Franklin Bynum Reporter
	Description • 17 years ago

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9a7pre) Gecko/2007072804 Minefield/3.0a7pre
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9a7pre) Gecko/2007072804 Minefield/3.0a7pre

After logging into Yahoo Mail beta, all seems to load properly but when the messenger feature at the top begins to stir, a dialog box pops onto the screen, and is followed by a crash.

Reproducible: Always

Steps to Reproduce:
1. Make sure Yahoo Messenger feature is turned on in Yahoo Mail beta
2. Log into Yahoo Mail beta at http://mail.yahoo.com
3. Dialog box will appear, click ok (only option)
5. Crash
Actual Results:  
Crash

Expected Results:  
Page load

http://crash-stats.mozilla.com/report/index/a9b08028-3d7b-11dc-828b-001a4bd46e84?date=2007-07-29-02

Tested on Gecko/2007072804 Minefield/3.0a7pre on Windows XP as well.  Same result.

	u88484
	Comment 2 • 17 years ago

This started with a 20070727 build

	Frank DiPalermo
	Comment 3 • 17 years ago

I can confirm that here as well.

	u88484
	Comment 4 • 17 years ago

And of course crash reporter doesn't work like normal so I can't provide and data.  Should note that we can not even read our email or click on the switch back option to go back to old version of yahoo so we are pretty much SOL for reading emails.

	Bob Clary [:bc] (inactive)
	Comment 5 • 17 years ago

probable dupe of bug Bug 389753

	-fullmetaljacket-
	Comment 6 • 17 years ago

i am getting the same error pop up but no crashes

	-fullmetaljacket-
	Comment 7 • 17 years ago

OS > ALL

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a7pre) Gecko/2007072922 Minefield/3.0a7pre ID:2007072922

	u88484
	Comment 8 • 17 years ago

Yeah I' am no longer getting any crashes but still get the dialog. Hmmmm.

	Peter van der Woude [:Peter6]
	Comment 9 • 17 years ago

see bug 389753#c14

	Blake Kaplan (:mrbkap) (inactive) Assignee
	Comment 10 • 17 years ago

The crash should be gone (reopen bug 389753 if it isn't), this bug is about the alert boxes that pop up -- I believe that the problem is that enumeration over cross origin wrappers no longer works.

	Blake Kaplan (:mrbkap) (inactive) Assignee
	Comment 11 • 17 years ago

This should fix things, the patch is actually just adding a:
do {
...
} while ((innerObj = JS_GetPrototype(cx, innerObj));

with some changes to make it compile. The problem was that we weren't enumerating up our prototype -- which confused XPConnect into not resolving any properties.

	Blake Kaplan (:mrbkap) (inactive) Assignee
	Comment 12 • 17 years ago

Oops, that last patch had some debugging code in it.

	Damon Sicore (:damons)
	Comment 13 • 17 years ago

Setting the milestone to M7 so people know why the tree is still closed.

	Blake Kaplan (:mrbkap) (inactive) Assignee
	Comment 14 • 17 years ago

Fix checked into trunk.

	Blake Kaplan (:mrbkap) (inactive) Assignee
	Comment 15 • 17 years ago

This patch creates a mochitest for all testable XOW bugs. I'm not sure what the right tone is for the strings in the is() and ok() tests, but these seem to be OK.

	Boris Zbarsky [:bzbarsky]
	Comment 16 • 17 years ago

Comment on attachment 275192 [details] [diff] [review]
Mochitest

So is window.location always wrapped in XOW, even when accessed from inside the window itself?

>+    try {
>+        is('PASS', window.eval("'PASS'"), 'window.eval throws an exception');
>+    } catch (e) {
>+        ok(false, 'window.eval does not throw an exception');

Those messages seem to be backwards...

	Boris Zbarsky [:bzbarsky]
	Comment 17 • 17 years ago

Comment on attachment 275192 [details] [diff] [review]
Mochitest

sr=bzbarsky.  Per Blake on irc, we do wrap window.location, since it lives on the outer window and hence its object principal might change.

	Blake Kaplan (:mrbkap) (inactive) Assignee
	Comment 18 • 17 years ago

I checked the mochitest in.