Closed Bug 350930 Opened 18 years ago Closed 18 years ago

crashes [@ SearchTable]

Categories

(Core :: DOM: Core & HTML, defect)

1.8 Branch
x86
Windows XP
defect
Not set
critical

Tracking

()

RESOLVED FIXED
mozilla1.8.1

People

(Reporter: dbaron, Assigned: dbaron)

References

Details

(Keywords: fixed1.8.1, topcrash)

Crash Data

Attachments

(2 files)

There have been continued crashes in pldhash code (in SearchTable) following the fix for bug 334177.  Some of the user comments mention shutdown; some don't.  The only theory I have so far is that they're all shutdown crashes, but some of the users got confused when reporting.  (I think I've seen that before for crashes that were clearly shutdown crashes by the stack.)

The talkback incidents all look pretty much like this:

Incident ID: 22723409
Stack Signature	SearchTable 69be2923
Product ID	Firefox2
Build ID	2006082101
Trigger Time	2006-08-31 03:05:45.0
Platform	Win32
Operating System	Windows NT 5.1 build 2600
Module	xpcom_core.dll + (0000147a)
URL visited	
User Comments	
Since Last Crash	477 sec
Total Uptime	477 sec
Trigger Reason	Access violation
Source File, Line No.	c:/builds/tinderbox/Fx-Mozilla1.8-release/WINNT_5.2_Depend/mozilla/xpcom/build/pldhash.c, line 414
Stack Trace 	
SearchTable  [mozilla/xpcom/build/pldhash.c, line 414]
PL_DHashTableOperate  [mozilla/xpcom/build/pldhash.c, line 557]
nsGenericDOMDataNode::~nsGenericDOMDataNode  [mozilla/content/base/src/nsGenericDOMDataNode.cpp, line 72]
nsTextNode::`scalar deleting destructor'

My one theory so far is that I'm forgetting to set the ops pointer to null after destroying the hash table -- a regression from some of my leak fixes, I think.
Attached patch trunk patchSplinter Review
Attachment #236321 - Flags: superreview?(bugmail)
Attachment #236321 - Flags: review?(bugmail)
Attached patch branch patchSplinter Review
This is a really trivial patch -- it nulls out a pointer that we null check before accessing the objects that we're destroying here.  I'm not sure that it's the cause of these crashes, but it could cause crashes (for a text node that's in the range list hash but not the event listener manager hash).
Attachment #236323 - Flags: superreview?(bugmail)
Attachment #236323 - Flags: review?(bugmail)
Attachment #236323 - Flags: approval1.8.1?
Flags: blocking1.8.1?
Hrm.  I think I should have used bug 330683 instead of filing this duplicate, but at this point I'll leave it separate in case it doesn't fix the crashes in talkback.
Blocks: 330683
Attachment #236321 - Flags: superreview?(bugmail)
Attachment #236321 - Flags: superreview+
Attachment #236321 - Flags: review?(bugmail)
Attachment #236321 - Flags: review+
Attachment #236323 - Flags: superreview?(bugmail)
Attachment #236323 - Flags: superreview+
Attachment #236323 - Flags: review?(bugmail)
Attachment #236323 - Flags: review+
Fixed on trunk.
Status: NEW → RESOLVED
Closed: 18 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla1.8.1
Comment on attachment 236323 [details] [diff] [review]
branch patch

a=beltzner on behalf of 181drivers
Attachment #236323 - Flags: approval1.8.1? → approval1.8.1+
Checked in to MOZILLA_1_8_BRANCH.
Keywords: fixed1.8.1
Have not seen any reports on branch talkback (Firefox2) since checkin.
*** Bug 330683 has been marked as a duplicate of this bug. ***
Crash Signature: [@ SearchTable]
Component: DOM → DOM: Core & HTML
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: