56240 - onMouseDown state gets stuck down with popup/app switch

Status: VERIFIED FIXED

(Core :: DOM: UI Events & Focus Handling, defect, P2)

Description

[Jesse Ruderman]

After cancelling an alert from an onmousedown event, moving the mouse selects 
text even though I'm no longer holding the mouse button down.  This happens 
whether I cancel the alert by clicking "ok" or by pressing enter/esc.

Comment 1

[Jesse Ruderman]

Attachment: testcase

Comment 2

[Johnny Stenback (:jst)]

Event problem, reassigning to Tom.

Comment 3

[Loco]

Hmmm...I think this is a different incarnation of a larger issue.  If you hover 
the mouse over a button that has an onMouseDown event (after dismissing the alert 
box) the button is depressed.  Until focus shifts elsewhere.  Changing Summary, 
Plat/OS to ALL, and taking ownership of the bug under Event Handling.  desale, 
you are free to go:).
Previous Summary:
After cancelling onmousedown alert, moving the mouse selects text

Comment 4

[Jesse Ruderman]

This bug can also be triggered by click-dragging out of the content area, alt-
tabbing to another app, and then alt-tabbing back.  (That happens for 
selectable text and html3/html4 buttons.)  There are already several bugs 
involving alt-tab and mouse events:

bug 33144 toolbar buttons stay depressed after click-drag, alt-tab

          Note: you don't have to drag out of anything to trigger 33144,
          contrary to the summary.  Click-hold works fine.

bug 35191 alt-tab, mouseup on other app while scrolling -> scrollbar grabs
          mouse focus

I wonder if some of these bugs are duplicates.

Comment 5

[joki (gone)]

I think this is a dupe of some others.  Its a basic issue that we need to 
synthesize a mousedown in a bunch of cases where we didn't get one.

Comment 6

[joki (gone)]

*** Bug 65344 has been marked as a duplicate of this bug. ***

Comment 7

[Hixie (not reading bugmail)]

Reassigning QA Contact for all open and unverified bugs previously under Lorca's
care to Gerardo as per phone conversation this morning.

Comment 8

[joki (gone)]

This fix seems like it might have some unintended consequences.  I'm gonna push 
it out to 0.9.1 rather than shove it into 0.9 tonight.

Comment 9

[clayton]

Moving out one milestone.  Joki has too many bugs.

Comment 10

[Madhur Bhatia]

QA contact updated

Comment 11

[joki (gone)]

Moving lower priority bugs out of .9.2 milestone.

Comment 12

[Blake Ross]

Doesn't look like this is getting fixed before the freeze tomorrow night.
Pushing out a milestone.  Please correct if I'm mistaken.

Comment 13

[saari (gone)]

->0.9.5

Comment 14

[joki (gone)]

Moving to 0.9.6

Comment 15

[joki (gone)]

*** Bug 102421 has been marked as a duplicate of this bug. ***

Comment 16

[joki (gone)]

0.9.6 has passed, moving to 0.9.7.  Load balancing 0.9.7 list shortly

Comment 17

[joki (gone)]

Moving bugs from 0.9.7 for triaging in 0.9.8

Comment 18

[R.K.Aa.]

*** Bug 120289 has been marked as a duplicate of this bug. ***

Comment 19

[R.K.Aa.]

*** Bug 102718 has been marked as a duplicate of this bug. ***

Comment 20

[Peter Trudelle]

nsbeta1+ per ADT triage

Comment 21

[joki (gone)]

This bug has a lot of incarnations since many of the problems areas track their
own state (selection state, button state, mouse capture, scroll state, etc).  I
have a fix for the test cases on this bug which I going foward with. It fixes
selection state adn mouse capture state. It still leaves 35191, the scrolling
issue and 33144, the button issue.

Comment 22

[joki (gone)]

Attachment: proposed patch

Comment 23

[saari (gone)]

Comment on attachment 70449 [details] [diff] [review]
proposed patch

As long as you've tested drag and drop and it works okay, r=saari

Comment 24

[Johnny Stenback (:jst)]

Comment on attachment 70449 [details] [diff] [review]
proposed patch

sr=jst

Comment 25

[joki (gone)]

Fix checked in to cover selection defects and mouse capture defects (see dupe
bug 102421).  A number of the dupes of this bug cover scrollbar defects which is
also covered in bug 35191.  Since the internal issues, though similar, must be
fixed in a number of locations I'm going to continue to track this as separate
bugs and fix them individually.  So for verification purposes, this fix does NOT
cover the scrollbar-based dupes listed on it.  Other issues listed within
should, however, be fixed and can be verified.

Comment 26

[Mike Kaply [:mkaply]]

This fix broke selection on SOME pages for OS/2 and I have no idea why.
We just discovered the regression.

An example page would be:

http://en.ecomstation.ru/

If I comment out the 

 mNormalLMouseEventInProcess = PR_TRUE;

in nsEventManager.cpp - NS_MOUSE_LEFT_BUTTON_DOWN, everything works as expected.

Does anyone have any ideas?

Thanks

Comment 27

[joki (gone)]

I couldn't find a new bug about the 0S/2 selection problem so I opened a new 
one, 138265.

Comment 28

[Rakesh Mishra]

Verified on the trunk build 2002-04-24-06-trunk on Windows 2000 .

Comment 29

[R.K.Aa.]

bug 144954 was just filed with 2002050521, linux.
Is this one really fixed?

Comment 30

[R.B.]

Re Comment #29: The fix does not cover all cases, in particular not the
scrolling/popup, see Comment #25.

Comment 31

[R.K.Aa.]

*** Bug 180359 has been marked as a duplicate of this bug. ***

Comment 32

[R.K.Aa.]

*** Bug 210333 has been marked as a duplicate of this bug. ***