Closed
Bug 440002
Opened 16 years ago
Closed 15 years ago
crash on exit if popups have been blocked
Categories
(Camino Graveyard :: Annoyance Blocking, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: jamadden, Assigned: stuart.morgan+bugzilla)
References
()
Details
(Keywords: crash, fixed1.8.1.22)
Attachments
(2 files, 1 obsolete file)
27.85 KB,
text/plain
|
Details | |
3.42 KB,
patch
|
mikepinkerton
:
superreview+
|
Details | Diff | Splinter Review |
User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.8) Gecko/20051111 Firefox/2.0.0.1 Build Identifier: Version 2.0a1pre (1.9pre 2008061223) Camino crashes on normal exit if you have visited a page that resulted in a blocked popup. This is true whether or not the page that triggered the popup is still displayed and whether or not the popup warning has been dismissed. Noticed in both 1.6.1 and nightly builds. The stack trace varies from crash to crash, and so I wasn't sure how to reproduce this. Then I saw a stack that mentioned a nsDOMPopupBlockedEvent, and that turned out to be a clue. Reproducible: Always
Reporter | ||
Comment 1•16 years ago
|
||
Reporter | ||
Comment 2•16 years ago
|
||
Reporter | ||
Updated•16 years ago
|
Attachment #325571 -
Attachment mime type: application/octet-stream → text/plain
Assignee | ||
Comment 3•16 years ago
|
||
Comment on attachment 325571 [details]
Crash report for same problem, nightly build, random stack
This stack isn't the same problem; it's a different trunk-only bug.
Attachment #325571 -
Attachment is obsolete: true
The first report is bug 349463; the second is bug 433644. If you can reproduce these types of crashes on 1.6.1 (and they aren't the PL_DHash... variety; that exists on 1.6.x and trunk), please do file another bug. Bug 433644 has a rather old Camino build with the patch to fix that crash linked in one of the comments; duping there, since that's more serious.
Severity: normal → critical
Status: UNCONFIRMED → RESOLVED
Closed: 16 years ago
Keywords: crash
Resolution: --- → DUPLICATE
Assignee | ||
Comment 5•16 years ago
|
||
While this does look like a dup of bug 349463, if this is a common source of the crash we could probably improve the situation by setting up an XPCOM shutdown observer in BW, and clearing out the popup blocking array at that point; reopening to consider that. However, I'm confused about this statement: > This is true whether or not the page that triggered the popup is > still displayed We clear the array on navigation, so I really doubt that this stack could actually happen when no open page had blocked popups. Could you give minimal steps that show that that case is really always reproducible, and attach the corresponding stack?
Status: RESOLVED → UNCONFIRMED
Resolution: DUPLICATE → ---
Reporter | ||
Comment 6•16 years ago
|
||
(In reply to comment #5) > We clear the array on navigation, so I really doubt that this stack could > actually happen when no open page had blocked popups. Could you give minimal > steps that show that that case is really always reproducible, and attach the > corresponding stack? Indeed, you are correct. I tried to reproduce this in 1.6.1 on a clean account and could not, after having navigated away from the page. Looking back at my crash reports for 1.6.1, I think I chalked most of the PL_DHashTableOperate crashes from bug 349463 to this. Sorry for the confusion.
Reporter | ||
Comment 7•16 years ago
|
||
I found out why this appeared to be always reproducible in my account but not in a different one. I had somehow ended up with a Camino window that was one pixel wide by about 50 pixels tall, and flush against the right side of the screen. This window was labeled "..." in the Window menu, and ultimately seemed to contain more than one tab. I hypothesize that this mini-window was loading a URL that tried to display a popup. After getting this window closed (not easy!) I haven't had a mysterious crash-on-exit.
Comment 8•16 years ago
|
||
Seeing this with v1.6.1 on OS X v10.5.3. CrashReporter data below. Process: Camino [4746] Path: /Applications/Camino/Camino.app/Contents/MacOS/Camino Identifier: org.mozilla.camino Version: 1.6.1 (1608.05.12) Code Type: X86 (Native) Parent Process: launchd [71] Date/Time: 2008-07-19 00:23:50.185 -0400 OS Version: Mac OS X 10.5.3 (9D34) Report Version: 6 Exception Type: EXC_BAD_ACCESS (SIGBUS) Exception Codes: KERN_PROTECTION_FAILURE at 0x000000000000000c Crashed Thread: 0 Thread 0 Crashed: 0 libxpcom_core.dylib 0x00c93ca8 PL_DHashTableOperate + 24 1 libxpcom_core.dylib 0x00c99498 AtomImpl::~AtomImpl() + 48 2 libxpcom_core.dylib 0x00c998cd AtomImpl::Release() + 47 3 libxpcom_core.dylib 0x00c94008 nsCOMPtr_base::~nsCOMPtr_base() + 20 4 org.mozilla.camino 0x00385269 nsCSSSelector::Reset() + 73 5 org.mozilla.camino 0x00385306 nsCSSSelector::~nsCSSSelector() + 20 6 org.mozilla.camino 0x0038535f nsCSSSelectorList::~nsCSSSelectorList() + 25 7 org.mozilla.camino 0x003853cb CSSStyleRuleImpl::~CSSStyleRuleImpl() + 39 8 org.mozilla.camino 0x00599f27 nsCSSRule::Release() + 41 9 libxpcom_core.dylib 0x00ca28c6 nsSupportsArray::Clear() + 44 10 libxpcom_core.dylib 0x00ca2133 nsSupportsArray::DeleteArray() + 19 11 libxpcom_core.dylib 0x00ca2297 nsSupportsArray::Release() + 47 12 org.mozilla.camino 0x0038a81f nsCSSStyleSheetInner::~nsCSSStyleSheetInner() + 59 13 org.mozilla.camino 0x003887cd nsCSSStyleSheet::~nsCSSStyleSheet() + 215 14 org.mozilla.camino 0x00387da4 nsCSSStyleSheet::Release() + 44 15 libxpcom_core.dylib 0x00c94008 nsCOMPtr_base::~nsCOMPtr_base() + 20 16 org.mozilla.camino 0x0081d76e DeviceContextImpl::GetAltDevice(nsIDeviceContext**) + 50582 17 libxpcom_core.dylib 0x00c93827 PL_DHashTableFinish + 73 18 org.mozilla.camino 0x0036e06e CSSLoaderImpl::~CSSLoaderImpl() + 84 19 org.mozilla.camino 0x0036de45 CSSLoaderImpl::Release() + 41 20 org.mozilla.camino 0x0042a31c nsDocument::~nsDocument() + 898 21 org.mozilla.camino 0x0048912a nsHTMLDocument::~nsHTMLDocument() + 398 22 org.mozilla.camino 0x0042023f nsDocument::Release() + 53 23 libxpcom_core.dylib 0x00c94008 nsCOMPtr_base::~nsCOMPtr_base() + 20 24 org.mozilla.camino 0x0046ba17 nsDOMEvent::~nsDOMEvent() + 109 25 org.mozilla.camino 0x0046fb8c nsDOMPopupBlockedEvent::~nsDOMPopupBlockedEvent() + 70 26 org.mozilla.camino 0x0046af01 nsDOMEvent::Release() + 41 27 libxpcom_core.dylib 0x00cad761 ReleaseObjects(void*, void*) + 21 28 libxpcom_core.dylib 0x00ca96b6 nsVoidArray::EnumerateForwards(int (*)(void*, void*), void*) + 56 29 libxpcom_core.dylib 0x00cad792 nsCOMArray_base::Clear() + 42 30 libxpcom_core.dylib 0x00cadbc8 nsArray::Clear() + 20 31 libxpcom_core.dylib 0x00cadbec nsArray::~nsArray() + 32 32 libxpcom_core.dylib 0x00cadcfc nsArray::Release() + 42 33 org.mozilla.camino 0x00003a6a -[BrowserWrapper dealloc] + 238 34 com.apple.AppKit 0x94789b83 -[NSTabViewItem dealloc] + 92 35 org.mozilla.camino 0x0004c595 -[BrowserTabViewItem dealloc] + 121 36 com.apple.CoreFoundation 0x91c425da CFRelease + 90 37 com.apple.CoreFoundation 0x91bd19e3 __CFArrayReleaseValues + 435 38 com.apple.CoreFoundation 0x91c42788 _CFRelease + 216 39 com.apple.Foundation 0x96af37bf NSPopAutoreleasePool + 1007 40 com.apple.Foundation 0x96afdc42 +[NSAutoreleasePool releaseAllPools] + 50 41 com.apple.AppKit 0x947e4183 -[NSApplication _deallocHardCore:] + 505 42 com.apple.AppKit 0x947e2e59 -[NSApplication terminate:] + 742 43 com.apple.AppKit 0x94703c23 -[NSApplication sendAction:to:from:] + 112 44 com.apple.AppKit 0x947b2660 -[NSMenu performActionForItemAtIndex:] + 493 45 com.apple.AppKit 0x947b2365 -[NSCarbonMenuImpl performActionWithHighlightingForItemAtIndex:] + 220 46 com.apple.AppKit 0x947b1feb -[NSMenu performKeyEquivalent:] + 866 47 com.apple.AppKit 0x947b0890 -[NSApplication _handleKeyEquivalent:] + 492 48 com.apple.AppKit 0x946cd853 -[NSApplication sendEvent:] + 3999 49 com.apple.AppKit 0x9462ae27 -[NSApplication run] + 847 50 com.apple.AppKit 0x945f8030 NSApplicationMain + 574 51 org.mozilla.camino 0x000034f4 main + 196 52 org.mozilla.camino 0x00002dc6 _start + 216 53 org.mozilla.camino 0x00002ced start + 41 Unlike the OP, I can't find a window that is triggering this. Confirming bug on the grounds that I am not the OP, and nsDOMPopupBlockedEvent is in my crash log, and I can reproduce this 100% of the time.
Status: UNCONFIRMED → NEW
Ever confirmed: true
Assignee | ||
Comment 9•16 years ago
|
||
Comment 8 looks more like bug 349463.
Assignee | ||
Comment 10•16 years ago
|
||
Taking so that I remember to implement comment 5.
Assignee: nobody → stuart.morgan+bugzilla
There's a new Talkback incident for the stack in comment 0 in today's 2.0 Talkback.
Since upgrading to 2.0b2, I've had this (the comment 0 variant of our old friend PL_DHashTableOperate) crash on quit every quit. I noticed on relaunch yesterday that one of the sites I had open in a minimized tab had the popup blocker bar visible, and it could well have been visible when I went to quit (this sort-of aligns with what Jason mentions in comment 7).
Assignee | ||
Comment 13•15 years ago
|
||
Comment 5 goodness.
Attachment #365136 -
Flags: superreview?(mikepinkerton)
Assignee | ||
Comment 14•15 years ago
|
||
This is really low-risk; shall we add it to 1.6.7?
Flags: camino1.6.7?
(In reply to comment #14) > This is really low-risk; shall we add it to 1.6.7? Yes, please.
Flags: camino1.6.7? → camino1.6.7+
Comment 16•15 years ago
|
||
Comment on attachment 365136 [details] [diff] [review] fix sr=pink
Attachment #365136 -
Flags: superreview?(mikepinkerton) → superreview+
Checked in for Stuart on cvs trunk and MOZILLA_1_8_BRANCH ahead of 1.6.7.
Status: NEW → RESOLVED
Closed: 16 years ago → 15 years ago
Keywords: fixed1.8.1.21
Resolution: --- → FIXED
Keywords: fixed1.8.1.21 → fixed1.8.1.22
You need to log in
before you can comment on or make changes to this bug.
Description
•