Closed Bug 336785 Opened 19 years ago Closed 19 years ago

Sunbird crashes when using context menu key after inline editing [@ nsEventListenerManager::PrepareToUseCaretPosition]

Categories

(Core :: DOM: UI Events & Focus Handling, defect)

x86
All
defect
Not set
normal

Tracking

()

RESOLVED FIXED

People

(Reporter: ssitter, Assigned: brettw)

References

()

Details

(Keywords: crash, fixed1.8.1)

Crash Data

Attachments

(1 file)

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20060505 Mozilla Sunbird/0.3a2 - Build ID: 2006050509 Sunbird crashes when using context menu key after inline editing [@ nsEventListenerManager::PrepareToUseCaretPosition] Steps to Reproduce: 1. Start Mozilla Sunbird with clean profile 2. Create a test event 3. Click on event to select it 4. Click a second time on event to start inline editing of event title 5. Change title and press Return keyboard key to end inline editing 6. Press the windows context menu keyboard key Actual Results: Sunbird crashes immediately Expected Results: Views context menu is displayed Additional Information: No crash when performing the same steps with Lightning 0.1+ (2006050506) and Mozilla/5.0 (Windows; U; Windows NT 5.0; de; rv:1.8.0.2) Gecko/20060308 Thunderbird/1.5.0.2. Last message before Sunbird crash in console is: ###!!! ASSERTION: You can't dereference a NULL nsCOMPtr with operator->().: 'mRawPtr != 0', file y:\tmp\sb-trunk-dbg\dist\include\xpcom\nsCOMPtr.h, line 849 ----------------------------------------------------------- Stack Backtrace: gklayout!nsEventListenerManager::PrepareToUseCaretPosition+0x1bd [y:\dev\sb-trunk\mozilla\content\events\src\nseventlistenermanager.cpp @ 2055] gklayout!nsEventListenerManager::FixContextMenuEvent+0x109 [y:\dev\sb-trunk\mozilla\content\events\src\nseventlistenermanager.cpp @ 1975] gklayout!nsEventListenerManager::HandleEvent+0xb7 [y:\dev\sb-trunk\mozilla\content\events\src\nseventlistenermanager.cpp @ 1673] gklayout!nsEventTargetChainItem::HandleEvent+0x8a [y:\dev\sb-trunk\mozilla\content\events\src\nseventdispatcher.cpp @ 335] gklayout!nsEventTargetChainItem::HandleEventTargetChain+0x99 [y:\dev\sb-trunk\mozilla\content\events\src\nseventdispatcher.cpp @ 429] gklayout!nsEventTargetChainItem::CreateChainAndHandleEvent+0x114 [y:\dev\sb-trunk\mozilla\content\events\src\nseventdispatcher.cpp @ 401] gklayout!nsEventTargetChainItem::CreateChainAndHandleEvent+0xde [y:\dev\sb-trunk\mozilla\content\events\src\nseventdispatcher.cpp @ 392] gklayout!nsEventTargetChainItem::CreateChainAndHandleEvent+0xde [y:\dev\sb-trunk\mozilla\content\events\src\nseventdispatcher.cpp @ 392] gklayout!nsEventTargetChainItem::CreateChainAndHandleEvent+0xde [y:\dev\sb-trunk\mozilla\content\events\src\nseventdispatcher.cpp @ 392] gklayout!nsEventDispatcher::Dispatch+0x254 [y:\dev\sb-trunk\mozilla\content\events\src\nseventdispatcher.cpp @ 575] gklayout!PresShell::HandleEventInternal+0x1de [y:\dev\sb-trunk\mozilla\layout\base\nspresshell.cpp @ 6148] gklayout!PresShell::HandleEvent+0x65b [y:\dev\sb-trunk\mozilla\layout\base\nspresshell.cpp @ 5919] gklayout!nsViewManager::HandleEvent+0x59 [y:\dev\sb-trunk\mozilla\view\src\nsviewmanager.cpp @ 1712] gklayout!nsViewManager::DispatchEvent+0xc7d [y:\dev\sb-trunk\mozilla\view\src\nsviewmanager.cpp @ 1665] gklayout!HandleEvent+0x46 [y:\dev\sb-trunk\mozilla\view\src\nsview.cpp @ 174] gkwidget!nsWindow::DispatchEvent+0xc1 [y:\dev\sb-trunk\mozilla\widget\src\windows\nswindow.cpp @ 1103] gkwidget!nsWindow::DispatchWindowEvent+0x26 [y:\dev\sb-trunk\mozilla\widget\src\windows\nswindow.cpp @ 1124] gkwidget!nsWindow::DispatchMouseEvent+0x6b3 [y:\dev\sb-trunk\mozilla\widget\src\windows\nswindow.cpp @ 5989] gkwidget!ChildWindow::DispatchMouseEvent+0x93 [y:\dev\sb-trunk\mozilla\widget\src\windows\nswindow.cpp @ 6172] gkwidget!nsWindow::ProcessMessage+0xc08 [y:\dev\sb-trunk\mozilla\widget\src\windows\nswindow.cpp @ 4538]
Brett, this crash happens in your code. Any idea what might have caused it?
Assignee: nobody → brettw
Retest with different Thunderbird versions: Lightning + Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8.0.4) Gecko/20060505 Thunderbird/1.5.0.4 - Build ID: 2006050504: No crash Lightning + Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8) Gecko/20060505 Thunderbird/2.0a1 - Build ID: 2006050504: Crash, Talkback Incident ID: TB18343711E Lightning + Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20060505 Thunderbird/3.0a1 - Build ID: 2006050506: Crash, Talkback Incident ID: TB18343864X Crash also happens in Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20051224 Mozilla Sunbird/0.3a1+. Therefore it does not look like a regression from a recent check in.
Attachment #220974 - Flags: first-review?
This is a general problem of not checking for NULL for a few functions. I didn't realize they could return NULL but NS_OK.
Component: General → Event Handling
Flags: first-review?
OS: Windows 2000 → All
Product: Calendar → Core
Attachment #220974 - Flags: review?(darin)
Attachment #220974 - Flags: approval-branch-1.8.1?(darin)
Comment on attachment 220974 [details] [diff] [review] Patch for this and a few other cases r=darin. this requires review from a contenet module peer.
Attachment #220974 - Flags: superreview?(bryner)
Attachment #220974 - Flags: review?(darin)
Attachment #220974 - Flags: review+
Attachment #220974 - Flags: superreview?(bryner)
Attachment #220974 - Flags: superreview+
Attachment #220974 - Flags: approval-branch-1.8.1?(darin)
Attachment #220974 - Flags: approval-branch-1.8.1+
Fixed on trunk and 1.8 branch
Status: NEW → RESOLVED
Closed: 19 years ago
Keywords: fixed1.8.1
Resolution: --- → FIXED
i question the value of the comments you're adding (watch out), that should be perfectly clear from the code you leave in the file and the interface. we don't normally include such comments, if we did, there would be millions of them.
(In reply to comment #7) > i question the value of the comments you're adding (watch out), that should be > perfectly clear from the code you leave in the file and the interface. we don't > normally include such comments, if we did, there would be millions of them. What on earth are you talking about?
Comment on attachment 220974 [details] [diff] [review] Patch for this and a few other cases >Index: nsEventListenerManager.cpp >+ // caret selection, watch out: GetCaretDOMSelection can return null but NS_OK
Crash Signature: [@ nsEventListenerManager::PrepareToUseCaretPosition]
Component: Event Handling → User events and focus handling
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: