81769 - browser urlbar is now updated with the URL of subframes of document

Status: VERIFIED DUPLICATE of bug 82236

(SeaMonkey :: Location Bar, defect)

Description

[John Morrison]

As of the evening builds for tonight (Linux 2001051820 and win2k 2001051820), 
which include jst's fix for bug 60173

    "Checking in support for onload handlers for frame and iframe elements, 
     and fixing .ownerDocument of orphan elements, fixes bug 60173, and also 
     fixes part of bug 27382."

With these builds, when you visit a frame set page, or a page with inline 
iframes (read "ad banners"), the urlbar will be updated with the URL of one
of the child frames. 

I'm assuming this is a consequence of jst's fix (which I'm very glad to see!),
and that the browser now receives multiple onload events (as the onload events
of each child frame now bubble up (jst correct me if that's wrong).

Steps to reproduce:
1) go to http://www.aol.com/ or http://www.wired.com/ (type in urlbar and hit  
 enter
2) watch the urlbar as the page loads; the location will be changed more than 
   once during the page load.
3) when the page has loaded, set focus in the urlbar and hit enter (e.g. the 
   page that will be loaded is not the url of the top-level document;  it is 
   the url of one of the subframes.

Comment 1

[John Morrison]

I hereby decorate with magic words (needs a fix for beta).

Comment 2

[John Morrison]

> I'm assuming this is a consequence of jst's fix (which I'm very glad to see!

Er, 'glad to see' that frame load events bubble, which can be useful. Too
bad about the side effect.

Comment 3

[R.K.Aa.]

This messes up history and makes it hard to set proper bookmarks at affected
sites. Some might not even realize what's happening and wind up bookmarking some
banner frame in good faith. Is severity "major" high enough?

Comment 4

[R.K.Aa.]

bug 50427 was once critical/P1

Comment 5

[Johnny Stenback (:jst)]

Is this caused by bubbling onload events in XUL? If so we can easily (and
should) fix that in the XUL code, lemme know, I whip up a fix for that.

Comment 6

[R.K.Aa.]

*** Bug 81806 has been marked as a duplicate of this bug. ***

Comment 7

[John Morrison]

(adding blizzard's notation from duplicate bug 81806).

jst: I haven't looked through this in detail. I'm not sure what you meant by
"caused by bubbling onload events in XUL", but I don't think it is a bad thing
to have these events bubble up from _content into the xul window. It's just 
that the app logic needs to distinguish between the onload of the 'top-level'
document, and the onload of subframes in its event listeners (or that's what 
I think it is).

Comment 8

[Johnny Stenback (:jst)]

This is unrelated to my frame/iframe onload event changes (which turned out to
not even work at the moment because of some nsIWebProgress fu), I backed out my
changes and I'm still seeing this problem.

Comment 9

[John Morrison]

Oops, sorry for jumping to that conclusion. And, If I had looked a bit closer, 
I would have seen that the frame loads were already being caught in the browser
js anyways. But, I don't know what changed (but it was between 12:00 noon 
and about 10pm Friday, based on the verification builds that came out.

Comment 10

[R.K.Aa.]

*** Bug 81858 has been marked as a duplicate of this bug. ***

Comment 11

[Gilles Durys]

*** Bug 81921 has been marked as a duplicate of this bug. ***

Comment 12

[timeless]

*** Bug 81992 has been marked as a duplicate of this bug. ***

Comment 13

[Peter of the Norse]

*** Bug 82070 has been marked as a duplicate of this bug. ***

Comment 14

[Gilles Durys]

*** Bug 82115 has been marked as a duplicate of this bug. ***

Comment 15

[Skewer]

Attachment: Frame testcase from bug 56062

Comment 16

[Skewer]

For window frames, bookmarking appears to work properly, despite the location
bar problem. The top (z) frame is bookmarked, not one of the children.

Comment 17

[Radha on family leave (not reading bugmail)]

Fix to bug 58906 caused this regression. taking from blake.

Comment 18

[John Morrison]

Er, the fix to bug 56062 caused this (that's the one you meant, right?)

Comment 19

[Gilles Durys]

*** Bug 82546 has been marked as a duplicate of this bug. ***

Comment 20

[djoham]

Adding self to cc list. I think this is the bug that is goofing up the reloading
of my frameset site. Everything gets horked on a reload. Wasn't a problem in .9

David

Comment 21

[doctor__j]

*** Bug 82675 has been marked as a duplicate of this bug. ***

Comment 22

[Radha on family leave (not reading bugmail)]

I'm marking this a dupe of 82236, even though this one came thro' earlier. 82236  
describes  another side effect that I want to monitor. 

*** This bug has been marked as a duplicate of 82236 ***

Comment 23

[doctor__j]

*** Bug 82784 has been marked as a duplicate of this bug. ***

Comment 24

[Tom Mraz]

Marking verified per comments.