Still happens with 1.4.0b4

Another test case:
http://www.kylehuey.com/moz/testRequest.php (via:
https://bugzilla.mozilla.org/show_bug.cgi?id=501962)

Please try Firebug 1.5Xa8, http://getfirebug.com/releases/firebug/1.5X. This is the
tracing version and it has one more fix for a problem related to watching ajax
events. If you still see the issue, open Firebug Icon Menu Open Tracing Console,
options ERRORS, NET, WINDOWS and see if anything fun happens.

I'm still able to reproduce this issue with 1.5Xa8. I've attached the log from the
tracing console.

firebug-tracing-logs.txt 5.3 KB   View   Download

Looks like  issue 1971 .

John -- this is a bug with potentially significant impact for Yahoo and other sites, given the number of 
erroneous defect reports we're likely to get -- what can we do to escalate this? --Eric (Eric Miraglia, Yahoo/YUI 
team)

I need to know what happens in other cases like this one when using Firebug 1.5X.0a9.  

John - did you take a look at the trace log I attached? It definitely shows errors in
the readystatechange handler in spy. Do you need something more than that?

Yes I looked at the trace; It looks like  issue 1971 . 

I need someone to try 1.5a9 and see what they get.

I am still having problems with this. Firing XMLHttpRequests works with onload
events, but when something else on the page triggers it - Firebug does not register it.

$$('input.test').addEvent('click', function() {
    // Firebug does not pick this up and this is a POST
    new Ajax().send();
});

I got this to work by turning Firebug on for all pages.

The problem in comment 11 sounds like a different one from  issue 1971 . If I had a
complete test case I could fix it.

I tried reproducing this with both example test cases and wasn't able to reproduce
it. I tried reloading the first one probably a dozen times.

The second test case in comment 3 I loaded up with a bunch of requests and they all
came back (with both 1 and 3s delays). If anyone has a more-easily reproducible test
case, I'd like to see it.

PS, I'm on OS X 10.5.7 with Firebug 1.4b5 and Firefox 3.5 final release.

A little more information:

I'm able to reproduce this using http://www.kylehuey.com/moz/testRequest.php but not
with the other one. If I set the delay to 10 seconds, and open the Firebug console, I
typically get between 2-5 occurrences of this for every 10 requests I try.

I'm using Firefox 3.5 final release with Firebug 1.5a8 on Windows XP. I'll give 1.5a9
a try and let you know.

(No comment was entered for this change.)

I just installed 1.5a9 and am able to reproduce the issue, too. I've attached the
trace log. Also, I see the following in the console:

onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]
onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]
onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]

firebug-tracing-logs.txt 1.8 KB   View   Download

Im having the same issue, very painfull to use on sites that have multiple ajax requests

Please be specific about the "same issue".

This bug is fixed in 1.4b6 and 1.5a9: the AJAX call back does fire as afar as I know.

But there is an exception some of the time.  I don't know the cause.

Here is the example:

http://revolutionremix.com/lag/a.html
With Firebug OFF:
Click on the button "clicker" till you have around 50 ajax queues, They will all
complete successfully.
Now, click another 50 times on the button "clicker" and usually in no more than 10
in-a-row ajax queries it will throw an exeption and die.

This issue sometimes happens instantly, so its not like im pushing it to give an
error, im just pushing to see that there is a random error.

This is the exeption you will get:

onreadystatechange FAILS Error: Permiso denegado para <http://revolutionremix.com>
para crear un wrapper para el objeto de clase UnnamedClass Error: Permiso denegado
para <http://revolutionremix.com> para crear un wrapper para el objeto de clase
UnnamedClass [xpconnect wrapped nsIDOMEventListener]


Tested on
Firefox 3.5 + FireBug 1.4.0b6 (with FB disabled, it works flawless)

Sorry about a mispell in the previous text, here is the fixed test-case:

With Firebug OFF:
Click on the button "clicker" till you have around 50 ajax queues, They will all
complete successfully.

ENABLE FUREBUG:
Now, click another 50 times on the button "clicker" and usually in no more than 10
in-a-row ajax queries it will throw an exeption and die.

It's definitely not fixed in 1.5a9, as I said in my comment, I'm still able to
reproduce it with some regularity.

  The call back either fires or the call back fails with an exception. The current
evidence is that this is a Firefox 3.5 problem.

How can that be possible if it only happens when FB is enabled?

https://bugzilla.mozilla.org/show_bug.cgi?id=502959

This was reported earlier to Firefox: https://bugzilla.mozilla.org/show_bug.cgi?id=501962

Probably the most important information now is: what is different between the cases
that work and the ones that don't?  Something is causing Firefox to take a different
code path, may be cache hits or net delays or other browser activity or ?

i vote for net delay issue.
go to the address in Comment 3
http://www.kylehuey.com/moz/testRequest.php
and write 10 in the 'Wait time' field.
every try will fail with long wait times.

i tested it with 1.5a10

I confirm that going to http://www.kylehuey.com/moz/testRequest.php and setting 10s I 
get the error on each run with winXP FF3.5 + fbug 1.4b7

I get

onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]

it works perfectly when firebug is closed.

Is it confirmed by firefox team that this is a firefox bug ?
reading through https://bugzilla.mozilla.org/show_bug.cgi?id=501962 it does not seem 
that clear to me.

We don't know the problem here. But we can not make progress without help from the
Firefox team, unless you can code that error message better than I can.

This may seem completely crazy (so feel free to tell me so), but would it be possible
to get a build of these Firebug versions to run on Firefox 3.0? That way, we could
determine if it's related to the Firefox version or not.

All 1.4 versions work on FF 3.0 except the ones on addons.mozilla.org. 
http://getfirebug.com/releases.

Okay, so with Windows XP Firefox 3.0 and Firebug 1.4.0b7 I'm unable to reproduce this
issue.

I think I cant add anything new here, just to confirm another test on 
http://www.kylehuey.com/moz/testRequest.php:
ff 3.5r and firebug 1.4.07b - same test result
onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]
onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]
onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]
onreadystatechange FAILS Error: Permission denied for <http://www.kylehuey.com> to
create wrapper for object of class UnnamedClass Error: Permission denied for
<http://www.kylehuey.com> to create wrapper for object of class UnnamedClass
[xpconnect wrapped nsIDOMEventListener]

I know there's a lot of conversation right over if this is a Firefox or Firebug
issue, I'd like to suggest another alternative: it's both. From my testing, using
Firefox 3.5 without Firebug works fine, and using Firebug 1.4.0b7 with Firefox 3.0
works fine, which suggests that neither one on its own is actually the problem.
Perhaps there was a change in the way the XHR objects are handled in Firefox 3.5 so
that the way they're handled in Firebug causes the error? (just thinking out loud,
sorry if this was obvious to everyone)

I can confirm this error occurs most of the time in our application running 1.4.0b7
on FF 3.5. (Mac OSX Leopard)  Prevents any AJAX requests from working and renders
FireBug useless for us.

It doesn't seem to be a multiple request issue, I have been able to reproduce it with
a single request.  It has to do with how long the request takes.  For example, if its
around 100-200ms for the request i never receive those errors, but if the request
takes around 1000+ms I receive it every time.  Also, If you want a QUICK fix for this
you can disable "Show XMLHttpRequests" from the options.

This is quite an annoying bug and it's also taking it's toll on several of our apps.
It typically happens when using comet (long-polling) requests via prototype and I can
confirm it's NOT a multiple-request issue.

http://www.jumpino.com/welcome - live link
On our dev machines running older versions of FF and FB, everything works fine, but
on mine (FF 3.5, FB 1.4X0b7), onreadystatechange FAILS errors are rampant. From what
I've read, this COULD be a Firefox bug on how it propagates events.

Perhaps another clue here -
http://www.mooforum.net/general12/firefox-5rc3-problems-with-request-solution-t1898.html
- stating that setting your onreadystatechange handler AFTER calling send() might fix
things.  But that seems to introduce a bit of a race condition.

To others looking for a workaround, it seems that the onload event handler is always
called appropriately regardless of what happens with onreadystatechange. I've used
this as a temporary fix while this issue is being resolved.

Also can confirm it's not a multi-request issue. In addition, I've also seen this
happen periodically on normal-length response times (< 1s). I don't think this is
unique to long responses, but it definitely pops up more frequently as the response
time increases.

I found turning off the net tab helps reduce the chances of getting the error.  Also
it seems that the error almost always happens when the combine time of HXRs exceeds
2-3 seconds on body onload.  

Our site is request heavy(3-6 onload) and I've reproduced the bug on Leopard, Ubuntu,
and XP.

As far as I am able to determine, all of the relevant Firebug code is:
    try
    {
        spy.context.onReadySpy = spy; // maybe the handler will eval(), we want the URL.
        if (spy.onreadystatechange)
            spy.onreadystatechange.handleEvent(event);
    }
...
That is, the exception here arises in the handleEvent() call. Firebug does not touch
the event. The code in question runs correctly "most" of the time. Based on this I
think it very unlikely that changes to Firebug code alone will solve this problem,
and certainly I am out of ideas. 

The workaround here is to ensure that the 'spy' code is not called, by setting
Firebug > Console > Show XMLHttpRequests OFF.

Showing XMLHttpRequests helps tremendously with debugging AJAX though :P -- it is,
however, an adequate fix. I just won't update FF on some machines as I'm still pretty
sure it's a Firefox issue, and not a Firebug one.

I'm still not sure if this has hashed out, but now I get no AJAX working with either
fbug 1.04 or 1.05 in FF 3.5... if you go to any website that has ajax comments (say,
gawker.com, i know, i know ) and click on a link, no comments load anymore. what can
this possibly be?

 Issue 1971  has been merged into this issue.

This issue cannot be resolved until Firefox 3.5.1 at the earliest.

https://bugzilla.mozilla.org/show_bug.cgi?id=502959

Maybe i found a temporary solution for my scripts.
If I put a random number on the ajax call it doesn't give me errors:

var randomnumber=Math.floor(Math.random()*11)
var myAjax = new Ajax.Updater('photos_list',
'ajax/album_photos.php?cache='+randomnumber});    

I too am getting these errors, Win2k3stdsp2, FF3.5, FB 1.4.0b7.  They happen three at
a time per request.  When I tried the temporary solution in comment 47, it seemed to
make the problem occur less often, but it still does occur, particularly if the
request takes longer than 10 s.

Same with FB 1.5X.0a16.

If you need help and workarounds, check out:
http://www.nczonline.net/blog/2009/07/09/firefox-35firebug-xmlhttprequest-and-readystatechange-bug/

Regarding comment 50, you might also like to point out that this problem can be
avoided by setting
   Firebug > Console > OptionsMiniMenu > ShowXMLHttpRequests OFF.

 Issue 2053  has been merged into this issue.

Yes, turning off ShowXMLHttpRequests does help for the individual developer, but it
doesn't help our users who are hitting our site and experiencing the issue. I don't
consider that to be a scalable workaround as compared to using the progress events.

 Issue 2056  has been merged into this issue.

Upgrading to 3.5.1 does not fix this issue.  although I don't know if that can be fix
with some code changes on fb's end.

Just to summarize the current status:
   This bug report is for all version of Firebug and for Firefox 3.5.
   It applies whenever Firebug > Console > Show XMLHttpRequests ON.
   The original AJAX callback problem reported here is fixed in Firebug 1.4 and 1.5.
   Occasionally Ajax callbacks throw "Permission denied for <site> to
create wrapper for object of class UnnamedClass Error"
   The occasions are 100% if garbage collection is forced, suggesting that the reason
the error tends to happen more with delayed requests is increased probability of GC.
   There is nothing that Firebug can do to fix this problem other than removing the
Show XMLHttpRequests feature.
   The firefox team has the bug report and the right folks over there know about it.
   The workaround is simple and effective: 
      Firebug > Console > Show XMLHttpRequests OFF

I have committed a workaround for this problem (1.5 R3662)

It's based on following:

- don't replace the xhr's onreadystatechange original callback by Firebug's internal
function since executing it later throws the exception.

- Use "load" and "error" events instead to monitor XHR.

This should be included in Firebug 1.5a18 available soon here:
http://getfirebug.com/releases/firebug/1.5X/

Please try it and let me know whether it's better for you.

I have also tried to verify this approach by asking here:
http://groups.google.com/group/mozilla.dev.platform/browse_thread/thread/d037e732e643717d?hl=en

Honza

 Issue 1539  has been merged into this issue.

Thanks for your report. This is one of 5 bugs fixed in Firebug 1.5a18,  see
http://blog.getfirebug.com/2009/07/22/firebug-1-5a18/. Please try it and
let us know.

Thank you all who helped debug and fix this problem.  It was sure killing debugging
on a local server, I suppose because any breakpoint executions would, of course,
delay the return to the AJAX callback.  You have saved me!  Ron

I appear to be having this issue with Firefox 3.5.5 and Firebug 1.5X.0b4. I disabled
Firebug and my callback fired as expected. Re-enabled Firebug and the callback no
longer fires.

Regarding comment 61, please see  issue 2501 

This happens to me too with Firefox 4.0b1 and Firebug 1.5.4. Unchecking "Show XMLHttpRequests" from the console menu solves the problem.

Regarding comment 63, please test with Firebug 1.6a16. Firebug 1.5 is not supported on Firefox 4.

Also please tell us which test case posted here that you used to reproduce the problem.
If you did not use on of the tests here, then please open a new bug report.

I am also seeing this issue with firefox 4 and Firebug 1.6X.0a16
using this test posted earlier  -
http://sendsome.org/kunden/XNG/test.html

THANK YOU  ziggyqubert!
Man, I love you so much. In a brotherly way that is. This issue has been driving me nuts and I spent all night debugging the Jquery api to dig out the issue. Damn firebug!

Firebug can't fix this problem due to a bug in Firefox 4
https://bugzilla.mozilla.org/show_bug.cgi?id=557791
See Firebug's  issue 3233  for more info

The workaround is to switch off "Show XMLHttpRequests" option in the Console panel.

Honza

i just got this issue reported by my users on firefox 3.5.3 WITHOUT firebug installed. Just told them to update, thanks

Just changed standard label for platform issues.

I am still receiving this error.  I am attempting to use AJAX to populate a drop down menu as people type.  When the user types slowly, there is no problem.  However, if they type a quick pace, the XMLHttpRequests will start showing an 'abort' next to some of them.  At that point the page will stop firing the onstatechange of the XMLHttpRequest objects.  I am using Firefox 8.0.1 with firebug 1.8.4.

Regards,

Jon

Jon, create a new issue report for the problem and please attach a simple test case that can be used to reproduce it on our machines.

Thanks!
Honza

(No comment was entered for this change.)