342179 - close a tab -> Error: tabbrowser.mTabs[i] has no properties

Status: RESOLVED FIXED

(Firefox :: Tabbed Browser, defect)

Description

[Peter van der Woude [:Peter6]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20060620 Minefield/3.0a1 ID:2006062002 [cairo]

repro:

1.open FF
2.open an extra tab
3.close that tab

Error: tabbrowser.mTabs[i] has no properties
Source file: file:///C:/Programmas/Firefox/ff30C/firefox/components/nsSessionStore.js
Line: 816


followed by
Error: Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIDOMXULElement.boxObject]
Source file: chrome://global/content/bindings/browser.xml

Comment 1

[Simon Bünzli]

This sounds to me like the regression of bug 338441 which should get fixed with bug 318168.

Comment 2

[Frederic Bezies]

Facing in under ubuntu 6.06 LTS.

OS -> All ?

Comment 3

[Peter van der Woude [:Peter6]]

(In reply to comment #1)
> This sounds to me like the regression of bug 338441 which should get fixed with
> bug 318168.
> 

it first showed up last night after bug 254021 was checked in though

Comment 4

[Simon Bünzli]

Attachment: fix (see bug 318168)

This fixes the issue for me. I don't see how bug 254021 could have caused this without SessionStore being enabled, so it's probably rather due to bug 337320 (as mentioned in bug 342155).

Comment 5

[aja+bugzilla]

(In reply to comment #4)
> Created an attachment (id=226366) [edit]
> fix (see bug 318168)
> 
> This fixes the issue for me. I don't see how bug 254021 could have caused this
> without SessionStore being enabled, so it's probably rather due to bug 337320
> (as mentioned in bug 342155).
> 

SessionStore has been enabled on trunk for several days, and worked fine until 254021 was checked in.

Comment 6

[Simon Bünzli]

Indeed. The problem is thus probably rather that SessionStore tries to capture the tab one last time while it's already being destroyed (however I don't see how that issue could be fixed before the new tab related events are added in bug 322898 or bug 318168).

Comment 7

[(not reading, please use [e-mail] instead)]

I think this is related to something I've seen with bug #318168, in that we need to clear the _browsers cache.  see bug #318168 comment #36.

Comment 8

[Simon Bünzli]

(In reply to comment #7)

I thought so, too, originally, but it in fact rather seems to be related to SessionStore trying to access a tab which has already been removed from mTabContainer. Apart of the _browsers fix, SessionStore must be able to observe the TabClose event and react to it (instead of DOMNodeRemoved) - see my comment #6.

Comment 9

[(not reading, please use [e-mail] instead)]

based on simon's comments, re-assigning to dietrich.

Comment 10

[Dietrich Ayala (:dietrich)]

Added dependency on bug 334839, which is for adding support for the new tab events to sessionstore, once they're available.

Comment 11

[Dietrich Ayala (:dietrich)]

Fixed by bug 318168 and bug 334839.