Closed Bug 414743 Opened 17 years ago Closed 17 years ago

Session Restore only restores 'untitled' tabs -

Categories

(Firefox :: Session Restore, defect, P1)

x86
Windows Vista
defect

Tracking

()

VERIFIED FIXED
Firefox 3 beta3

People

(Reporter: jmjjeffery, Assigned: mrbkap)

References

Details

(Keywords: regression)

Attachments

(1 file)

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9b3pre) Gecko/2008012917 Minefield/3.0b3pre Firefox/3.0
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9b3pre) Gecko/2008012917 Minefield/3.0b3pre Firefox/3.0

Session Restore is not saving any info, file size stays at 1kb.  No errors in the console. 



Reproducible: Always

Steps to Reproduce:
1. Open a few tabs
2. Close/reopen Minefield- all tabs are 'untitled' 
3.



Latest hourly , Vista HP.

No idea when this regressed, but the nightly this morning was OK, but will all the checkins someone will have to find a window.
Flags: blocking-firefox3?
Keywords: regression
Version: unspecified → Trunk
I'm seeing
nsSessionStore could not be initialized: [Exception... "Component returned failure code: 0x80040111 (NS_ERROR_NOT_AVAILABLE) [nsIDOMJSWindow.setTimeout]"  nsresult: "0x80040111 (NS_ERROR_NOT_AVAILABLE)"  location: "JS frame :: file:///home/mrbkap/src/3mozbuild/firefox/dist/bin/components/nsSessionStore.js :: sss_restoreWindowFeatures :: line 1713"  data: no]

on startup. It looks like somebody's managing to call setTimeout on a window that doesn't have an inner window yet.
Status: UNCONFIRMED → NEW
Ever confirmed: true
Regresses for me between... and...
http://hourly-archive.localgho.st/linux/20080129_1727_firefox-3.0b3pre.en-US.linux-i686.tar.bz2
http://hourly-archive.localgho.st/linux/20080129_1738_firefox-3.0b3pre.en-US.linux-i686.tar.bz2

https://bugzilla.mozilla.org/process_bug.cgi has the same regression range for me.

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9b3pre) Gecko/2008012917 Minefield/3.0b3pre
This  was caused by bug 387706. In particular, CallerInnerWindow doesn't deal with the caller being a JS component or some other type of caller whose scope doesn't end with a window.
Blocks: postMessage
This is the smallest patch I could make. With it, if the caller scope chain doesn't end in a window, we'll just use our current inner window. I don't think there are security implications for this as if the caller is not from a window, we should probably trust it (at least, I think we do!).
Assignee: nobody → mrbkap
Status: NEW → ASSIGNED
Attachment #300253 - Flags: superreview?(jst)
Attachment #300253 - Flags: review?(jwalden+bmo)
Attachment #300253 - Flags: review?(jst)
Comment on attachment 300253 [details] [diff] [review]
Proposed minimal patch

It's really a bit whacky this code didn't crash before in this case.
Attachment #300253 - Flags: review?(jwalden+bmo) → review+
(In reply to comment #2)
> https://bugzilla.mozilla.org/process_bug.cgi has the same regression range for
> Should read https://bugzilla.mozilla.org/show_bug.cgi?id=414758

Flags: blocking-firefox3? → blocking-firefox3+
Priority: -- → P1
Target Milestone: --- → Firefox 3 beta3
Attachment #300253 - Flags: superreview?(jst)
Attachment #300253 - Flags: superreview+
Attachment #300253 - Flags: review?(jst)
Attachment #300253 - Flags: review+
Checked in.
Status: ASSIGNED → RESOLVED
Closed: 17 years ago
Resolution: --- → FIXED
*** VERIFIED

Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9b3) Gecko/2008020514 Firefox/3.0b3
Status: RESOLVED → VERIFIED
This was originally filed as bug 413259, but I didn't get a chance to look at that bug before bug 387706 exposed it in a more noticeable way here.
Depends on: 418378
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: