Closed Bug 639237 Opened 13 years ago Closed 3 years ago

Mouse pointer disappears/flickers when the Adobe Flash plugin is loaded

Categories

(Core Graveyard :: Plug-ins, defect, P3)

x86
All

Tracking

(e10s-)

RESOLVED INCOMPLETE
Tracking Status
e10s - ---

People

(Reporter: giana_, Unassigned)

References

Details

Attachments

(1 file)

User-Agent:       Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:2.0b13pre) Gecko/20110305 Firefox/4.0b13pre
Build Identifier: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:2.0b13pre) Gecko/20110305 Firefox/4.0b13pre

Once the Flash plugin is loaded on start-up of Firefox, the mouse pointer will disappear at times.

It is highly annoying and a major bug.

Reproducible: Always

Steps to Reproduce:
1. Get the latest version of Firefox/Minefield and install the latest Adobe Flash plugin.

2. Enable the Flash plugin. (Everything will look alright)

3. Restart Firefox/Minefield.

4. Open up a few tabs with some sites loaded. At least one of them should have a Flash ad/video running.

5. Go to any of those sites and watch the mouse pointer disappear and reappear randomly.
   Once the mouse pointer stops moving it will disappear from the screen.

6. Disable the Flash plugin.

7. Restart Firefox/Minefield.

8. Now all is back to normal function except that Flash is disabled.
Actual Results:  
Please read above :)

Expected Results:  
The mouse pointer should not disappear!

I have tried these combinations:

Firefox 3.6.15 32Bit & Adobe Flash 32Bit = Fine

Firefox 4.0 Beta 12 32Bit & Adobe Flash 32Bit = Broken

Minefield 4.0b13pre (2011-03-05) 32Bit & Adobe Flash 32Bit = Broken

Minefield 4.0b13pre (2011-03-05) 64bit & Adobe Flash 64Bit = Broken

A lame workaround would be to have the Flash plugin disabled at start-up and then load it manually. When you finish browsing you have to disable the plugin before exiting the browser.
Does this occur if you disable hardware acceleration in prefs (advanced -> general), then restart Firefox for the changes to take effect?
Component: Extension Compatibility → Plug-ins
Keywords: regression
Product: Firefox → Core
QA Contact: extension.compatibility → plugins
Version: unspecified → Trunk
Also, please only try with x86 builds for now; 64bit Windows builds aren't officially supported.
(In reply to comment #2)
> Also, please only try with x86 builds for now; 64bit Windows builds aren't
> officially supported.

Hi :)

1. Yes, i already tried 32Bit and 64Bit versions of Firefox 4 and Adobe Flash. It doesn't matter which one i use.

2. Tools > Options > Advanced > General > Browsing > Use hardware acceleration when available
   Turning this off and restarting the browser doesn't help either.

Firefox 3.6.15 is fine btw.
Whilst it occurs for you with both 64bit and 32; please test just with 32bit for now, to reduce the chances of any of the open 64bit bugs affecting the testing.

Could you see if the issue occurs if using Firefox in safe mode:
http://support.mozilla.com/kb/Safe+Mode

How about with a new, empty profile?
http://support.mozilla.com/kb/Basic+Troubleshooting#w_make-a-new-profile
Hardware: x86_64 → x86
Sorry, should have added:

Please paste the contents of about:plugins here.
Minefield 64Bit 4.0b13pre (2011-03-06):

Enabled plugins

Shockwave Flash

    File: NPSWF64_10_3_162.dll
    Version: 10.3.162.28
    Shockwave Flash 10.3 d162

MIME Type 	Description 	Suffixes
application/x-shockwave-flash 	Adobe Flash movie 	swf
application/futuresplash 	FutureSplash movie 	spl

Microsoft Office 2010

    File: NPAUTHZ.DLL
    Version: 14.0.4730.1010
    Office Authorization plug-in for NPAPI browsers

MIME Type 	Description 	Suffixes
application/x-msoffice14 	14.0.4730.1010 	*

-----------------------------------------------------------

Minefield 32Bit 4.0b13pre (2011-03-05):

Enabled plugins

Shockwave Flash

    File: NPSWF32.dll
    Version: 10.2.152.32
    Shockwave Flash 10.2 r152

MIME Type 	Description 	Suffixes
application/x-shockwave-flash 	Adobe Flash movie 	swf
application/futuresplash 	FutureSplash movie 	spl

Microsoft Office 2010

    File: NPAUTHZ.DLL
    Version: 14.0.4730.1010
    Office Authorization plug-in for NPAPI browsers

MIME Type 	Description 	Suffixes
application/x-msoffice14 	14.0.4730.1010 	*

Tried running in Safe Mode but it doesn't help.
Tried a new profile but it doesn't help either.

I have now narrowed down the problem to this site:

http://bohuslaningen.se/

Disable the Flash plugin on that site and the problem goes away...

I'm not sure how many sites that are affected by this :)
Small update: The bug is visible in:

IE8

But not in:

Minefield 4.0b13pre (2011-03-05) Mac OS X
Opera (Windows)
Safari (Windows)

And before someone asks, yes i have updated my mouse drivers as well :)
I cannot reproduce this with Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.0) Gecko/20100101 Firefox/4.0 with HW acceleration on or off
Just downloaded Minefield x32 4.0b13pre (2011-03-11) and the issue is still there.

Was your Flash plugin enabled? I assume it was though.

Did you wait 5 seconds or so? It takes a few seconds before the mouse pointer disappears.

Which mouse do you use?

I use a Logitech MX 518
Ok I've rechecked on youtube (can't reproduce) and also tried bbc news videos, where if I right click on the video and then leave the mouse it will disappear until I move it out of the video area.
So you didn't try this?

"I have now narrowed down the problem to this site:

http://bohuslaningen.se/

Disable the Flash plugin on that site and the problem goes away..."
(In reply to comment #11)
> So you didn't try this?
> 
> "I have now narrowed down the problem to this site:
> 
> http://bohuslaningen.se/
> 
> Disable the Flash plugin on that site and the problem goes away..."

I can't reproduce any problem with the mouse pointer disappearing on that site with flash enabled.

The disappearing cursor with flash issue has been noticed before, see Bug 582622
I can reproduce the bug you linked to so that issue isn't fixed either.

I'm beginning to wonder if Adobe should take a look at these issues. Might be a bug in their player. Either that or something is going on with the plugin-container.exe :)
Reporter, please can you confirm whether this issue still occurs using Firefox 4.0.1 (http://www.mozilla.com/firefox/new/) or higher, in Firefox safe mode (http://support.mozilla.com/kb/Safe+Mode) and/or with a clean profile (http://support.mozilla.com/kb/Basic+Troubleshooting#w_8-make-a-new-profile). Ideally, also check using the latest nightly: http://nightly.mozilla.org/

If this issue no longer occurs, please close as "Resolved Worksforme".

It it in fact still occurs, please provide as much extra information as possible, including what versions tried, whether safe mode/new profile tested etc. Thanks! :-)

(Template reply to inactive UNCO bugs)
I have this issue occuring intermittently to me in Firefox 11.0 on OSX 10.6.8 using Flash 11.1.102.64 

I have noticed it after the ABC iView plugin has opened up to fullscreen and I have put it back to windowed mode.

Steps that might help reproduce the bug:

* Open http://www.abc.net.au/iview/ in one tab 
* Play a movie in fullscreen
* After it has gone to full screen reduce it to windowed mode using ESC. The button provided to switch back to windowed mode doesn't seem to trigger the issue from my experimentation.
* Open other tabs
* Split off iView tab into a new window next to the current window (which sometimes crashes Firefox or Flash for me)
* In some cases the cursor now disappears whenever it is over Firefox, which makes it impossible to do anything, although it appears if you hover over the desktop

Although I can't reliably reproduce the cause, the cursor immediately appears properly again when I close the iView tab so I am fairly sure it is related to Flash.
Build identifier: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:14.0) Gecko/20100101 Firefox/14.0

Similar problem - removing focus from the main window brought the cursor back. Putting focus on the main window disappeared the cursor. Disabling the flash plugin solved the problem.

I tried retracing my recent history, or some of the sites above, but can't make it come back now that I've fixed it via flash disable.
(In reply to Steve Scott (pxbugz) from comment #12)
> The disappearing cursor with flash issue has been noticed before, see Bug 582622

An even older report: bug 243718.
Unable to reproduce:
    File: NPSWF32.dll
    Version: 10.2.152.32
    Shockwave Flash 10.2 r152
Firefox Version      4.0 or 4.0b1
Is this error still occurring?
Also Cannot Reproduce with latest Adobe Flash and Firefox Version 	43.0.1
Build ID 	20151216175450
User Agent 	Mozilla/5.0 (Windows NT 5.1; rv:43.0) Gecko/20100101 Firefox/43.0
Flags: needinfo?(giana_)
I was all set to reply and say I'd not seen it happen for a long time. Then trawled through all the example website links and found it DOES still occur. This link started the cursor vanishing/flickering for me: http://www.kongregate.com/games/ImpendingRiot/alice-is-dead-ep-2
Forgot to add: 

