Closed Bug 239223 (ghostproc) Opened 21 years ago Closed 3 years ago

[Meta] firefox.exe doesn't always exit after closing all windows; session-specific data retained

Categories

(Firefox :: General, defect)

x86
Windows XP
defect
Not set
critical

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: k.m.duncan, Unassigned)

References

(Blocks 1 open bug)

Details

(Keywords: meta, privacy, Whiteboard: security impact in comment 4. may be several bugs.)

Attachments

(2 files)

User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.6) Gecko/20040206 Firefox/0.8 Build Identifier: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.6) Gecko/20040206 Firefox/0.8 Hidden window can store username/password called via .htaccess Exiting the application apparently completely does not "reset" the application. Upon further examination (CTR+ALT+DEL) it appeared there was a "hidden" Firefox window/component running. Suspect this may be an intermittent fault - difficult to debug :-( Problem with the OS or a problem with Firefox? Perhaps future releases could include some sort of check on exit/shutdown? Reproducible: Didn't try Steps to Reproduce: 1. 2. 3. Sincere apologies if this is already known about - somehow I suspect not. To the end user it *appears* as if everything is ok... he/she closes the application - normally the system would then "forget" all about usernames/password. However with a hidden window running in the background (difficult to replicate) it perhaps leads the user into a false sense of security. I discovered this problem when I was setting up a secure directory... I couldnt understand what was going on.. I checked the server directory permissions..etc.. One directory was asking me for userID - whilst the other was letting me in no problem. It was only when I looked at background processes/applications I saw a suspicious "Firefox" entry - I closed this down and checked the website and sure enough htaccess was performing as I would expect. For the stuff I am developing it is maybe not so important - but in some instances security could be compromised - the only real workaround I can think of in the meantime is to reset your PC after entering "secure" data ;-)
If Firefox sticks around after you close all windows, that's a bug. Does Firefox always stick around after you close the last window or have you only seen it happen once?
(In reply to comment #1) > If Firefox sticks around after you close all windows, that's a bug. Does > Firefox always stick around after you close the last window or have you only > seen it happen once? Prior to this I had only seen this once, however in I *think* I may have been able to replicate this...if the user does a lot of frantic mouse clicking... triple clicking to launch successive sessions (perhaps 1 or 2 seconds between triple clicks) - then as soon as the window appears..before it has a chance to open properly closes the window. Repeat the cycle 2 or 3 times maybe... eventually I am left with no windows visible. but a "hidden" Firefox application running in the background. Obviously *normally* people might not be triple clicking or opening/closing windows so quickly in such a manner. Then again if the environment is quite busy... multiple windows open.. multi-tasking etc. Depending on system loading and responsiveness of software it *may* leave one of these hidden windows. Normally not a problem - seems to clear itself when a new instance of Firefox is launched thereafter... but *may* be a security issue in *some* circumstances. -- Keith
For reference - dont know if it helps any, I ran Fabertoys - there are two diagnostic files: 621KB: http://www.personal.dundee.ac.uk/~kmduncan/firefox_bug/Firefox.html 30KB: http://www.personal.dundee.ac.uk/~kmduncan/firefox_bug/Firefox2.html I am running Sygate Personal Firewall with kernel32.dll blocked and rpcss. Firefox *appears* to run ok thru Firewall ;-)
Summary: Hidden window data retention issue for htaccess request → firefox.exe doesn't always exit after closing all windows; basic auth data retained
This would be expected behavior if the app isn't shutting down all the way--all sorts of private stuff could still be available to the next guy to sit down at the machine (internet cafe or kiosk?). If an attacker could put a machine into this state the next victim would inherit any logins. Since we can't log out of http auth this could prevent the victim from using some site correctly, though we can hope most interesting sites don't rely on http auth. The victim could be pre-logged on to interesting sites that use cookies, but all of those do have mechanisms for signing off (if the victim notices they're not getting the right data). If the victim then shuts down does the hidden window keep running, or is that the end of it? If the hidden window keeps running then the attacker could come back at the end of the day and potentially glean things from history, cookies, the cache, etc. This is all very vague and speculative... is this bug really one that needs the security sensitive flag? It might get a lot more attention if it were visible to more people
Whiteboard: [sg:needinfo]
(In reply to comment #4) >though we can hope most interesting sites don't > rely on http auth. We can always *hope* but at the end of the day how can the user tell which sites use http auth and which ones dont? We can hope that user agents will have a low no. of bugs We can hope that most webpages will be accessible and validate to W3C stds. We can hope few people will learn of this "bug" and perhaps use to their advantage... > If the victim then shuts down does the hidden window keep running, or is that > the end of it? From what I have seen this hidden window only ends via CTRL+ALT+DEL and termination of application *OR* through a system shutdown. >If the hidden window keeps running then the attacker could come > back at the end of the day and potentially glean things from history, cookies, > the cache, etc. > My gut feeling is this may be possible....difficult for me to test... I am not a programmer/hacker ;-) > This is all very vague and speculative... is this bug really one that needs the > security sensitive flag? It might get a lot more attention if it were visible to > more people Perhaps I misunderstood the purpose of the "security sensitive flag".
*** Bug 261245 has been marked as a duplicate of this bug. ***
Confirming bug based on number of reports, clearing security flag because it's not protecting anyone.
Group: security
Severity: enhancement → critical
Status: UNCONFIRMED → NEW
Ever confirmed: true
Alias: ghostproc
*** Bug 246942 has been marked as a duplicate of this bug. ***
*** Bug 248033 has been marked as a duplicate of this bug. ***
*** Bug 248320 has been marked as a duplicate of this bug. ***
*** Bug 259144 has been marked as a duplicate of this bug. ***
*** Bug 260331 has been marked as a duplicate of this bug. ***
*** Bug 249120 has been marked as a duplicate of this bug. ***
From bug 260331 > Installed extensions : Tabbrowser, Adblock, SingleWindow, Download Statusbar, > Download Sort, Linkification > Installed Theme : Noia 2.0 This similar to any of the other dupes? Were some of the duplicate reports a straight clean install with no extensions or themes?
I've experienced this (only on win2k/xp, and only using downloaded nightly builds) in what I believe to be clean (no extensions) situations. Though with my development environments, I can't guarantee it. In my situation, however, the hung firefox process doesn't allow new windows to be opened, so to the end-user it appears that firefox just isn't starting at all.
Assignee: firefox → nobody
Flags: blocking-aviary1.0?
QA Contact: bsmedberg
You can see my description from (now closed) Bug 249120. I was able to consistently replicate this on clean installs as well as not-so-clean.
From bug 249120 Reproducible: Always Steps to Reproduce: 1.Browse to http://www.pogo.com 2.Play a game (which opens in a new window). I reproduced this consistently with "Word Whomp". 3.Exit all browser windows
who cana take ownership of this bug. dan, bryner, bsmedberg, any ideas?
Flags: blocking-aviary1.0? → blocking-aviary1.0+
Well my Prerelease Version sometimes does not exit properly as well. There is still a process of firefox running after closing window, that contained a Java Applet running with Java 1.5.0 beta. This does not happen every time. But I can see, that firefox is still open, when the console of Java does not disappear!
To summarize, there are two cases here (after the last window is closed and firefox.exe process still shows up in the Task Manager): 1. Any attempt to start the browser silently fails. Firefox appears dead. 2. Any attempt to start the browser succeeds, and all previous session information is still available. #1 is similar to a standard program crash, and has no easily exploitable security repercussions. (That's how Acrobat Reader plugin behaves on my system all the time.) If not for profile locking, users would not be noticing the problem. The "dead" process is waisting RAM until it's shut down. So, it's mostly an annoyance rather than a security issue. #2 is the real problem, IMO. Somehow, browser keeps a hidden window open. It happened to me a few times in the past. The problem was usually cured by re-opening browser and doing File | Exit. It's my impression, that I get #2 when a plugin fails to load, or FF doesn't know how to deal with some fancy/malformed markup, or a popup failed to show up.
Well it is really true, that Firefox remains open, when a plugin had an error. Java says in the console, that "Cookie Service not Available. Looking for Cookie Cache". Everytime this error occurs, Firefox can't be closed regularly.
(In reply to comment #20) > 1. Any attempt to start the browser silently fails. Firefox appears dead. > If not for profile locking, users would not be noticing the > problem. Sounds like an issue I've seen from to to time on my sister's computer, where Firefox shows the profile manager on startup because the profile is locked, and restarting the computer [usually] fixes it. I know that she sometimes uses sites like www.pogo.com. She runs Windows XP.
Flags: blocking-aviary1.0+ → blocking-aviary1.0-
I have this behaviour in connection with another problem: I can't open any new windows (File/New Window, Ctrl-N, Tools/Extensions/Get More Extensions, ...) Where it connects to this bug: o if i didn't try to open any new windows, the process stops after I close the window o if i tried to open a new window the process doesn't stop when I close the only visible one (maybe it thinks there are other windows -- even though they had never appeared on the screen) I run 1.0PR on Win2000, with many-many extensions. I disabled/removed/reinstalled/... all that can have anything to do with tabs/windows and almost all the rest.
I've just started noticing this issue today as well. It seems to happen when I have Firefox open for a couple hours on end with multiple tabs and multiple windows. I'm using XP SP2, Firefox PR with the following extensions: ieview, webdev, minit, linkification, close tab on dbl click, adblock, livehttpheaders, dict search. I am still able to start Firefox with the "ghost" process there (I end up with 2 firefox.exe procs). I haven't yet tried to figure out if session data is still active, so I'll give that a shot next time it happens. I'll see if I can't figure out a way to replicate this problem reliably.
perhaps one of those extensions is misbehaving. have you tried disabling all extensions?
(In reply to comment #25) > perhaps one of those extensions is misbehaving. have you tried disabling all > extensions? Not yet. I just upgraded to 1.0 RC1 and I'll keep an eye on things and see if the problem shows up again. If it does, I'll kill my extensions and dig around.
After I posted #23 I started with a clean profile and the problem went away. I installed all extensions and it still worked fine. This morning I noticed that the ChromeEdit extension didn't appear in Tools so I removed it. Many of the extensions disappeared from Firefox (GoogleBar, WebDeveloper, ...) and the "new window" problem is here again. (but I'm getting wiser: this time I have a copy of my working firefox profile!)
I've reproduced this problem using versions 1.0RC and 1.0 release edition with the following situations: P3, Windows XP Home P4 HP, Windows XP Pro Both using clean installs of Firefox 1.0PR and Firefox 1.0 release. The zombie procs appear consistently and must be killed. No extensions or themes are loaded and the profile is clean. To reproduce, install Firefox and navigate to news.google.com and click 4-6 links. In an open window, navigate to a site requiring a username/password (and enter it correctly). Quitting firefox should leave a zombie.
I can reproduce this reliably on Windows 2000 SP2, Firefox 1.0 release, by enabling the Bookmarks Synchronizer extension. With Bookmarks Synchronizer enabled: - no new windows can be opened, either by Ctl-N or links; - the ghost process remains after the program exits; - the program cannot be started until the process is killed. Without Bookmarks Synchronizer, none of this happens.
(In reply to comment #26) > (In reply to comment #25) > > perhaps one of those extensions is misbehaving. have you tried disabling all > > extensions? > > Not yet. I just upgraded to 1.0 RC1 and I'll keep an eye on things and see if > the problem shows up again. If it does, I'll kill my extensions and dig around. Hmmm.. I dunno if I can live without my extensions. It still happens sometimes though, so maybe I'll get around to it. This afternoon I had 4 ghost processes! Here are the extensions I'm running: - ieview - web developer - liveHttpHeaders - adblock - linkification - close tab on double click - miniT - dictionary search - view cookies (newly installed) My best guess would be that it has something to do with ieview somehow since it has crashed my browser a few times. The problem only seems to happen after extended use.
OS: Windows 98 → Windows XP
*** Bug 274363 has been marked as a duplicate of this bug. ***
*** Bug 274078 has been marked as a duplicate of this bug. ***
*** Bug 275973 has been marked as a duplicate of this bug. ***
a rather unscientific survey reveals 20% of users share this problem http://forums.mozillazine.org/viewtopic.php?t=196962
A possible solution would be to include a counter for how many windows there are open, and every time one is closed, the counter decrements. Once it hits 0, it shuts itself down all the way.
Ever since I've upgraded to Web Developer extension 0.9.2 (I was at 0.8.x), firefox.exe doesn't linger for me. So that may just be the problem. Can anyone else confirm?
*** Bug 275931 has been marked as a duplicate of this bug. ***
*** Bug 280119 has been marked as a duplicate of this bug. ***
I hate this new version of mozilla- I want the old one back at least it had a "user profiles" and "switch user". I have lost all my bookmarks and quick links! Im so ready to just go back to Netscape Navigator, at least my bookmarks and buttons are ALWAYS there!
*** Bug 277074 has been marked as a duplicate of this bug. ***
*** Bug 286918 has been marked as a duplicate of this bug. ***
*** Bug 267327 has been marked as a duplicate of this bug. ***
*** Bug 274907 has been marked as a duplicate of this bug. ***
*** Bug 288231 has been marked as a duplicate of this bug. ***
*** Bug 286174 has been marked as a duplicate of this bug. ***
Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b2) Gecko/20050309 Firefox/1.0+ Still happens for me, after visiting this site that uses java: http://sazraceforthecure.kintera.org/faf/home/default.asp?ievent=92540
*** Bug 263669 has been marked as a duplicate of this bug. ***
*** Bug 288211 has been marked as a duplicate of this bug. ***
I uninstalled Firefox, rebooted, and reinstalled, and the problem did not go away. Starting Firefox in 'safe mode' did not help either. However, when I uninstalled, rebooted, and then deleted my entire profile directory, then reinstalled, the problem was fixed. Perhaps a corruption of the profile, or one of the settings therein, caused the problem.
*** Bug 265270 has been marked as a duplicate of this bug. ***
(In reply to comment #49) > Perhaps a corruption of the profile, or one of the settings therein, > caused the problem. Looks likely. I just installed the latestly nightly on a machine that has never had Firefox on it - the problem doesn't appear there. Having said that, with so many dups, telling users to delete their profile seems like a less than ideal resolution.
The problem for me occers almost ALWAYS when I use all-in-one guestures to close the last tab, so it seems to me that when the TAB is closed instead of the whole window, via top X, it will stay open.
*** Bug 291926 has been marked as a duplicate of this bug. ***
I wrote a small Win32 replacement for firefox.exe that calls the original firefox.exe (renamed to firefox1.exe) after ensuring that all running firefox1.exe processes have at least one visible window. Any firefox1.exe processes that have no visible windows are forcibly terminated. Of course, this is really a hack and not a proposed fix for this issue. I'm thinking that perhaps the fix mentioned in bug #214675 attachment #148991 [details] [diff] [review] can be amended to include a variant of the file I'm attaching instead of ShowOSAlert in that patch at: +#ifndef MOZ_ENABLE_PROFILE_MANAGER + ShowOSAlert("No profile specified. Please start with -P <profile name>\n"); + return NS_ERROR_FAILURE; +#else return ShowProfileManager(profileSvc, aNative); +#endif Just a thought.. I may have time to experiment with it a home and, if so, I'll attach a patch to obsolete this separate program. Here is what I originally wrote in bug #214675 when I initially posted this program to that bug by mistake: ============================================================ We, like many people out there, occasionally have users who "lose" their bookmarks and start page due to something causing Firefox to uncleanly shut down and thus keep the profile locked. I'm not going to tell our end-users to go into the Task Manager to end the process (think "Auntie Mabel" or "Grandma" type of users), so I wrote a small tool that does that for them. I'm hoping that perhaps a part of this can go into the real firefox.exe so that it can kill any previous instances that have no windows open and thus prevent the Profile Manager from appearing (at least, that's one idea). I understand that fixing the root cause of freezing problems rather than working around them is the best approach, but this workaround can handle cases of rogue plug-ins that I suspect are causing Firefox to hang rather than Firefox itself. BTW, thanks for an excellent web browser! The IT Department of a non-profit that I am in, with my help, rolled out Firefox to all of our 100 or so computers as a part of our Windows XP migration. I'll send more details to the appropriate place if I find out what caused one particular user's Firefox to "lose" his bookmarks on two occasions. For now, I've just manually edited profiles.ini to switch his default back to his original to "restore" his bookmarks and make him happy, though I'm hoping that this workaround will prevent this from happening on our agency's computers.
Well, I do have the same problem with firefox running background since i updgraded to 1.0.3. The main problem is that each time I close Firefox, I have to kill it (CTRL+ALT+DEL) to be able to run it again ...
For simplicity, I put the ancillary files and the compiled firefox.exe file up on http://www.moonlightdesign.org/startfirefox/ , though I recommend compiling it yourself after reviewing the code as you should do with random software :-). Because it's structured as a firefox.exe wrapper (after renaming the real firefox.exe to firefox1.exe in the same folder), it includes the three icons that the real firefox.exe has, though I'll definitely take it offline if any objections to including the icons in there come up (just in case if it's a sticky legal situation). BUT, I stress that this is not really a 'solution' but merely a temporary workaround to save the hassle of force-closing zombie firefox.exe processes in Task Manager.
*** Bug 293965 has been marked as a duplicate of this bug. ***
(In reply to comment #23) > I have this behaviour in connection with another problem: I can't open any new > windows (File/New Window, Ctrl-N, Tools/Extensions/Get More Extensions, ...) > > Where it connects to this bug: > > o if i didn't try to open any new windows, the process stops after I close > the window > > o if i tried to open a new window the process doesn't stop when I close the > only visible one (maybe it thinks there are other windows -- even though they > had never appeared on the screen) This is exactly what I'm experiencing. - Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.7.8) Gecko/20050511 Firefox/1.0.4 - Noia 2.0 (eXtreme) 2.82 - Extensions: + User Agent Switcher 0.6.6 + Web Developer 0.9.3 + googlebar 0.9.5.06 + AdBlock v.5 d2 * nightly 39 + BugMeNot 0.6.2 - Options: + Default Browser + Home page is several tabs + Block popup windows + Open links from other applications in: a new tab in the most recent window + Do not hide the tab bar when only one web site is open + Do not select new tabs opened from links + Select new tabs opened from bookmarks or history + Warn when closing multiple tabs I've just included what I think may be pertinent, I could attach a sanitised prefs.js if it would help. I've also noticed that from time to time, when I start Firefox, I see a single, empty tab, even though I get prompted for HTTP Basic auth for one of my start pages. When I attempt to close Firefox, it prompts me that I have 5 tabs already open, am I sure? I click OK and it does shutdown fine, no residual firefox.exe process, and starts up fine next time. I disabled Forecastfox extension as it seemed at one time to be related to that, but it's still occurring without it.
(In reply to comment #56) I rolled out my firefox.exe wrapper at work and, so far, it has worked well, though I should mention a side-effect that came up. When Firefox started in users' Windows XP profiles, it noticed that the default browser was not firefox1.exe, so it asked users if they want to change the default browser to Firefox. Fortunately, our users don't have write access rights to HKCR in the registry, so the answer had no effect (I should actually search for or possibly file a bug on that issue so that it won't ask users if they don't have the rights to change it anyway). The solution that I implemented to this problem was to force browser.shell.checkDefaultBrowser to false in our custom, hopefully soon-to-be-open-source, OS user profile system. It's also possible to change this in the defaults on installed computers within Program Files/Mozilla Firefox. The wrapper apparently does not work in Windows NT 4 (missing functions in NT4's version of kernel32.dll). I haven't tried it yet in Windows 2000, though I do know that it works in Windows XP and Windows Server 2003.
(In reply to comment #59) Do you really encounter this bug with every OS you tested ? Have you (the people who encounter this bug) tried to delete your profile and/or the Firefox main directory (and reinstall it of course for this last suggestion) ? If you did so, did the problem reappeared ? If it did, it would be interesting to know what extensions you use or if you changed some settings in about:config.
*** Bug 294108 has been marked as a duplicate of this bug. ***
(In reply to comment #60) It's a rare issue at work, and only two people have reported it, though it happened twice to one person. Nevertheless, some people may not report the problem to us at work and instead beleive that it was something they did, and I'd rather not put users through that pain. At work, we're using Windows XP, though I did experience it once on my Fedora Core 4test3 system (I had updated Firefox in the background on my Linux system while Firefox was open. I then browsed around until Firefox got confused and would not let me click in the window even though the keyboard worked. When I closed Firefox, the firefox-bin process was still running and thus reloading Firefox prompted me to create a new profile, so I did a `killall firefox-bin` and was able to get back in and with the new version). I'm thinking that attacking this problem from both the front-end and the back-end is the way to go, meaning that while the root causes should be fixed, Firefox should either kill itself if it doesn't shut down after the last window closes or kill other Firefox processes running as the same user within the same terminal/X11/OSX session that have the default profile locked if they are non-responsive. Plug-ins will freeze, extensions may do wacky things, and other bad things can happen. Firefox should ideally handle such conditions gracefully and not confuse the user by making them think that they "lost" their bookmarks. I also advocate for fixing the root cause(s), though this condition should be handled properly if/when it happens. When I have time, I'm thinking of implementing what I wrote in comment #54.
(In reply to comment #60) > Do you really encounter this bug with every OS you tested ? I've only seen it on WinXP, but then I have a different profile on my Linux box. I'll try copying the profile across when I get time. > Have you (the people who encounter this bug) tried to delete your profile Yes, just now, right after I had to kill firefox.exe from Task Manager. The problem went away. This seems to suggest that the problem is profile-related, but we can't go telling users "if you have this problem, delete your profile, re-install all your extensions and themes, remember all your passwords, reset all your settings...." > to know what extensions you use or if you changed some settings in about:config. I can try re-adding the extensions one at a time and see if the problem pops up again. I don't *think* I changed anything in about:config, but I may be wrong.
(In reply to comment #63) The ideal would be to backup the faulty profile (by renaming the Mozilla folder by example) to be able to do a diff between it and a fresh profile with the same extensions (and theme(s)).
(In reply to comment #64) > (In reply to comment #63) > The ideal would be to backup the faulty profile (by renaming the Mozilla folder > by example) to be able to do a diff between it and a fresh profile with the same > extensions (and theme(s)). Yep, I'm already on that (lots of difference though, trying to prune it by removing .bak files and the like). However, I happen to have installed DOM inspector in the meantime and noticed something interesting: - Start Firefox - Open DOM Inspector - Do File -> Inspect a Window + Two windows are listed -- my main Firefox window and DOM Inspector - To verify, close DOM Inspector, re-open it and repeat -- still two windows - Close DOM Inspector - Do Tools -> Extensions -> Get More Extensions (no window displayed) - Open DOM Inspector - Do File -> Inspect a Window + Four windows are listed -- my main Firefox window, Extensions, a window without a name, and DOM Inspector Can I do anything in DOM Inspector with this window without a name that will help diagnose the problem?
(In reply to comment #65) Well, I'm not familiar with the DOM inspector. The problem is that Firefox clears the cache when it is forced to close, so if the problem is cache related, it will be hard to find. But does the problem still occurs now ? The ghost window is very strange, so you can't access the mozilla update page by clicking on Get More Extensions ? Did you already try to clear all in the privacy tab from the options ? (if it fixes the problem, you can try to clear one thing each time to find where does it come) If it still occurs after that, you can zip your profile and send it to me.
(In reply to comment #66) > But does the problem still occurs now ? Yep. > The ghost window is very strange, so you can't access the mozilla update page by > clicking on Get More Extensions ? Exactly -- any attempt to open a new browser window fails. > Did you already try to clear all in the privacy tab from the options ? (if it > fixes the problem, you can try to clear one thing each time to find where does > it come) I hadn't tried this, no, but I have now. I still can't open a new window, and having attempted to do so, the firefox.exe process still lingers in the process list. > If it still occurs after that, you can zip your profile and send it to me. Send my profile and all that sensitive information to someone I've never met? Only if you promise not to disclose the pr0n sites in my browser history. :-P
(In reply to comment #67) Actually, if you clicked on clear all in the privacy, all your private infos have been erased, even the pr0n sites :-) So, it seems to be in relation with the extensions or the pref.js file ?
(In reply to comment #68) > Actually, if you clicked on clear all in the privacy, all your private infos > have been erased, even the pr0n sites :-) > So, it seems to be in relation with the extensions or the pref.js file ? Heheh, yes that would seem logical... except I didn't do it right, did I? I've just cleared all info on the privacy tab, then quit Firefox. Firefox exits normally because I haven't attempted to open a new window. I then restarted Firefox and now I can magically open windows again, e.g. Get More Extensions or File -> New Window. Close all windows and no firefox.exe process in Task Manager. So what is this telling us? That my pr0n history isn't safe after all? :-P
I hope you've made a copy of your profile before to do that. If so, you can try (using an other copy of your profile) to erase each time one thing more, close Firefox and then reopen it and check if the thing you erased (in the privacy tab) was the problem.
(In reply to comment #70) > I hope you've made a copy of your profile before to do that. I did. :) > If so, you can try (using an other copy of your profile) to erase each time one > thing more, close Firefox and then reopen it and check if the thing you erased > (in the privacy tab) was the problem. OK, now I'm seeing inconsistent behaviour. Using the same backed-up profile as previously, incrementally clearing each of the stored data areas didn't make a difference. So I decided to try the clear all trick again, and that failed too. I still can't open a new window, either by File -> New Window or by Tools -> Extensions -> Get More Extensions (well apparently I can, but it's invisible), and consequently the firefox.exe process persists. So now I'm totally confused. I restored the profile each time from the backup and just cleared one extra data store each time. I even tried not completing the HTTP Basic Auth data for the startup page that requests it, but that didn't make a difference. I'm going to try removing that page from my startup links and see if anything changes. Any other suggestions gratefully received.
(In reply to comment #71) > startup page that requests it, but that didn't make a difference. I'm going to > try removing that page from my startup links and see if anything changes. It made no difference. :(
*** Bug 295091 has been marked as a duplicate of this bug. ***
*** Bug 248866 has been marked as a duplicate of this bug. ***
I use the profile manager on start up. But because of a another bug with it I have disabled it, After that I have had no problems. Does anyone else use the profile manager?
*** Bug 298036 has been marked as a duplicate of this bug. ***
*** Bug 300506 has been marked as a duplicate of this bug. ***
100% reproducible. Close firefox window .. and there is an instance running using 104+Meg memory and using 30% cpu .. This is a serious issue ... I don't think it was there before .. just noticed it now in release 1.05 and in 1.0 Secondly .. every new instance of firefox uses 104Meg .. that major time sucks.
Whiteboard: [sg:needinfo] → [sg:needinfo] security impact in comment 4. may be several bugs.
*** Bug 296470 has been marked as a duplicate of this bug. ***
To all reporters of the problem of this bug: (A) If your problem is Java related, your problem is possibly DUP of Bug 172129 or Bug 275783. - Bug 172129 is possibly resolved by newest Java 1.5.0_04. (I couldn't reproduce problem reported to Bug 172129 comment #12 any more.) - Bug 275783 seems to be still exist even after newest Java 1.5.0_04. See both bugs, and post a comment to these bugs, if same problem. (B) If your problem is extension related, contact with the developer of the extension, and analyze your problem with him(or her), which fault - Firefox's fault or extension's fault or both, then open independent bug for the problem if Firefox's fault is involved. (C) If your problem doesn' have relation to Java nor extension, and if your problem occurs even when new profile and newest nightly/latest-trunk, please describe about your problem with evidence. Saying "My Firefox became zombie process" only is too vogus. Please exlain about when problem occurs and attach data for problem anaysis. (such as log, trace, dump, detailed operations to reproduce, JavaScript Console log, characteristics of your prefs.js if no problem when new profile, specific URL, HTTP headers by LiveHTTPHeaders, cache entry listing by about:cache, ...)
Actuly I thought it was Java too. I can't reproduce it without it.
(In reply to comment #81) > Actuly I thought it was Java too. I can't reproduce it without it. If Java related, see also Bug 300981(Hang problem when Java 1.5.0 update 4, DUP of Bug 299024).
*** Bug 308413 has been marked as a duplicate of this bug. ***
*** Bug 308537 has been marked as a duplicate of this bug. ***
Blocks: 261877
See also bug 282889, which seems to be one of the causes, unless it now works.
No longer blocks: 261877
*** Bug 261877 has been marked as a duplicate of this bug. ***
I think we should try to get this nominated to block v1.5 milestone if it's possible to get this fixed. The problem happens less often than before but it still happens for unknown reasons sometimes.
Flags: blocking1.8.1?
Alright, I've been doing quite a bit of testing on this. The bug will only work if 2 situations are satisfied (at least fore me this is true) When Java is disabled It will close properly regardless, also if the profile manager is disabled (of show this at start up is not checked) it will close properly, the only way I can reprodued it is when bot are enabled.
I'm turning this into a metabug. Dependencies with reliable steps to reproduce should probably have severity set to critical and be marked as [sg:low]. See comment 4 for why they're minor security holes. I'm also adding the qawanted keyword because it would be nice if someone would go through the comments and dups, making sure there are open bugs for any set of steps to reproduce that still works. I'm CCing bsmedberg because he has proposed changes to the XPCOM shutdown sequence (http://benjamin.smedbergs.us/blog/2005-11-14/xpcom-shutdown/) and might be interested in some of these bugs.
Depends on: 275783, 299024
Keywords: meta, privacy, qawanted
*** Bug 308491 has been marked as a duplicate of this bug. ***
*** Bug 276614 has been marked as a duplicate of this bug. ***
Depends on: 317942
The same bug appears on Linux as well when run on my FC3 box. I disabled all cache usage via about:config and via configuration dialog. I wonder why this bug has no milestone set.
Depends on: 318945
Depends on: 300999
Depends on: 316623
Depends on: 319280
*** Bug 306797 has been marked as a duplicate of this bug. ***
*** Bug 319166 has been marked as a duplicate of this bug. ***
*** Bug 317942 has been marked as a duplicate of this bug. ***
No longer depends on: 317942
*** Bug 320759 has been marked as a duplicate of this bug. ***
I just created another profile and now Firefox works fine. If I go back to the original profile, it will hang again.
*** Bug 321808 has been marked as a duplicate of this bug. ***
I am running WinXP SP2. At odd intervals, Firefox does not terminate but is not visible or accessable. This loads down the computer. The only method to stop Firefox is to use Win Task Manager and end the process. I have not found anyway to make this occur.
This issue has been going on since version 1.5 came out. Is anyone working on this? I do not know anything about browsers and have to rely on others in this area. Please fix this.
Depends on: 323026
I can confirm that this bug is configuration/profile related. I created an additional profile via "firefox -ProfileManager" and configured this profile like I did in the Firefox 1.0.7 release and now the application behaves as expected. Maybe there was an incompatible change thus I give more data about what I change usually: certificates: ask every time update: unchecked everything general/browsing: checked everything downloads: defined default directory downloads: quit download manager if complete tabs: default content: images for originating sites only cache size: 0 cookies: for originating sites only history: clear when exit passwords: never remember saved forms: never save history: 0 homepage: about:blank about:config: referer:0 adblock.nozdev.org dev branch installed Maybe this setting helps to reproduce it
I just followed the comments in #101. However, I had left the "clear private data" setting to "ask me..." When I closed Firefox, it still stayed resident. However, when I unchecked the "ask me..." setting, which would ask when I opened Firefox, the program closed as it should with no resident processes in task manager.
I just followed the comments in #101. However, I had left the "clear private data" setting to "ask me..." When I closed Firefox, it still stayed resident. However, when I unchecked the "ask me..." setting, which would ask when I opened Firefox, the program closed as it should with no resident processes in task manager.
*** Bug 323638 has been marked as a duplicate of this bug. ***
Depends on: 323672
Summary: firefox.exe doesn't always exit after closing all windows; basic auth data retained → firefox.exe doesn't always exit after closing all windows; session-specific data retained
I re-install 'FireFox.exe' to get the program to shut down. The '.exe.' file cannot install into a running program - it asks me to shut down FireFox TWICE (it makes me think that 2 programs are running), and then installs the program. It has the power to shut down the runaway program(s). Re-booting the computer cannot shut it down, but re-installation can...... Somebody knows the answer - do we know who that is?
I'm very new at this (first bug reported) and I don't completely understand how tihs process works. Is some one working on this bug? How is new information about it disseminated? Via this website or actually sent out to the email addresses? Please advise... -M
*** Bug 325834 has been marked as a duplicate of this bug. ***
Prepending '[Meta]' to summary, according to change to 'meta bug' on 2005-11-21 by Comment #89.
Summary: firefox.exe doesn't always exit after closing all windows; session-specific data retained → [Meta] firefox.exe doesn't always exit after closing all windows; session-specific data retained
*** Bug 327875 has been marked as a duplicate of this bug. ***
I've seen Firefox 1.5 (both versions) stay in memory on an XP box a few times. It never caused problems (it doesn't seem to consume any additional CPU cycles) so I only ever noticed it if I happened to go into Task Manager and wondered why Firefox still showed up even though I'd closed it. Today I installed SeaMonkey 1.0 on the same box, and at the end of the day, there were three SeaMonkey processes running even though I'd closed the browser. I wouldn't have noticed except that I thought it odd that free memory didn't increase when I shut down the browser. In both cases, Adblock Plus 0.6x (whatever the most recent version is) was installed. Java (JRE 5 update 6) installed, but I relatively infrequently visit sites which use Java. Multiple profiles are not used on this box.
*** Bug 329421 has been marked as a duplicate of this bug. ***
Depends on: 330157
Similar bug here also -- sometimes closing FF leaves a FF process running that must be closed manually via task manager (WinXP, 1.5.0.1)
It hasn't done this to me in a long while, but maybe in the past 2 or 3 days, it has again. Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.2) Gecko/20060326 Firefox/1.5.0.2 ID:2006032605
Depends on: 332651
*** Bug 332332 has been marked as a duplicate of this bug. ***
The new 1.5.0.1 has resolved the bug for me. Thank you for your efforts.
Doing it frequently now. Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.2) Gecko/20060407 Firefox/1.5.0.2 ID:2006040705 Seems mostly when in java (javascript?) sites.
Also in Win2K SP2, Ffx 1.5.0.1. Happens with both Java and non-Java sites. Email would be welcome on how to trap session information to assist.
*** Bug 333978 has been marked as a duplicate of this bug. ***
Depends on: 333978
I'm not sure if I have two bugs or if its a worse version to this bug but I am having the same problem (xp user, FF1.5.0.1).Can only open 1 FF page, very slow running, always have to close window And End Task is the only way to get back into FF. But also the bottom inch or so of the page is greyed out, if I try to alter the size of the FF page the page size changes, but the greyed out area remains. This is my first time posting on here. Is ther a fix for this yet? If so where? Thank you.
Depends on: 334943
(In reply to comment #119) > I'm not sure if I have two bugs or if its a worse version to this bug but I am > having the same problem (xp user, FF1.5.0.1).Can only open 1 FF page, very slow > running, always have to close window And End Task is the only way to get back > into FF. But also the bottom inch or so of the page is greyed out, if I try to > alter the size of the FF page the page size changes, but the greyed out area > remains. This is my first time posting on here. Is ther a fix for this yet? If > so where? Thank you. UPDATE: Tried to make a new profile on advise given, worked forst time but 2nd time firefox.exe was still running after closing all windows. On additional info to above I have uninstalled & re-installed FF but with no luck, still the same problem, still 1" grey boarder on bottom edge & still only able to open 1 FF page at a time. Any suggestions much appriciated. Thanks
*** Bug 334822 has been marked as a duplicate of this bug. ***
If the same type of memory management is being done by windows that is being done by Linux (using the glibc malloc()/free()) the problem here is *NOT* that Firefox isn't closing properly. It is that Firefox users are simply not waiting long enough for all of the firefox processes to exit. I've routinely seen Firefox take 5-10 minutes to "Quit" on an unloaded Pentium 4 running the latest release of Linux. The maximum amount of time I've seen it take is 25 minutes. A significant fraction of that time (perhaps 15-40%) may be taken *after* all of the windows have been closed. During that time one can even observe the Firefox resident memory set actually *increase* due to the extensive paging that is being done while it tries to coalesce all of the small memory fragments that are in the heap. It maxes out the memory paging capability of the O.S. due to the numerous page faults and page-ins that are required to merge all of the small memory pieces. If one waits long enough it *will* eventually exit cleanly. I suspect this may also prevent one from restarting firefox until it has completed this task and cleanly exited (the subject of other bug reports). I suspect this problem is only occuring with people who have <= 512MB of main memory on their systems. It requires *significant* effort to push Firefox's heap memory usage up into the 800+MB area which might start to cause problems on 1GB machines. It is a bug in bugzilla that they don't require people to list the memory on the machines which are having various problems. The fundamental problem here is heap memory fragmentation and/or insufficient memory self-management by Firefox of the various memory pools (i.e. overreliance on the C Library memory management and Operating System VM manager). When that is properly resolved I suspect this problem will disappear. In the meantime the fastest way to kill a firefox which has been running a long time (and has a large memory footprint) is to kill all of the processes.
I have to disagree with comment #122, at least for my particular instance. I've shut down Firefox at the end of the day and come back the following day, some 15 hours later, tried to start Firefox and it failed because the old instance was still running, often at 99% CPU utilisation.
Depends on: 337592
*** Bug 338207 has been marked as a duplicate of this bug. ***
*** Bug 338588 has been marked as a duplicate of this bug. ***
The only solution for that problem is to desinstall firefox and reinstall it.
*** Bug 300999 has been marked as a duplicate of this bug. ***
No longer depends on: 300999
It appears to be a FF Javascript problem in many cases. The following site is coded to use EITHER ActiveX (and always runs fine in IE) or Javascript (and is a disaster in FF): http://tweencam.tween-waters.com/twcam/twcam.html At least half the time it hangs FF, especially if I do anything before it's done, like even moving my mouse. When that happens, the ONLY way out is with Task Manager or reboot - the manner of attempted exit has nothing to do with it. BTW, I can't even get http://validator.w3.org/ to check it the coding is so bad!
Update to Comment #128: Site now works fine, possibly because its developers saw this thread. However, IE now fails repeatedly on it (ROTFLMAO)!!!
This batch file(for Xp users) will close and restart Firefox.
Firefox version 1.0.5.5 Windows XP, Home Edition, pathed to most recent patches available from Microsoft I didn't read the whole bug thread.. But I am experiencing a problem that may (or may not) be related. I'll let the developers decide if it's realted to this bug or start a new bug report.. Either way, I hope I can help narrow down where the problem(s) might be.. Anyhow, I've observed the following behaviour: - After browsing for some time, especially during sessions with lots of images or downloads (but not limited to such sessions), Fx becomes unresponsive to new page loads -- Fx itself responds to all commands, etc. -- but any page loads, including refreshes, seem to consistantly get stuck at 0% or around 50%. In one case, I let it run for two hours without progress. - I have an extension loaded, "Cache Status 0.6.3", which gives me the option to force clearing of the Fx RAM and Disk caches. I've noticed that if I use this extension to force the cache's (both RAM and Disk, doing either alone won't work) clear, the very next page load completes successfully, but subsequent page loads once again freeze. (Unless I clear the cahce prior to each page load.) - Any time I encounter this "freezing" problem, if I exit Fx, it remains in memory as a zombied process and will not unload unless I kill the process tree in Task Manager. I've often left the process overnight to see if it will still exit on it's own, and has never done so. It does seem (for me, at least) that when zombied, the process never uses any CPU. - On occasion (unpredictable, seems random), when any of this occurs, Fx will sometimes stop and ask if I'm sure I want to quit because it's still got downloads queued. When I cancel and check my download queue, there is nothing there. If I close again, and choose to continue exiting, sometimes Fx will exit properly, other times it will remain in memory as a zombied process. It seems there's a 50/50 chance of it exiting cleanly or becoming a zombie. - Obviously, if reloading Fx without killing the zombie, the Fx window (usually, but not always) comes back up, acting as if it was loading fresh, and sometimes (maybe 15%-20% of the time?) will work correctly -- for awhile, before the same problems occur again. If it doesn't load "fresh," the same problems occur immediately. - All these same problems occur even in a "clean install", with nothing installed but Fx. Susequently adding the "Cache Status" extension to the clean install does work as described above. - In one case, Fx had been using 125MB RAM, and in another, it was using nearly 500MB RAM. - Generally, which page/url I try to load, and anything within it (javascript, activex, etc.) has little effect, although externally accessed objects do seem to increase the chances of the problem. I've encountered the problem while loading pure HTML only pages as well as mixed. Likewise, pages with many images or files to download do seem to increase the likelyhood of the problem occuring, but I have encountered it on occasion with purely text-only pages. Again, clearing the caches always seems to "fix" the problems -- At least for the very next page load. Clearing the caches also has limited (maybe 25%?) success in Fx exiting cleanly, as well.
I'm having the same problem on my XP Pro machine. About half the time Firefox won't completely close. It may be helpful to know that this started the same time as the following: I am no longer able to install any extensions. I tried to get the latest version of "Download Them All" I click on "install now" an install box opens, counts down from 3, install button activates, I click it, another box opens and stays completely blank. No extensions will install.
I began noticing that the firefox.exe doesn't always close since I installed version 1.5
(In reply to comment #129) > Update to Comment #128: > > Site now works fine, possibly because its developers saw this > thread. However, IE now fails repeatedly on it (ROTFLMAO)!!! ooops - not true; site still fails.
Have experienced almost everything that C. Mitchell has (posted 2006-07-29 18:52 PDT) except that I do not use the extension called 'Cache Status' so cannot comment on issues related to that extension. I have experienced and/or concur with the following points he made; 1. If reloading Fx without killing the zombie, the Fx window (usually, but not always) comes back up, acting as if it was loading fresh 2. It remains in memory as a zombied process and will not unload unless I kill the process in Task Manager. I've often left the process overnight to see if it will still exit on it's own, and has never done so. It does seem that when zombied, the process never uses any CPU. 3. After browsing for some time, especially during sessions where Fx has been running for a few weeks without closure, Fx becomes unresponsive to new page loads (Fx itself responds to all commands, etc) but any page loads, including refreshes, seem to consistantly get stuck at 0% or around 50%. In one case, I let it run for a few hours without progress. What I can add to this bug is the following; Bug-specific Fx comments; * Reproducible (AFAIK) - happened three times. I usually have Fx open for a long time, ie: 2 weeks minimum and usually a month or more (only restart Fx (and PC) when M$ update requires me to) and when it has happened Fx has been constantly open for at least 3 weeks * Every time it has happened Fx has been using over 250MB of RAM * Has happened with the last 2 releases of Fx, possibly the last three but I can't be 100% sure on that * When it happened the first two times I had no extensiond installed. The third time I had installed 5 extensions so nothing to do with those * Upon 'closing' this time, Fx was using approx 265MB RAM and recovered about 75MB to the OS while 175MB remains visible in the task manager (see pic below) General Fx usage comments; * When clearing cache and then closing all but a few tabs (5-10) it is still using 150-200MB RAM which is excessive IMO Here is a screenshot of the two processes in tandem. Have included as much info as I thought relevant such as process ID, max mem usage etc. http://i5.photobucket.com/albums/y188/pbkt01/FFResidentProcess01.png Through observing the behaviour of Fx and it's memory usage (I very often check the Task Mgr and note what apps are using RAM and in what amounts - Fx often uses up to 300MB RAM) I really think that it is a RAM problem, somehow accumulating/leaking after a long period of regular and heavy (content-wise) browsing. That's my 2c worth. Hope it is of some help to the devs. Regards, Med OS: XP Pro [5.01.2600 SP2 (all updates applied except for IE, Outlook & WMP)] RAM: 1.5GB DDR400 Build Info: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7
Sorry for the double post but had to do it to try something. Killing the zombie process; 1. Didn't kill the second process 2. Returned all 175MB of RAM it was using 3. Made me feel a lot better... ;-)
(In reply to comment #135) ... > What I can add to this bug is the following; > > Bug-specific Fx comments; > * Reproducible (AFAIK) - happened three times. I usually have Fx open for a > long time, ie: 2 weeks minimum and usually a month or more (only restart Fx > (and PC) when M$ update requires me to) and when it has happened Fx has been > constantly open for at least 3 weeks 3 weeks? I can't last 3 HOURS without forcing a quit and using "Session Manager" extension to put me back where I was without the slowness. Either you have a much better system than I or more patience. I just looked at a 2.0 nightly build and it does seem a little snappier, but that was on a new (duplicate) profile also. I think not only memory, but some other things have optimized for speed on the latest 2.0 which is noticeable. Isn't there a bugzilla site with graphs of startup or rendering times somewhere? What about memory use? Thanks.
Update: It looks as if Fx is leaving **** in RAM that should be cleared. Since I posted I have since closed and reatarted Fx and been paying close attention tot the RAM usage, with 2 windows and a total of 17 tabs open I took a reading (87MB of RAM). I have the same number of tabs open now (3 days have elapsed) and, before clearing the cache for the first time just a minute ago, it was using 182MB. After clearing cache it went down to using 162. Upon clearing the cache a slightly less amount than 22Mb was returned to the total available memory shown in Task Manager. This is of course not necessarily Fx withholding/leaking RAM and not freeing it up because I have a lot of other apps running. Nevertheless it could well be that it's not doing what it should. If you look at difference in RAM usage (same number of tabs, windows and approximately the same content, ie: almost consistently surfing the same webpages for 3 days) between the time I took the reading and 3 days later and the clearing of cache, you see a 75MB discrepancy. Note: I have the disc cache set to 100MB which should REDUCE the RAM usage considerably - but it seems not to be making much difference, even when I have the disc cache set at 50MB. Of course I could be talking out of my **** because I'm not a dev but it seems to me that, after clearing the RAM cache, that there is way too much **** left over and this only gets worse. When I use Fx for 3 or 4 weeks like I mentioned in my first post, the RAM usage exceeds 300MB WITH THE SAME AMOUNT OF TABS/WINDOWS OPEN. Now that's friggin ridiculous and anyone with half a brain can see that there is a problem. Lets hope the devs are reading this and find the problem cos I'm just gonna end up restarting it every week or so to regularly free up my RAM that it's chewing through...
Depends on: 355399
Depends on: 360534
I'm getting this problem on my end quite a bit. It seemed to start happening after installing the Facebook toolbar but after removing the toolbar the problem occurs.
I'm getting this problem on my end quite a bit. It seemed to start happening after installing the Facebook toolbar but after removing the toolbar the problem occurs.
Seeing this all the time now on 2.0.0.1 and XP Professional in nightly builds since New Year's or so.
This kind of thing happens to me on XP Pro and firefox 2.0.0.1 too, only the details are different, though it's really annoying. I have dual core, and firefox.exe process that is left after shutting the broswer starts to eat away more and more of CPU until it hits its limit of 50% in Process Explorer (i.e. jams the other core). I need to manually terminate the process, and these days I check every time to make sure the process terminates. RAM usage doesn't grow with time, but on restarting it acts as if it crashed, i.e. it displays the dialog to restore the old session, which it does if you click ok.
The problem occurs many times after activated a kind of flash - Like for every 10 flash, Firefox need to manually terminated.
Whiteboard: [sg:needinfo] security impact in comment 4. may be several bugs. → security impact in comment 4. may be several bugs.
QA Contact: benjamin → general
Seeing it a lot on updates and fresh installs of Firefox latest release version. It will hang and not open anything at the install, then shows up in task manager. Killing the process and then trying to open again it starts up. Maybe this is a different bug, but it seems like it starts but then restarts after an install, so I'm putting comment here.
Firefix 2.0.0.3 doesn't quit after closing all the firefox windows under WinXP SP2. Even under task manager, the process can't be killed. Besides, after opening a single firefox window, and doing something like managing the bookmarks, firefox starts to eat all the CPU, reaching 100%.
I've seen this bug after the last update to Firefox 2.0.0.6 (Mozilla/5.0 (X11; U; Linux i686; de; rv:1.8.1.6) Gecko/2007073000 SUSE/2.0.0.6-6.1 Firefox/2.0.0.6) with Linux (!). After closing the browser window the task is still there and can't be killed via KDE task manager. It doesn't happen every time I start Fx, but I didn't found a way to reproduce this bug yet.
https://bugzilla.mozilla.org/show_bug.cgi?id=272452 and https://bugzilla.mozilla.org/show_bug.cgi?id=361060 are duplicate of this bug. I have the same problem on a fresh Windows XP install with a raw firefox 2.0.0.6 (no flash/java yet). While casually browsing pages, suddenly all browser activity stops completely. Firefox acts as if it has lost the network connection while IE and other alternative browser work normally. In task manager there is 2 instance of firefox.exe (should be only 1) and I have to end process and start firefox again to get back to browsing.
Depends on: 396339
Depends on: zombieproc
Bug seems to be solved in Mozilla/5.0 (X11; U; Linux i686; de; rv:1.8.1.8) Gecko/2007101500 SUSE/2.0.0.8-1.1 Firefox/2.0.0.8 under opensuse Linux 10.3.
Blocks: zombieproc
No longer depends on: zombieproc
I had the same issue with Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b4) Gecko/2008030714 Firefox/3.0b4. I closed the browser after modifying a preference (about:config), when restarting (with the user-manager) a message appeared saying that FF was already started (the process appeared in the task manager). I had waited for about 1 minute before I tried to restart the browser. The problem occured on a recent 3.0b4 install, but with a profile that came from 3.0b3. I only had about 5 open tabs, and I had used the browser for maybe an hour, not in an intensive way... The memory used was about 30MB (memory)/60MB (virtual memory).
I have the come across this bug as well Browser: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-GB; rv:1.8.1.13) Gecko/20080311 Firefox/2.0.0.13 Cause: Watching a streamed video from BBC iplayer ( http://www.bbc.co.uk/iplayer ). After I've finished and closed the browser, it can still be seen running a process in Windows Task Manager
I discovered this issue when my computer started running very slowly. I looked in task manager, and discovered that there were 7 different instances of firefox processes running at the same time. Once I closed all of them, my computer ran a lot faster. I'm going to go back to Internet Explorer until this problem is fixed. I can't be calling of task manager and manually ending firefox processes every time I browse the internet, but that's what I have to do now.
I had this problem 'forever' with the windows version - - now have the same problem using Firefox 2 on an iMAC, and IT'S NOT THAT EASY TO LOCATE AND SHUT DOWN PROCESSES ON A MAC. This has been going on for many years, and it's about time someone fixed it.
I downloaded & installed FireFox Beta 3.0b5, and the problem seems to have been fixed, at least for my system (Win XP SP 2). Of course, none of my add-ons work now, so I hope that that is taken care of soon. I don't suppose the add-ons developers will start putting out new versions until the final Firefox 3 is announced.
Depends on: 411258
Had the same problem with 3.0b5 and still having it with RC1. I'm running Vista X64 with FireFTP, FireBug and ColourZilla add ons.
This problem must relate to interaction with some other application on the OS, especially those that weigh heavy on CPU, e.g. resident application running. It is a bug that both happened to FF 2.0 or Flock 1.1 for that matter, when I had the COMODO 3.0 firewall installed on Windows XP SP3 (restarting the browser was impossible because another instance of the browser was still running, one is still able to activate it through task manager 'though, while it (firefox.exe) was found running hidden in the background. With another firewall installed the problem did not re-appear once. I am beta testing Minefield (Build 2008060204) very stable in these respects, only able to get it on its knees with the Crash Me add-on, as far as I can see it can be narrowed to a Mem race problem, polonus
Same problem here with my firefox 3 installation (windows vista 32-bit). Installed addons: - Greasemonkey (0.8.20080609.0) - HTML Validator (0.8.5.2) - oldbar (1.2) - Url Params (2.2.0) - Web Developer (1.1.6) Mozilla/5.0 (Windows; U; Windows NT 6.0; de; rv:1.9) Gecko/2008052906 Firefox/3.0
I have the same problem. Sometime when i close firefox 3.0.1 seems ok, but the process still running in the task manager and when i open sometime appear a message that FF is just open or the FF is open but without home page and i have to close the windows, go to the task manager and close it. I have noticed that this not happen always (I have Win XP Sp3).
I upgraded 4 computers from FF2 to FF3. This issue happens on 2 out of these 4: On random occasions (more often than 50%), Firefox 3 comes up with all of the bookmarks missing. The only way to fix this TEMPORARILY is to kill all instances of firefox.exe from Task Manager (there are always more than 1 copy under Task Manager when this happens), and then run Firefox 3 again. This issue still exists with FF3.01. I have NOT been able to figure out how to fix this. I have uninstalled Firefox, deleted the Mozilla Application Data directories (also under Local Data) as well as Mozilla Firefox under Program Files. I then reinstalled a clean FF3.01, restored ONLY the bookmarks, extensions, and saved passwords file, and I get the same results. Same as bug 284099 maybe?
I was running the Beta version of Firefox 3 with very little instance of this problem, then installed the new Firefox 3.0.1, and it happens every time I use Firefox. I'm no computer expert, but I wrote a short batch file that I run now instead of closing Firefox normally. It works with Windows XP, but there must be another way of doing the same task for other OS's. Usually, the message says that the system is closing a child PID. Anyway, the text of the batch file is this: @echo off taskkill /f /t /fi "imagename eq firefox.exe" taskkill /f /t /fi "imagename eq firefox.exe" taskkill /f /t /fi "imagename eq firefox.exe" taskkill /f /t /fi "imagename eq firefox.exe" taskkill /f /t /fi "imagename eq firefox.exe" taskkill /f /t /fi "imagename eq firefox.exe" taskkill /f /t /fi "imagename eq firefox.exe" taskkill /f /t /fi "imagename eq firefox.exe" I run taskkill this many times to make sure all instances of Firefox are closed. The /f parameter forces closing, /t kills the tree: processes and child processes. /fi filters the command to target firefox.exe. I titled the file "KillFireFox.cmd", saved it as a text file, and put a shortcut to it in my quick-launch area. Works like a charm with XP.
I get the same thing. FireFox doesn't close properly in Win XP and in Win Vista. I also get errors like this one: Fout: [Exception... "Component returned failure code: 0x80004003 (NS_ERROR_INVALID_POINTER) [nsIDOMHTMLTableSectionElement.appendChild]" nsresult: "0x80004003 (NS_ERROR_INVALID_POINTER)" location: "JS frame :: https://bugzilla.mozilla.org/js/sorttable.js :: anonymous :: line 50" data: no] Bronbestand: https://bugzilla.mozilla.org/js/sorttable.js Regel: 50 I don't know what this means but when i was trying to update to 3.0.3 this error occured. I think this relates to the problem.
