664720 - onmouseup event not fired when onmousedown generate a synchronous XHR

Status: NEW

(Core :: DOM: Events, defect)

Description

[Nicolas Faugout]

User-Agent:       Mozilla/5.0 (Windows NT 6.0) AppleWebKit/534.30 (KHTML, like Gecko) Chrome/12.0.742.91 Safari/534.30
Build Identifier: 4.0.1

When an HTML element such as a DIV has onmousedown and oumouseup events each associated with a javascript function and the onmousedown code creates a synchronous ajax request, then the onmouseup function is not called if user released the mouse button before the ajax request ends.

The URL show the problem. The file that is pointed by the AJAX request waits 200ms before ending. So if you make a mouse down/up shorter than 200ms, the mouseup function is not triggered.

Reproducible: Always

Steps to Reproduce:
1. Make a mouse down/up on the blue DIV element shorter than 200ms

Actual Results:  
The mouse up function is not called

Expected Results:  
The mouse up function should be called

There is no workaround that I could find

Comment 1

[Olli Pettay [:smaug][[e-mail]]]

Yeah, sync XHR messes with event handling in all the browsers.
It re-orders key events in non-gecko engines, and gecko has its own
problems like this.

Sync XHR is in general bad for UX, and I hope we could deprecate it.

But, however, this bug is perhaps something I could fix.

Comment 2

[Olli Pettay [:smaug][[e-mail]]]

I'll try to look at this soonish. But please remind me if I don't manage
to do anything in a week or so.

Comment 3

[Olli Pettay [:smaug][[e-mail]]]

Hmm, this should work.
Something has regressed or something...

Comment 4

[Nicolas Faugout]

It works fine on IE and Chrome.

Thanks for looking at it.

(In reply to comment #1)
> Yeah, sync XHR messes with event handling in all the browsers.
> It re-orders key events in non-gecko engines, and gecko has its own
> problems like this.
> 
> Sync XHR is in general bad for UX, and I hope we could deprecate it.
> 
> But, however, this bug is perhaps something I could fix.

Comment 5

[Nicolas Faugout]

Hi Olli,

You told me to remind you after one week, that's what I do.

I see that Firefox has released version 5 and insist for people to upgrade from version 4 to 5.

Do I should consider that Firefox 4 will soon be deprecated ?

I will test under Firefox 5 if the BUG still exists.

Comment 6

[Olli Pettay [:smaug][[e-mail]]]

(In reply to comment #5)
> You told me to remind you after one week, that's what I do.
Thanks!


> Do I should consider that Firefox 4 will soon be deprecated ?
There won't be any updates to Firefox 4. Firefox 5 is the update to
Firefox 4.

Comment 7

[Nicolas Faugout]

I installed Firefox 5. The BUG is still here. Do I need to file a new BUG for it ?

Comment 8

[Elbart]

Attachment: xhr_synchronous_e10s.png

Is an intermediate fix for this bug – and its dupes – on anybody's agenda, or is e10s going to be the designated bugfix? See screenshot.

Possible dupes:
https://bugzilla.mozilla.org/show_bug.cgi?id=636171#c4 (testcase)
https://bugzilla.mozilla.org/show_bug.cgi?id=672677
https://bugzilla.mozilla.org/show_bug.cgi?id=685971
https://bugzilla.mozilla.org/show_bug.cgi?id=687451
https://bugzilla.mozilla.org/show_bug.cgi?id=810691
https://bugzilla.mozilla.org/show_bug.cgi?id=867938
https://bugzilla.mozilla.org/show_bug.cgi?id=934422
Several testcases

Comment 9

[Olli Pettay [:smaug][[e-mail]]]

/me starts thinking how to remove support for sync XHR from main thread.

But I think this bug may need to be fixed before that. The problem is that all the browser
mess with the event ordering when sync XHR is involved. It just varies which events.

Comment 10

[Olli Pettay [:smaug][[e-mail]]]

See http://xhr.spec.whatwg.org/#sync-warning and recent discussion on
http://lists.w3.org/Archives/Public/public-webapps/

Comment 11

[Elbart]

> Warning: Synchronous XMLHttpRequest on the main thread is deprecated because of
> its detrimental effects to the end user's experience. For more help
> http://xhr.spec.whatwg.org/

That plea regarding the main-thread sounds a bit like "Your website breaks our browser and we don't want to or can't fix it. Please stop breaking our browser."

Comment 12

[Olli Pettay [:smaug][[e-mail]]]

No. It is about user experience.

Fortunately use of sync XHR is declining.

I would review a fix for this bug, but personally this isn't a high priority bug given that
sync XHR is deprecated and all the browser vendors want to get rid of it.
See for example http://blogs.msdn.com/b/wer/archive/2011/08/03/why-you-should-use-xmlhttprequest-asynchronously.aspx