Build identifier: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:45.0) Gecko/20100101 Firefox/45.0
Flash Version: 21.0.0.182
Yes - Thanks I can reproduce and see the problem
Adobe Flash Player Version 21.0.0.182
Version 	45.0
Build ID 	20160303134406
User Agent 	Mozilla/5.0 (Windows NT 6.1; WOW64; rv:45.0) Gecko/20100101 Firefox/45.0
Setting to New/Confirmed, I don't think MozRegression will be much help as this effects a lot of Adobe and Firefox versions. Would output from Browser Console help?
Status: UNCONFIRMED → NEW
Ever confirmed: true
Keywords: regression
Hello, I was able to reproduce this issue on Firefox 45.0 and Firefox 47.0 with following Shockwave versions:
Shockwave Flash 21.0 r0
Shockwave Flash 22.0 r0 
 
I am observing continious vanishing/flickering with the flash plugin.
Jim I think this is worth investigating. Agree, and can you find an owner?
Flags: needinfo?(giana_) → needinfo?(jmathies)
Priority: -- → P2
Is anyone seeing this on any site other than http://www.bohuslaningen.se/ ? Currently that's site isn't loading properly for me.
This issue is still valid in FF 47.0 Stable and the bug is also present in FF Aurora 49.0a2 (2016-06-20) and I updated the Flash plug-in a few minutes ago. You can see the issue on this website (as mentioned by  the user; WKDPOWER):

http://www.kongregate.com/games/ImpendingRiot/alice-is-dead-ep-2
Edit: This can also be reproduced on OS X 10.11.5 so it's not a Windows-specific issue.
Flags: needinfo?(jmathies)
OS: Windows 7 → All
Not e10s specific either, I can reproduce in both modes.
This is just flash calling the standard SetCursor api ignoring that the cursor isn't within the bounds of the plugin drawing area or window.

I put together a patch that hooks SetCursor and checks the cursor position against known plugin windows and was able to fix this for the test case mentioned here. However the fix breaks cursor changes in windowless flash since there's no child window. The check looks something like this:

POINT cursorPos;
HWND hwnd;
PluginInstanceChild* inst;
if (::GetCursorPos(&cursorPos) &&
    (hwnd = ::WindowFromPoint(cursorPos)) &&
    (inst = reinterpret_cast<PluginInstanceChild*>(
      ::GetProp(hwnd, kPluginInstanceChildProperty))) &&
    inst->mPluginWindowHWND == hwnd) {
    return sUser32SetCursorHookStub(hCursor);
}
return NULL;

I'm looking at moving this to PluginModuleChild so that I can iterate all plugins, check the type, and then do appropriate bounds checks.
(In reply to P A from comment #28)
> Edit: This can also be reproduced on OS X 10.11.5 so it's not a
> Windows-specific issue.

I almost forgot, I need to look at osx as well. It's pretty clear what's going on in Windows. osx might be a similar issue.
I looked into moving the SetCursor hook over to PluginModuleChild. To make this approach work I'd need access to the list of plugin instances stored in mManagedPPluginInstanceChild. I can't access this from within the SetCursor callback without a static pointer to the parent PluginModuleChild. In single process e10s we have two of these, a chrome side instance that's connected to the chrome process through the plugin bridge, and a content instance that spawns child instances for content. I could store a static pointer and get at mManagedPPluginInstanceChild for this mode, however with multiple content processes we'll have multiple content side PluginModuleChild(ren). When that gets introduced a single static PluginModuleChild pointer will break.

So far I'm not seeing a way to address this.
tracking-e10s: --- → -
Blassey, do we have any contacts at adobe we can reach out to? They should be able to address this pretty easily on their end by restricting SetCursor calls when the mouse is within the area of their plugin's drawable area.
Flags: needinfo?(blassey.bugs)
Flags: needinfo?(blassey.bugs)
This is low priority. Long standing issue, applet specific vs. a global issue with all flash, and is reproducible with e10s and non-e10s. The plan is to ask adobe about this once they finish up with more pressing updates to the flash runtime.
QA Update: I tested this with Flash Beta v 23
E10s enabled can still see flickering
E10s disabled I am not seeing mouse flickering.
Reference attached "FXAFB23 file for full details and "ni" me if additional info is needed.
Flags: needinfo?(jmathies)
Attached file FXAFB23.rtf
(In reply to Michelle Funches - QA from comment #35)
> QA Update: I tested this with Flash Beta v 23
> E10s enabled can still see flickering
> E10s disabled I am still seeing mouse flickering. (miss reported earlier)
> Reference attached "FXAFB23 file for full details and "ni" me if additional
> info is needed.
Flags: needinfo?(jmathies)
Moving to p3 because no activity for at least 1 year(s).
See https://github.com/mozilla/bug-handling/blob/master/policy/triage-bugzilla.md#how-do-you-triage for more information
Priority: P2 → P3

Flash is no longer supported.

Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → INCOMPLETE
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: