHandleEvent for the same event. aTargetFrame is a frame in the actual HTML document. All the scrolling stuff is done in terms of aTargetFrame.
2. (In e10s builds) We enter nsEventStateManager::PostHandleEvent and call HandleCrossProcessEvent. It looks at mCurrentTargetContent, which is still the URL bar, and ignores the event because mCurrentTargetContent isn't a remote target.
My patch just changes HandleCrossProcessEvent so that it uses aTargetFrame instead of mCurrentTargetContent. With this change, the mouse wheel works. I didn't notice any other problems, but I didn't do much testing.
I don't really understand the purpose of mCurrentTargetContent. Maybe you guys can explain it?
Attachment #8362736 -
Flags: feedback?(felipc)
Attachment #8362736 -
Flags: feedback?(bugs)
Assignee | ||
Updated•12 years ago
|
Component: Untriaged → DOM: Events
Product: Firefox → Core
Assignee | ||
Updated•12 years ago
|
Comment 3•12 years ago
|
||
Comment on attachment 8362736 [details] [diff] [review]
mouse-wheel-guess
Review of attachment 8362736 [details] [diff] [review]:
-----------------------------------------------------------------
Olli knows for sure, but this looks reasonable. I wonder if it's worth to keep the opposite logic, i.e. if aTargetFrame = null use mCurrentTargetContent
Attachment #8362736 -
Flags: feedback?(felipc) → feedback+
Comment 4•12 years ago
|
||
Could you make that code
nsIFrame* frame = GetEventTarget();
nsIContent* target = frame ? frame->GetContent() : null;
if (IsRemoteTarget(target)) {
targets.AppendElement(target);
}
(I need to cleanup mCurrentTargetContent handling a bit. That stuff is _old_.)
Updated•12 years ago
|
Attachment #8362736 -
Flags: feedback?(bugs)
Assignee | ||
Comment 5•12 years ago
|
||
Thanks for the help!
Assignee: nobody → wmccloskey
Attachment #8362736 -
Attachment is obsolete: true
Status: NEW → ASSIGNED
Attachment #8363884 -
Flags: review?(bugs)
Updated•12 years ago
|
Attachment #8363884 -
Flags: review?(bugs) → review+
Assignee | ||
Comment 6•12 years ago
|
||
Status: ASSIGNED → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla29
Comment 8•11 years ago
|
||
Reproduced in Nightly 2014-01-17, Win 7 x64.
Verified fixed FF 31 RC 2.
Status: RESOLVED → VERIFIED
status-firefox31:
--- → verified
You need to log in
before you can comment on or make changes to this bug.
Description
•