In the 12-17 builds (the first business day after Pavlov's landing), topcrashes at PR_EnterMonitor started appearing on Windows: PR_EnterMonitor [../../../../pr/src/threads/prmon.c line 93] PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 499] USER32.dll + 0x3a5f (0x77d43a5f) browseui.dll + 0x710eb (0x75ff10eb) More information that I can't get to should be available on the internal talkbalk servers, which I don't have access to. A sampling of user comments: (1538311) Comments: opened mail got a "failed to connect to server" message (though I was online). I clicked OK and it crashed. (1501638) Comments: reading mail (1501348) Comments: reading mail (1486704) Comments: Crashed using 'n' to advance to the next unread newsgroup. (1486078) Comments: I had started mozilla mail 5 minutes earlier. I have 5 email accounts. In the right-hand window pane I had right-clicked on the email addresses and selected "Get new messages..." or something like that for the first 3 of my email addresses. When I did (1486078) Comments: this with the fourth email address I apparently typed in the wrong password or at least the software gave the window box saying PASS blah blah blah (password not accepted). Then I got the Application error. Dont' remember whether the error happened (1486078) Comments: by itself or after i clicked OK on the passwd error box. I might have had some email downloading when I was right clicking on other email accounts to download them simultaneously but frankly I think this time my mailboxes were all pretty light. And (1486078) Comments: in any case that has worked fine previously. I only mention it in case you can rule out other causes it might hint at a threading bug of some kind (yah i'm a programmer). (1476397) Comments: Crashed on cancellation of password prompt for sending email (1473517) Comments: Trying to send e-mail kept prompting for password entering correct password brought up dialog box again clicking on cancel caused mozilla to crash (1440515) Comments: As the site was loading ZDNet asked if I wanted to set a cookie. I said yes and it crashed. (1426276) URL: www.vvamsterdam.nl (1426276) Comments: clicked on a link (1416191) URL: http://uptime.netcraft.com/up/graph?mode_u=off&mode_w=on&site=direct.where2getit.com&submit=Examine (1377000) URL: http://vadorguy.multimania.com/ (1377000) Comments: browsing site (1376800) URL: http://vadorguy.multimania.com/down.htm (1376800) Comments: loading page (1298804) Comments: Clicked OK on the dialog that Mozilla couldn't connect to a server and it crashed. (1296816) URL: http://www.chip.de/dooyoo/index.html?itemID=51235 (1264596) Comments: pressing exit in account creation after starting addressbook (1238906) Comments: Crash using 'n' to advance to next newsgroup. (1216263) URL: home.netscape.com (1200563) URL: . (1162124) Comments: work with an internally developed IE centric application (but I want to use Mozilla rather than IE....) (1161583) Comments: Crashed using the 'n' key to advance to the next newsgroup with unread messges.

Stacks available? /be

It's lines 4-8 of comment #0.

Sorry, I skimmed too fast. Who should be cc'd on this bug who has access to the internal-to-netscape.com talkback servers? /be

the eip is null. Not sure how we could have gotten into this state. I was thinking that the queue could have been being deleted (popped) and somehow a timer event was posted to it... Aren't these the best bugs to work on? :-/

Talkback doesn't show much of a stack for this one: ==================================================================================================== Count Offset Real Signature [ 30 PR_EnterMonitor 9c120599 - PR_EnterMonitor ] [ 18 PR_EnterMonitor 38f2bb6d - PR_EnterMonitor ] [ 5 PR_EnterMonitor b3a5baa3 - PR_EnterMonitor ] [ 3 PR_EnterMonitor a99bf8bf - PR_EnterMonitor ] [ 2 PR_EnterMonitor 5339c59a - PR_EnterMonitor ] [ 1 PR_EnterMonitor fbd25ea6 - PR_EnterMonitor ] [ 1 PR_EnterMonitor 948b18fb - PR_EnterMonitor ] [ 1 PR_EnterMonitor 795b67a3 - PR_EnterMonitor ] Crash date range: 2002-01-11 to 2002-01-21 Min/Max Seconds since last crash: 6 - 433949 Min/Max Runtime: 12 - 444840 Keyword List : Count Platform List 39 Windows NT 5.0 build 2195 23 Windows NT 5.1 build 2600 Count Build Id List 22 2002011106 10 2002011509 5 2002011809 5 2002011709 5 2002011409 3 2002011506 2 2002012109 2 2002012011 2 2002011911 2 2002011406 1 2002011706 1 2002011615 1 2002011609 1 2002011606 No of Unique Users 55 Stack trace(Frame) PR_EnterMonitor [../../../../pr/src/threads/prmon.c line 93] PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 499] USER32.DLL + 0x2e98 (0x77e12e98) 0x5e204689 (1954871) URL: shoplet.com (1936268) URL: http://sfgate.com (1936268) Comments: deleted buddy from buddy list in aim (1934540) URL: http://hardware.mp3.com (1934540) Comments: viewing hardware.mp3.com and refusing to set a cookie (1922783) Comments: Installed mozilla; first-time startup as part of installation process (1902068) Comments: I was finnishing the setup of a mail account (1885396) URL: http://www.geocities.com/whatsup66us/ (1885396) Comments: I had pasted this address from a SPAM into the URL bar and hit enter. (1834477) Comments: crashing after switching between aol accounts and then sending not sure if this is related to the crash or not. (1797826) URL: shopping.yahoo.com (1784413) URL: www.rtl.de (1738970) URL: http://www.geeknik.net:2082/frontend/hrcp/mysql/index.html (1720090) Comments: hitting the offline icon when downloading articled (1705056) URL: www.match.com (1701499) URL: http://www.espn.com (1689778) URL: http://www.ritlabs.com/ (1689778) Comments: loading page (1672335) Comments: trying to save an e-mail which was not being delivered so I could shut down mozilla and try again. v. annoying. this was after repeating the experiment several times. (1631481) Comments: Died on first startup (1609742) Comments: Clicked a link (1597583) Comments: I just clicked on the address field and this crash occured. But we have plenty of user comments! ;-)

nominating topcrash bugs for nsbeta1.

jpatel: my talkback searches show this stopping around 2002-02-20 or so. Can you confirm this? Is this fixed now?

This was a topcrasher with Mozilla 0.9.8 but since then there have only been a few incidents on the MozillaTrunk. The last one being with build 2002022605 on Windows. If we don't have a clue about what might have fixed this after that date...I think we should look into it some more before marking it fixed or worksforme. Adding qawanted to see if anyone can reproduce this with a recent Trunk build. I'll attach all the user comments from M098 and if we can't reproduce and this doesn't show up in Mozilla 0.9.9, I'll mark it worksforme.

I've attached all the user comments and urls submitted by M098 users...we should try to reproduce this from that info if possible.

per adt, not critical for nsbeta1. hence minus.

Since this doesn't seem to be happening very often anymore, can we remove topcrash? I expect that various fixes between me and brendan have helped to resolve the problems.

I would rather just keep the topcrash keyword and mark this worksforme...we can always reopen it if this shows up again in topcrash reports. I was just waiting for more M099 data to verify that this wasn't happening anymore.

There have been a few of these crashes with recent MozillaTrunk builds. Here is some Talkback data: Count Offset Real Signature [ 4 PR_EnterMonitor 38f2bb6d - PR_EnterMonitor ] [ 2 PR_EnterMonitor a99bf8bf - PR_EnterMonitor ] [ 2 PR_EnterMonitor 9c120599 - PR_EnterMonitor ] [ 1 PR_EnterMonitor d8d2d12a - PR_EnterMonitor ] [ 1 PR_EnterMonitor 795b67a3 - PR_EnterMonitor ] Crash date range: 2002-03-20 to 2002-03-24 Min/Max Seconds since last crash: 31 - 73008 Min/Max Runtime: 31 - 192924 Keyword List : Count Platform List 6 Windows NT 5.1 build 2600 3 Windows NT 5.0 build 2195 1 Windows 98 4.10 build 67766446 Count Build Id List 4 2002032211 4 2002032205 1 2002032110 1 2002031711 No of Unique Users 8 Stack trace(Frame) PR_EnterMonitor [../../../../pr/src/threads/prmon.c line 93] PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 526] USER32.dll + 0x3a5f (0x77d43a5f) 0x5e204689 (4385106) Comments: went to a web site and *Bam* everything dissapeared. ==================================================================================================== Count Offset Real Signature [ 3 PR_EnterMonitor aef27f05 - PR_EnterMonitor ] Crash date range: 2002-03-20 to 2002-03-23 Min/Max Seconds since last crash: 4271 - 42229 Min/Max Runtime: 13825 - 90235 Keyword List : Count Platform List 3 Windows NT 5.0 build 2195 Count Build Id List 2 2002031809 1 2002032103 No of Unique Users 2 Stack trace(Frame) PR_EnterMonitor [prmon.c line 93] PL_ProcessPendingEvents [plevent.c line 526] USER32.DLL + 0x1b60 (0x77e11b60) 0x5e204689 (4308652) Comments: xslt again (4262085) Comments: testing a page with xml+xslt that includes another xslt ==================================================================================================== Count Offset Real Signature [ 2 PR_EnterMonitor 808082d7 - PR_EnterMonitor ] [ 1 PR_EnterMonitor fccc724d - PR_EnterMonitor ] [ 1 PR_EnterMonitor 85ea588b - PR_EnterMonitor ] Crash date range: 2002-03-25 to 2002-03-25 Min/Max Seconds since last crash: 11 - 18525 Min/Max Runtime: 9537 - 18525 Keyword List : Count Platform List 3 Windows NT 5.0 build 2195 1 Windows 98 4.10 build 67766446 Count Build Id List 3 2002032505 1 2002032411 No of Unique Users 3 Stack trace(Frame) PR_EnterMonitor [../../../../pr/src/threads/prmon.c line 93] PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 526] _md_EventReceiverProc [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 1072] nsAppShellService::Run [d:\builds\seamonkey\mozilla\xpfe\appshell\src\nsAppShellService.cpp line 309] main1 [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp line 1366] main [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp line 1701] WinMain [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp line 1719] WinMainCRTStartup() KERNEL32.DLL + 0xd326 (0x77e8d326) (4461223) Comments: Was composing an email and new mail arrived the notifier popped up and was retreating when Moz went BOOM!.

This is still the #3 topcrash on the trunk (based on data for 3-23 to 3-27, and excluding a smoketest blocker for one day) and, as such, probably needs to be fixed for mozilla1.0. Any chance of reconsidering the target milestone here?

I think the biggest problem is there is no reproducable testcase. If you can give me one, I will certainly do my best to figure out what the problem is.

This continues to be a major topcrasher on the MozillaTrunk (currently #4). It seems that this stack signature is being reported for a lot of different crashes (based on user comments)...so it might be difficult to debug. But maybe we can figure out what's going on if someone can find a single testcase to reproduce this crash. Here are recent Trunk crashes groups by unique stack traces: ==================================================================================================== Count Offset Real Signature [ 18 PR_EnterMonitor 808082d7 - PR_EnterMonitor ] [ 17 PR_EnterMonitor fccc724d - PR_EnterMonitor ] [ 6 PR_EnterMonitor 85ea588b - PR_EnterMonitor ] [ 4 PR_EnterMonitor 80f8fe1b - PR_EnterMonitor ] [ 3 PR_EnterMonitor 2e48d18f - PR_EnterMonitor ] Crash date range: 2002-03-25 to 2002-04-02 Min/Max Seconds since last crash: 2 - 567356 Min/Max Runtime: 2 - 567356 Keyword List : mail(8), Count Platform List 40 Windows NT 5.0 build 2195 7 Windows 98 4.10 build 67766446 1 Windows NT 4.0 build 1381 Count Build Id List 10 2002032810 8 2002032711 5 2002032610 4 2002040110 4 2002032918 4 2002032505 3 2002033111 3 2002032510 2 2002040210 2 2002032411 1 2002040106 1 2002033011 1 2002032605 No of Unique Users 34 Stack trace(Frame) PR_EnterMonitor [../../../../pr/src/threads/prmon.c line 93] PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 532] _md_EventReceiverProc [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 1078] nsAppShellService::Run [d:\builds\seamonkey\mozilla\xpfe\appshell\src\nsAppShellService.cpp line 309] main1 [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp line 1366] main [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp line 1701] WinMain [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp line 1719] WinMainCRTStartup() KERNEL32.DLL + 0xd326 (0x77e8d326) (4758151) URL: http://www.njtransit.com (4758151) Comments: planning an itinerary (4752124) URL: https://akoim1.us.army.mil/login.cfm (4734391) Comments: 2002-03-28 Win32 trunk build. I tried to log in on AOL Doc page at AOL internal site. It asked for user ID password etc. WhenI supplied all the info and clicked on OK it crashed. (4732632) URL: ftp://ftp.rfc-editor.org/in-notes/rfc3252.txt (4732632) Comments: Got an alert box "250 users already connected. try again in 10 mins". Clicked OK then crash. (4724921) URL: www.pocketgear.com (4722828) URL: ftp.idsoftware.com (4722828) Comments: deleted an e-mail account and clicked OK - crash (4721738) Comments: send msg. I have spell checker come up automatically on send. (4719368) Comments: Crash inserting a link into a mail message. (4708638) Comments: I just changed the IMAP address for my mail server from its IP to its domain (4703792) Comments: Prompted for login and entered login information (4699426) URL: http://www.inmagic.com/ (4699426) Comments: Trying to open location in new Nav. Window (4650243) Comments: I changed the address for my mail sever from its "real name" to its IP Address :-/ (4631754) URL: http://www.geocities.com/cartoon_studio2/CDRR-BOTB_13-24.htm (4631648) URL: http://www.geocities.com/cartoon_studio2/CDRR-BOTB_13-24.htm (4631648) Comments: Opened an AvantGo page (account management I believe). (4616746) Comments: Mail Account preferences (4597668) Comments: open using open web location dlg (4595351) URL: ftp.idsoftware.com (4587637) URL: www.sonique.com (4587595) URL: www.sonique.com (4575166) Comments: I was edit a received email as new and addressing it to a bunch of friends. When I had finished selecting the addresses I pressed OK and wammo (4544197) URL: http://www.kicxfm.com/ (4544182) URL: http://www.kicxfm.com/ (4544127) URL: http://www.kicxfm.com/ (4544127) Comments: Tried to add a new address to my address book in Mozilla Mail while I was composing a note. (4527816) URL: verizon.com (4527816) Comments: using the function to find a phone number and walah!! (4526832) Comments: tabbed browsing (4503473) Comments: adding people to bcc via address button (4461223) Comments: Was composing an email and new mail arrived the notifier popped up and was retreating when Moz went BOOM!. ==================================================================================================== Count Offset Real Signature [ 11 PR_EnterMonitor 98ab1501 - PR_EnterMonitor ] [ 4 PR_EnterMonitor 4e0f4514 - PR_EnterMonitor ] Crash date range: 2002-03-25 to 2002-04-02 Min/Max Seconds since last crash: 255 - 437467 Min/Max Runtime: 255 - 437467 Keyword List : compose(6), Count Platform List 15 Windows NT 5.1 build 2600 Count Build Id List 4 2002040106 3 2002032810 2 2002040210 2 2002032711 1 2002033011 1 2002032906 1 2002032806 1 2002032510 No of Unique Users 9 Stack trace(Frame) PR_EnterMonitor [../../../../pr/src/threads/prmon.c line 93] PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 532] _md_EventReceiverProc [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 1078] USER32.dll + 0x3c076 (0x77d7c076) USER32.dll + 0x3c076 (0x77d7c076) _except_handler3() kernel32.dll + 0x3bb86 (0x77e9bb86) (4748369) Comments: inserting an image into a new Composer page (4747061) Comments: inserting image into new Composer page (4737014) Comments: attempting to insert an image in composer. (4710585) URL: http://www.virtualgamestation.com (4709590) Comments: adding hyper link to a page in Composer. (4708398) Comments: publishing a page in Composer (4704745) Comments: imbedding tables within tables in Composer. ==================================================================================================== Count Offset Real Signature [ 4 PR_EnterMonitor 768d60ee - PR_EnterMonitor ] Crash date range: 2002-03-25 to 2002-03-31 Min/Max Seconds since last crash: 1516 - 8159 Min/Max Runtime: 1516 - 8159 Keyword List : Count Platform List 4 Windows NT 5.1 build 2600 Count Build Id List 2 2002032918 1 2002033011 1 2002032411 No of Unique Users 4 Stack trace(Frame) PR_EnterMonitor [../../../../pr/src/threads/prmon.c line 93] PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 532] _md_EventReceiverProc [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c line 1078] USER32.dll + 0x3c076 (0x77d4c076) USER32.dll + 0x3c076 (0x77d4c076) _except_handler3() kernel32.dll + 0x3bb86 (0x77e7bb86) (4657113) Comments: Launching ChatZilla while downloading a mozilla talkback in the same time ! (4439629) Comments: Launching Chatzilla ! :-( I'll try to reproduce this, but it might be a good idea to get more QA people cc'd on this to see if they can somehow get this crash too. Pav: Can you add a few QA people that know this component?

Is bug 121532 related to this by any chance ? I still see the hangs from time to time and the stack trace there also shows hang in PR_EnterMonitor ().

Talkback staff: Can I get a list of users (at Netscape) who have reported a crash in this area?

This isn't, by any wild change, related to bug 134070, is it? I see a lot of 'password dialog' comments in this bug, as I saw in that bug.

Lisa, here is a list of internal crashers on the Trunk at PR_EnterMonitor. (You're on it, BTW ;-)

*** Bug 135693 has been marked as a duplicate of this bug. ***

Stuart: The dupe has a 100% reproduceable testcase

On M099 I have been able to reproduce this crash once (Incident 4896293), but it is not 100% for me. In fact, for my crash I only had to click on "Kontakt" in the top links to get the crash. All the same, good place to start. Renominating and boosting to topcrash+

This looks (from talkback data) like it may well have been mostly fixed by the fix for bug 134070 (there's only been one report of this crash since). Does the reproducable testcase still crash in current builds?

i still get a crash with build 20020408.. and bug 134070 is fixed at 200200405 i dunno if it's the same crash because my debug is out of date...

Using today's Windows Trunk build (20020410xx) I was able to crash using steps similar to those in the duped bug 135693. 1. Go to http://www.golfinfo.at/test/start.asp 2. Click on "Kontakt" at the top 3. Click around quickly on the links in the right column -> Crash

I just created a new profile (blank), set up my mail preferences, start up the mail client, and press the 'Enter' key while the focus is in the text input box, and crash. Clicking on the 'OK' button with the mouse or tabbing to 'OK' and pressing enter enables the user to log into mail without problem. So it appears that the event handling/generation (or some call therein) in the text box seems to be the gremlin. see bug 136939

What are the chances, tht what is left to resolvbe for this crash, will be done by Friday, 04.19?

considering pavlov is out this week, I would say zero to none. :-/ I can try looking at this after my blocking bugs.

how interesting is this.... It looks like the nsScriptLoader is being passed around as the context of a async file read. The context is being release during the onStopRequest leaving a single reference owned by the nsFileTransport. The neFileTransport make the assuption that any context can be released on any thread and releases the object after it makes the onStopRequest.

Updating summary M098 -> M100 (branch)

Has anyone been able to reproduce this on the MozillaTrunk or Mozilla1.0 Branch? Talkback data shows that this last crashes in big numbers with Trunk builds on 4/10. Since then there has only been one crash on the Trunk and Mozilla1.0 RC1 only has about 10 crashes reported. I'm thinking dbaron's comment #24 might be valid, the fix for bug 134070 helped fix this crash also (at least most of them). If no one is able to reproduce this crash anymore, I think this bug might be worth marking worksforme.

it seems that the testcase work for me now with 20020430.. (no crash)

what jpatel said. wfm.

v.wfm

Reopening for trunk and M120B. Trunk: Total blackboxes in this sample: 50 Total unique users: 34 PR_EnterMonitor (50 crashes) Crash data range: 2002-11-22 to 2002-11-25 Build ID range: 2002112204 to 2002112308 Stack Trace: PR_EnterMonitor [prmon.c line 93] PL_ProcessPendingEvents [c:/builds/seamonkey/mozilla/xpcom/threads/plevent.c line 553] USER32.dll + 0x3a68 (0x77d43a68) 0x5e204689 Source File : prmon.c line : 93 14332655 OS: Windows NT 5.1 build 2600 Comments: right click highlighted text for web search crash! then a windows pops up saying talkback.exe No disk in drive F: 14321712 OS: Windows NT 5.0 build 2195 Comments: Just clicking on the link above. URL: http://photoalbum.powershot.de/fotos/58/30//6e7a4a2ee200cf4ae18640c10ec7e91c.jpg 14318825 OS: Windows NT 5.1 build 2600 URL: http://www.moxi.com 14318819 OS: Windows NT 5.1 build 2600 URL: http://www.moxi.com 14317566 OS: Windows NT 5.1 build 2600 URL: http://www.ifilm.com/ifilm/product/film_multimedia/0 4470 2446644 00.html 14316117 OS: Windows NT 5.1 build 2600 Comments: Highlited text cut text into buffer then crash. Then a message poped up on my screen: "Please insert disk into drive F:" Drive F: is my CD drive 14293828 OS: Windows NT 5.1 build 2600 Comments: Clicked the play button: javascript:launchEP('/ifilm/media/player2/1 4566 2446661 00.html?fid=2446661'); URL: http://www.ifilm.com/ifilm/product/film_multimedia/0 4470 2446644 00.html 14281184 OS: Windows NT 5.0 build 2195 URL: It always does this... 14280572 OS: Windows NT 5.1 build 2600 URL: something on xbox.com 14280330 OS: Windows NT 5.1 build 2600 URL: something on xbox.com 14277118 OS: Windows NT 5.1 build 2600 Comments: Opening a trailer for NFL Fever 2003 URL: something on xbox.com 14273897 OS: Windows NT 5.1 build 2600 URL: yahoo.groups.com 14256779 OS: Windows NT 5.1 build 2600 Comments: sending an email :) URL: www.geocities.com/moz_blubber M120B: Crash date range: 2002-11-14 to 2002-11-24 Min/Max Seconds since last crash: 118 - 1830421 Min/Max Runtime: 118 - 2254604 Keyword List : browser(4), click(13), crash(6), java(7), link(9), load(5), mail(7), site(5), popup(6), Count Platform List 49 Windows NT 5.1 build 2600 28 Windows NT 5.0 build 2195 14 Windows 98 4.10 build 67766446 8 Windows 98 4.90 build 73010104 4 Windows 98 4.10 build 67766222 Count Build Id List 103 2002101612 No of Unique Users 89 Stack trace(Frame) PR_EnterMonitor [prmon.c line 93] PL_ProcessPendingEvents [c:/builds/seamonkey/mozilla/xpcom/threads/plevent.c line 553] USER32.dll + 0x3a68 (0x77d43a68) 0x5e204689 (14311218) URL: salon.com (14292982) URL: MAIL.YAHOO.COM (14259996) URL: http://www.business2.com/articles/web/0 1653 43198|2 FF.html (14259996) Comments: One tabbed page set to http://webmail.shaw.ca - not logged in just sitting at the login page. The other tabbed page was at the link mentioned above. SEQUENCE: The failure occured when I clicked on the 'Next Page' link at the bottom-right of the (14253527) URL: www.osnews.com (14253527) Comments: Going from slashdot.org to www.osnews.com. (14252532) Comments: Opeining the Druge Report in a new tab. (14200518) URL: http://www.axioma.es (14198847) Comments: opening 10 (more or less) websites tabs. (14194026) Comments: playing a game (14192637) URL: www.pclinuxonline.com (14192637) Comments: switching from browser to mail module (14132916) URL: www.sprintpcs.com (14132916) Comments: just stopped when I went to shopping (14089754) URL: sbcglobal.prodigy.net (14089754) Comments: Sending form info. "Can be read by outsiders" box popped up hit OK then crash. (14076888) URL: www.toptronics.nl (14052532) URL: http://www.cable-modems.org/articles/ (14052532) Comments: many tabs open but Resource Meter still shows all resources over 45% free. (14004629) URL: www.vietcong-game.com (14004629) Comments: opening a picture...i had another window open at www.sharereactor.com (13985997) URL: http://jjaf.de/eci/hi-focus/atu-r/ (13985997) Comments: pressing N to get to next message 3 times then hitting a return-recipt one asking. clicking no. (13978746) Comments: Browsing the web. Just clicked to an http:// URL and browser crashed. (13978631) URL: http://www.cuo.org.uk/ (13976630) URL: www.delphi-center.com (13976630) Comments: reading mail (13974482) URL: http://www.fandango.com/movie_page.asp?mv=42066 (13974482) Comments: Clicked a link to see pictures about some lame Segal movie (was probably trash anyway). It may have involved a plugin. The link was to a javascript popup.... javascript:launchRemote('media_popup.asp?type=PHOTOS&mv=42066' 'MediaPhoto' 542 400 (13940824) URL: www.misjump.com (13933121) Comments: Accessing routing table of Linksys firewall box (on internal interface) (13932781) URL: www.fastpath.de

:( i was using imap to read the (bug)mail i got over the past ~10 days, i was almost done. i also occasionally loaded bugrelated pages. i didn't realize mozilla was dead at first (msvc was behind the windows, and i had switched to my irc deskstation). w2k, cvs build from ~2weeks ago PR_EnterMonitor(PRMonitor * 0xcdcdcdcd) line 92 + 3 bytes PL_ProcessPendingEvents(PLEventQueue * 0x111e4bb8) line 551 + 13 bytes _md_EventReceiverProc(HWND__ * 0x00dd0690, unsigned int 49306, unsigned int 0, long 287198136) line 1335 + 9 bytes USER32! 77e13eb0() USER32! 77e1401a() USER32! 77e13f0f() nsAppShellService::Run(nsAppShellService * const 0x01047898) line 472 main1(int 4, char * * 0x002844c0, nsISupports * 0x00276f08) line 1541 + 32 bytes main(int 4, char * * 0x002844c0) line 1902 + 37 bytes mainCRTStartup() line 338 + 17 bytes KERNEL32! 77e87903() _md_EventReceiverProc(HWND__ * 0x00dd0690, unsigned int 49306, unsigned int 0, long 287198136) line 1335 + 9 bytes queue: 111E4BB8 08000080 01000000 68CFB510 CDCDCDCD CDCDCDCD ...А....h╧╡.════════ 111E4BCC CDCDCDCD CDCDCDCD CDCDCDCD 00000000 CDCDCDCD ════════════....════ 111E4BE0 FDFDFDFD 0DF0ADBA ABABABAB ABABABAB 00000000 ¤¤¤¤.Ён║лллллллл.... - queue 0x111e4bb8 |+ name 0x80000008 "" \- queue {...} |+ next 0x00000001 \- prev 0x10b5cf68 <- points to garbage |+ next 0x025867b0 <- points to garbage \+ prev 0x00000001 Queue is mostly uninitialized garbage, there are visible endpoints for the linked list, but the other two entries are pure garbage. Could we add extra logging to help catch this? my build has lots and lots of debug stuff enabled and i wouldn't mind adding a bit more. (note that i don't think i've hit this before, i'm just very upset because i suspect my imap session activity was never sent to the server).

Adding M121 to summary since this is also a topcrash with Mozilla 1.2.1. Not sure when this started, but it is killing us on the Trunk...currently the #1 crasher (and it has been for a few weeks).

timeless mentioned imap, cc'ing bienvenu. /be

In timeless's case, it looks like he has an event queue with an uninitialzied monitor (0xcdcdcdcd). Is that even possible? As far as IMAP is concerned, something like this stack tends to happen when shutting down, if I remember correctly.

my crash might be related to two other threads in bug 183497.

*** Bug 183814 has been marked as a duplicate of this bug. ***

*** Bug 184615 has been marked as a duplicate of this bug. ***

not sure what the conclusion is, but this is not only imap related. I crash whenever I click on link that shows an image in a new window. TB14926445Z TB14926413X TB14926335H TB14926310Q TB14926191Z TB14926153Z TB14926127K TB14926081H 20021210 on WinXP

ok. another mail crash, stack lost due to tired user... I composed an email to scc/dbradley, pressed ctrl-enter(send), looked at it, realized one digit of my phone number was missing in one of two instances, pressed escape a lot (nothing happened), realized that mail wasn't going to send it because smtp needed a password, went to that dialog, canceled, got the mail could not be sent, clicked ok, >died sometime here<. self->monitor was 0 this time. That's a solvable case. I think we should write code which handles that case, i'm not sure what portion of the crashes it would fix but given how many hits we have I believe it's worth it. -- that said, i'm gone so if anyone is interested in explaining why this is wrong, i'll read the bugmail when i return from my trip. -- now i have to rewrite the email.

Here's a certain crasher... [porn] warning! 1) go to http://www.martl.sixeven.com/sophie.html 2) click on a image, view image, close image 3) repeat 2 until you crash which you quickly will especially: 5247054.jpg 5247063.jpg seems to be good crasher...

Updating summary with M130A since this is the #1 topcrasher for Mozilla 1.3 Alpha. If anyone is interested in seeing the Talkback data, let me know and I'll attach it.

I have this to offer: TB: 15482743 Call Stack: (Signature = PR_EnterMonitor d8d2d12a) PR_EnterMonitor [prmon.c, line 93] PL_ProcessPendingEvents [c:/builds/seamonkey/mozilla/xpcom/threads/plevent.c, line 572] KERNEL32.DLL + 0x363b (0xbff7363b) 0x5e204689 is this the same as this bug? Is bug 179976 maybe a duplicate of this one?

*** Bug 186805 has been marked as a duplicate of this bug. ***

*** Bug 187135 has been marked as a duplicate of this bug. ***

*** Bug 187345 has been marked as a duplicate of this bug. ***

*** Bug 188113 has been marked as a duplicate of this bug. ***

*** Bug 188356 has been marked as a duplicate of this bug. ***

*** Bug 188352 has been marked as a duplicate of this bug. ***

This testcase seems to work for me _really_ reliably on Windows 2000's 2003-01-13-08 trunk build: 1. Launch www.millertime.com 2. Do the age verification 3. Click on TV Ads 4. Click on 'Launch Site' 5. Close the window that comes up 6. Click on 'Launch Site' Keep repeating steps 5 and 6 until you crash.

I've got yet _another_ testcase ;-) 1. Load www.hungryminds.com 2. Click on the logo at the top. This usually crashes for me.

I ran into this under Purify. The attachment contains the stack traces for, allocation, free, and access. Looks like some window fired an event and tried to access a dead queue.

*** Bug 189002 has been marked as a duplicate of this bug. ***

i found the problem. patch and explaination in a bit.

the problem, in a nut shell, is that the notification window callback is called after we destroy the event queue. At the point where we destroy the event queue, we would hope that all native notification stop. :-) It is clear with the purify data that dbradley provided (thank you) that the last reference to the event queue is being held by the file transport. These transports exist on a non-ui thread. Furthermore, all native event queues are created on the ui thread. We destroy the window with a call to DestroyWindow. However, the documentation about this function states that it does nothing to the window handle if the window was created on a different thread. Therefore, our attempt to stop the notification in this case does nothing. The workaround is to send a close message to the window handle instead of calling DestroyWindow. The DefWindowProc will do the right thing when WM_Close is handled (DefWindowProc will call DestroyWindow). The patch is very simple and should not cause any performance problems.

Comment on attachment 111571 [details] [diff] [review] Send WM_Close instead of DestoryWindow sr=jst

Comment on attachment 111571 [details] [diff] [review] Send WM_Close instead of DestoryWindow r=darin (nice catch!)

Checking in plevent.c; /cvsroot/mozilla/xpcom/threads/plevent.c,v <-- plevent.c new revision: 1.30; previous revision: 1.29 done Lets see if this works. Please reopen this bug if we still get reports of crashes with this signature. Also, I am sure that there are a ton of dups. Please dup them against this bug.

*** Bug 179976 has been marked as a duplicate of this bug. ***

*** Bug 186973 has been marked as a duplicate of this bug. ***

marking fixed.

*** Bug 189459 has been marked as a duplicate of this bug. ***

*** Bug 189124 has been marked as a duplicate of this bug. ***

*** Bug 189124 has been marked as a duplicate of this bug. ***

verifying fixed. no crashes in talkback data since 1/14 builds.

*** Bug 190385 has been marked as a duplicate of this bug. ***

Comment on attachment 111571 [details] [diff] [review] Send WM_Close instead of DestoryWindow simple fix that, 1.0 branch topcrash

*** Bug 192939 has been marked as a duplicate of this bug. ***

*** Bug 190602 has been marked as a duplicate of this bug. ***