732593 - mTabProgressListener onStateChange error if iframe onload moves the iframe

Status: NEW

(Firefox :: Tabbed Browser, defect)

Description

[Jesse Ruderman]

Attachment: testcase

Loading the testcase triggers:

JavaScript error: , line 0: uncaught exception: [Exception... "Component returned failure code: 0x80004002 (NS_NOINTERFACE) [nsIWebProgress.DOMWindow]"  nsresult: "0x80004002 (NS_NOINTERFACE)"  location: "JS frame :: chrome://browser/content/tabbrowser.xml :: <TOP_LEVEL> :: line 545"  data: no]

Seems to be this line:
http://hg.mozilla.org/mozilla-central/annotate/343ec916dfd5/browser/base/content/tabbrowser.xml#l549

Comment 1

[Dão Gottwald [:dao]]

http://mxr.mozilla.org/comm-central/search?string=.DOMWindow&case=on&filter=DOMWindow\b

Hardly anywhere do we expect DOMWindow to throw. Can it return null instead? Alternatively, can nsIWebProgress get a flag for whether it relates to a top-level window?

Comment 2

[Huseyin]

I have encountered the same problem today. I was directly doing something in a function that called from onload() of an iframe.

In an iframe:
window.onload=function() { parent.aFunctionOnParent(); }
In parent:
function aFunctionOnParent() { codes... }

Then I have used setTimeOut and then worked without exception. These changes fixed my problem:
function aFunctionOnParent() { setTimeOut(){ function() { codes... }, 1} }

Comment 3

[Jesse Ruderman]

Still happens on trunk. 

http://hg.mozilla.org/mozilla-central/rev/a8a547fb44a8 changed the error message to:

JavaScript error: chrome://browser/content/tabbrowser.xml, line 562: NS_NOINTERFACE: Component returned failure code: 0x80004002 (NS_NOINTERFACE) [nsIWebProgress.isTopLevel]

Comment 4

[:Gavin Sharp [email: [e-mail]]]

(In reply to Dão Gottwald [:dao] from comment #1)
> Hardly anywhere do we expect DOMWindow to throw. Can it return null instead?

I guess this would be relatively simple to implement. But I wonder whether it would instead be better to suppress firing web progress notifications in the cases where it would return null.

Comment 5

[:Gavin Sharp [email: [e-mail]]]

Bug 866748 filed for the case that is affecting tabbrowser. I guess we can maybe morph this into the more general case of "what should the nsIWebProgressListener behavior be" in this situation?

Comment 6

[Firefox Bug Husbandry Bot]

https://wiki.mozilla.org/Bug_Triage/Projects/Bug_Handling/Bug_Husbandry#Move_fix-optionals