Lately, I've noticed it takes a longer time to completely quit. Sometimes, if you wait a bit, then try to start it, it will start OK. I guess it was trying to start while the dying one wasn't quite finished quitting yet.
@#163: I also noticed, that firefox will close automatically after a while. But most times my firefox won´t close itself after viewing flash videos on Youtube, MyVideo etc. The videos won´t stop anymore and I have to kill the process manually. Maybe firefox has problems with Flash?
This is still a huge problem for me. I'm using version 3.0.6. I bet I have Ctrl-Alt-Delete and kill the firefox.exe process every other time I open the program. Windows XP Pro Service Pack 2 Extensions: CustomizeGoogle 0.76 Firebug 1.3.0 Forecastbar Enhanced 0.9.6 GooglePreview 3.17 IE Tab 1.5.20081203 Java Quick Start 1.0 McAfee SiteAdvisor 26.5 Microsoft .NET Framework Assistant 1.0 Open in Browser 1.4 PDF Download 2.1.0.1 Ubiquity 0.1.5 UpdateNotifiter 0.1.5.4 Themes: Chromifox 1.0 Default 3.0.6
This problem has been cropping up recently for me. It seems related to bug #474022, but I'm not a programmer by any stretch of the imagination. I have the latest version of firefox as of this date. It happens constantly, and most notably, today, I was watching hulu.com, finish watching, opened new tab (same browser + only thing running), went to abc.go.com to watch something else, I had a lag in my computer, closed abc's tab and the commercial's sound continued. closed the entire browser(hulu.com), sound still continued. browser appeared closed. ctrl+alt+delete, firefox.exe process was still sitting there. killed it, and every thing was happy again. Reopened browser, said, not shutdown properly, restore session? so, hulu's player?, abc's player?, flash?, my computer?, browser? Any thoughts?
I am having this problem as well. Have to go processes and cancel firefox. Then subsequently it tells me it's been terminated abnormally and asks about restoring tabs. Will keep system (XP) from shutting down if I don't go in and override the process to shut it down. For heaven's sake, fix this - I can see it goes back to 2004. Ken
I seem to be having a similar problem. I have Windows Vista SP1. After exiting Firefox and trying to restart it, I get a message saying that Firefox did not exit properly, that I need to kill the process or retart the computer. When I open the Task Manager I do see that there is a Firefox process still running. But when I try to kill it nothing happens (the process remains there). My only option is then to restart the computer. But (probably because of this remaining process) that takes a very long time: around 6 or 7 minutes. This is a very annoying problem. I am happy to provide more information about the system I use if that can help. Thanks !
having this same issue since the upgrade to FF3.5 (also lost the majority of my saved logins and other random data - but that is another problem) - was using RC3 prior with no issues. this happens in multiple profiles (no extensions or many). I can browse the web for 1min or several but the same problem will occur. (mem usage jumps up to 400mb, firefox window will close but still running in the background, able to launch the profilemanager and start a new profile - but nothing will appear besides another firefox.exe process)
I've had the same problem... but only when using flash (say youtube).... I have newest flash and XP SP3
Flags: blocking-firefox3.6?
This is a metabug, so it doesn't make a lot of sense for it to block a release. Instead, make sure there's a bug report with clear steps to reproduce in this bug's dependency list. If you want, you can nominate one or more of those bugs to block a release.
Flags: blocking-firefox3.6?
I have been encoutereing this problem for a while on Vista. If I open and close Firefox repeatedly within a short period of time, some times when I try and open it again it doesnt open it just says firefox is already running and then I have to go to task manager to kill the process first and then open it again before I can se he browser.
It doesn't matter if the FF is opened for short or long time. Recently (I'm using FF 3.5.2) it happens almost all the time when I close the application, the process remains alive and must be killed using the Task Manager.
Firefox sometimes starts with just a blue bar or in minimized mode, with another window acting normally. Sort of a junk window in addition to the regular one. Hope this helps somehow.
growing zombie condition seems to follow a script failure on a page. taskmgr will show the zombie process growing in resources. I use version 3.5.4
Firefox.exe does not closes after closing all windows. It uses too much CPU resources and slows down my computer. I am using windows vista home basic. I found this problem on two computers.
Depends on: 507903
I get this occasionally too even on xp machines but on the visa machine i found it on i put together a script that killed firefox and reopened it but the cmd dialog was messy so i made a vbs that ran the cmd invisibly i think that it could be integrated into the original programming that a small program will invisibly open check if firefox is running, if it is running then check wether its responding or not, if not kill firefox and reopen, if responding simply open a new window
(In reply to comment #20) > To summarize, there are two cases here (after the last window is closed and > firefox.exe process still shows up in the Task Manager): > 1. Any attempt to start the browser silently fails. Firefox appears dead. > 2. Any attempt to start the browser succeeds, and all previous session > information is still available. > > #1 is similar to a standard program crash, and has no easily exploitable > security repercussions. (That's how Acrobat Reader plugin behaves on my system > all the time.) If not for profile locking, users would not be noticing the > problem. The "dead" process is waisting RAM until it's shut down. So, it's > mostly an annoyance rather than a security issue. > > #2 is the real problem, IMO. Somehow, browser keeps a hidden window open. It > happened to me a few times in the past. The problem was usually cured by > re-opening browser and doing File | Exit. > > It's my impression, that I get #2 when a plugin fails to load, or FF doesn't > know how to deal with some fancy/malformed markup, or a popup failed to show up. This is happening to me every time, first it was with ff 3.6 beta, then changed to 3.5.5, and after either 3.5.6 or 3.5.7 update this came to the non-beta too
Hi there! I have been a loyal and dedicated Firefox user ever since the split from Netscape. I have used it on many computers and had few problems. Since the recent update to Version 3.5 however, (This version Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9.1.7) Gecko/20091221 Firefox/3.5.7 (.NET CLR 3.5.30729)) I have had a problem. On this laptop, when I end a Firefox session, the main window disappears but the process seems to hang as it’s winding up. The computer becomes sluggish and unresponsive. Fire up TaskManager and I find processor duty cycle at 100%. Memory usage is usually 100 ~ 400 MB higher than I’d expect. A check of the Applications tab shows nothing, so I go to the Processes tab and order by memory or processor usage. And guess what? Taking up 98% processor and several hundred MB of ram, is firefox.exe. Select it and ‘end process’ generates a popup box ‘Do you know what you’re doing?’ and then it exits as neatly as any other process. The question is why does it freeze on shutdown and hang around in the background taking up all my resources? The problem can’t be that bad because when it gets a shutdown order from XP it does so in a quick and orderly manner – unlike a program that has had a serious problem like a page fault or a general protection fault. The sequence of events does not generate a dump file or a crash report or any other sign that something bad happened. The event viewer (application) log clearly shows several other programs hanging over the last few days but no mention of firefox at all. I’m not deeply troubled or concerned about any of this but it is inconvenient and irritating. I’d love to know what’s causing it. Yours respectfully, Mike Thompson.
I had this problem last year. It was caused by one of the "add-ons" I was using. Wish I could tell you which one. Anyway, I experimented with disabling them and my problem has not reoccurred. Ken
For me, this problem began on multiple XP machines with 3.6. The most annoying aspect of the problem is that it takes so bloody long to send back a "firefox is already running and you must close it before restart" message -- as long as a normal startup! It seems to me that this should be a trivial fix and check; if it's that important to ensure that firefox.exe isn't already running, shouldn't the error message pop up immediately as a result of one of the first "is this computer ready?" checks in the startup sequence? I have confirmed that this is a core-system problem by doing a fresh install of 3.6 on a fully-patched XP/SP3 machine with NO add-ons. However... if there is, in fact, an interference with another component, my guess would be either the Acrobat or the Java platform, as each of them fails to respect the "disable" option when those arrogant SOBs use their closed update system... and the problem seems to have been worse since the last updates to Acrobat Reader and Java.
(In reply to comment #182) > For me, this problem began on multiple XP machines with 3.6. The most annoying > aspect of the problem is that it takes so bloody long to send back a "firefox > is already running and you must close it before restart" message -- as long as > a normal startup! It seems to me that this should be a trivial fix and check; > if it's that important to ensure that firefox.exe isn't already running, > shouldn't the error message pop up immediately as a result of one of the first > "is this computer ready?" checks in the startup sequence? > > I have confirmed that this is a core-system problem by doing a fresh install of > 3.6 on a fully-patched XP/SP3 machine with NO add-ons. However... if there is, > in fact, an interference with another component, my guess would be either the > Acrobat or the Java platform, as each of them fails to respect the "disable" > option when those arrogant SOBs use their closed update system... and the > problem seems to have been worse since the last updates to Acrobat Reader and > Java. Interestingly enough, this problem seems to have disappeared for me starting with Firefox 3.6 although I suppose it could have something to do with the fact that I am now using Windows 7 64-bit instead of XP.
I am having this problem in Windows 7 (32bit) after installing Firefox 3.6 [Never occurred in earlier versions of Firefox.] The process takes in excess of 30 seconds to close after shut-down every time. Occasionally, the process does not quit, and must be terminated. No error messages until I try to re-start Firefox. Add-ons: Adblock Plus Better Gmail 2 Better Privacy CoolPreviews Download Statusbar DownThemAll! Forecastbar Enhanced Ghostery Google Toolbar fo rFirefox InvisibleHand TACO Xmarks The new Add-ons since the installation of Firefox 3.6 are: CoolPreviews TACO Better Privacy Ghostery (IIRC)
Does the wrapper designed to terminate the program without windows stop executing if a window is present and responding? (In reply to comment #54) > Created an attachment (id=182186) [details] > Wrapper for firefox.exe that terminates old firefox.exe processes that did not > cleanly shut down.c > > I wrote a small Win32 replacement for firefox.exe that calls the original > firefox.exe (renamed to firefox1.exe) after ensuring that all running > firefox1.exe processes have at least one visible window. Any firefox1.exe > processes that have no visible windows are forcibly terminated. > > Of course, this is really a hack and not a proposed fix for this issue. I'm > thinking that perhaps the fix mentioned in bug #214675 attachment #148991 [details] [diff] [review] can > be amended to include a variant of the file I'm attaching instead of > ShowOSAlert in that patch at: > > +#ifndef MOZ_ENABLE_PROFILE_MANAGER > + ShowOSAlert("No profile specified. Please start with -P <profile name>\n"); > + return NS_ERROR_FAILURE; > +#else > return ShowProfileManager(profileSvc, aNative); > +#endif > > Just a thought.. I may have time to experiment with it a home and, if so, I'll > attach a patch to obsolete this separate program. > > Here is what I originally wrote in bug #214675 when I initially posted this > program to that bug by mistake: > ============================================================ > We, like many people out there, occasionally have users who "lose" their > bookmarks and start page due to something causing Firefox to uncleanly shut > down and thus keep the profile locked. I'm not going to tell our end-users to > go into the Task Manager to end the process (think "Auntie Mabel" or "Grandma" > type of users), so I wrote a small tool that does that for them. > > I'm hoping that perhaps a part of this can go into the real firefox.exe so that > > it can kill any previous instances that have no windows open and thus prevent > the Profile Manager from appearing (at least, that's one idea). > > I understand that fixing the root cause of freezing problems rather than > working around them is the best approach, but this workaround can handle cases > of rogue plug-ins that I suspect are causing Firefox to hang rather than > Firefox itself. > > BTW, thanks for an excellent web browser! The IT Department of a non-profit > that I am in, with my help, rolled out Firefox to all of our 100 or so > computers as a part of our Windows XP migration. I'll send more details to the > appropriate place if I find out what caused one particular user's Firefox to > "lose" his bookmarks on two occasions. For now, I've just manually edited > profiles.ini to switch his default back to his original to "restore" his > bookmarks and make him happy, though I'm hoping that this workaround will > prevent this from happening on our agency's computers.
@Bobby: Ahh, you've found the workaround in this long thread :-). The wrapper looks at all Firefox processes by looking for firefox1.exe and kills all that do not have at least one visible window. If it finds one with a visible window, it will keep on going. All properly working firefox1.exe processes should have at least one visible window (it could be minimized, and that's still considered visible), though that might not be the case if Firefox is starting up. A user who gets impatient with the Firefox startup process on a very slow computer could try to open Firefox multiple times, and that would basically cause the wrapper to kill the currently loading firefox1.exe and then start a new one, which is not ideal. That, however, is still a generally better user experience than a lingering Firefox process and possibly "losing" one's bookmarks on the next Firefox process launch (users will just click through whatever on the profile manager will get them into Firefox without thinking about the consequences). The wrapper does not check to make sure that the window response to messages. It won't kill an unresponsive Firefox window. It only checks for window visibility. An unresponsive window could just be executing a long-running JavaScript, for example, and killing that wouldn't be nice, so the wrapper doesn't.
Would it be feasible to implement something that would identify a window as unresponsive, loading, working, idle, or non-existent? Such as a marker byte that would be the identifier of such. I understand that the firefox executable knows that it is open but not responding but reacts in an unfavorable way. This marker byte would be used to determine status and once the status is determined a decision is made based on it (kill for unresponsive/non-existent, new window for idle, a external program saying "please wait" that closes once firefox is loaded) or something of the such. (In reply to comment #186) > @Bobby: Ahh, you've found the workaround in this long thread :-). The wrapper > looks at all Firefox processes by looking for firefox1.exe and kills all that > do not have at least one visible window. If it finds one with a visible window, > it will keep on going. All properly working firefox1.exe processes should have > at least one visible window (it could be minimized, and that's still considered > visible), though that might not be the case if Firefox is starting up. A user > who gets impatient with the Firefox startup process on a very slow computer > could try to open Firefox multiple times, and that would basically cause the > wrapper to kill the currently loading firefox1.exe and then start a new one, > which is not ideal. That, however, is still a generally better user experience > than a lingering Firefox process and possibly "losing" one's bookmarks on the > next Firefox process launch (users will just click through whatever on the > profile manager will get them into Firefox without thinking about the > consequences). > > The wrapper does not check to make sure that the window response to messages. > It won't kill an unresponsive Firefox window. It only checks for window > visibility. An unresponsive window could just be executing a long-running > JavaScript, for example, and killing that wouldn't be nice, so the wrapper > doesn't.
Depends on: 479373
No longer depends on: 479373
Keywords: qawanted
Well an additional in-browser taskmanager for Fx would be appreciated, GoogleChrome has something like this - press shift & escape to bring it up, there one could see what process is causes unresponsiveniss or causes hanging and the user then could close the appropriate process that is consuming all the CPU or mem, and go on with the rest of the browser fully functioning. With this already entered in the Fx 3.7 version, something similar could be performed and I think will be greatly appreciated, could coders and nightly builders have a go at this? polonus
I took all these ideals and tried them all...none worked. I even tried root kits, nothing worked. I downloaded a disk cleaner to completely clear my hard drive and restored my system to its factory state. I am not having this problem now but I don't think the issue has been resolved....I will bet this is still happening to other Firefox users. And please if I have any more problems I hope I don't get solutions from 2004!!!!!!! I don't want GoogleChrome....they store personal information about you for a very long time.
Since upgrading to 3.6.3, Firefox does not stop running when application is closed--ranging from every time to frequently. The FF process is running in Task Manager and must be shut down before Firefox will relaunch. Prior to upgrade, Firefox closed normally (except for crashing bug). Have been at this site two weeks ago and did not report same bug. Am worried about how much time has elapsed without the bug being fixed. (Windows XP Pro SP3)
Sorry to abuse this bug to contact everyone that has been reporting slow shutdown times; are you still seeing this with Firefox 4 or newer? Trying to gauge whether we need to start looking at this (and at least measure it in the wild) for the upcoming releases. Bug 546366 has a theory on this, too.
Steps to reproduce: 1. log into your gmail account in firefox. 2. close the browser 3. open the browser 4. your gmail account is already signed into. This is a security problem, because you may be at a guest computer, think that you signed out, but anyone who opens the browser can access your gmail account.
trttrgtrg
I'm seeing this on my laptop on all the latest releases. At least once a day firefox becomes very slow. Using Firefox -> Exit menu all the windows are closed but firefox.exe remains in memory (and usually consumes 50% CPU) until I kill it.
I'm having the same problem as dvirlahat@gmail.com. It consumes exactly 50% CPU and I have to manually terminate. It happened a few times with the application normally installed, but happens more often when I use a portable version of Firefox.
Well, I must add myself to this list. I can't believe people have been seeing this exact error for as long as they have. I only just started noticing it in the latest version 11.0. The scenario I am seeing is as follows: 1) Open Firefox i. Doesn't seem to matter how many tabs/pages I have open, what pages I am browsing to, or how long I have had FF open. 2) I close or restart Firefox i. On a restart it just seems to sit there and I usually try to relaunch Firefox ii. I have given Firefox plenty of time to close, sometimes even an hour as I am working on other projects. 3) I relaunch Firefox and get the error: "Firefox is already running etc..." To correct this I have to open task manager and kill Firefox.exe, then I am able to start it up. I am able to reproduce this easily, as it happens 99% of the times I close Firefox. I suspect that it has something to do with the sync process, with the addition off add-on sync, but have not been able to confirm that yet. If anyone knows of a way for me to collect more data, or find logs to view, please let me know and I will gather anything you need.
(In reply to Christopher S. Bates from comment #197) > Well, I must add myself to this list. I can't believe people have been > seeing this exact error for as long as they have. I only just started > noticing it in the latest version 11.0. > > The scenario I am seeing is as follows: > 1) Open Firefox > i. Doesn't seem to matter how many tabs/pages I have open, what pages I > am > browsing to, or how long I have had FF open. > > 2) I close or restart Firefox > i. On a restart it just seems to sit there and I usually try to relaunch > Firefox > ii. I have given Firefox plenty of time to close, sometimes even an hour > as I > am working on other projects. > > 3) I relaunch Firefox and get the error: "Firefox is already running etc..." > > To correct this I have to open task manager and kill Firefox.exe, then I am > able to start it up. I am able to reproduce this easily, as it happens 99% > of the times I close Firefox. > > I suspect that it has something to do with the sync process, with the > addition off add-on sync, but have not been able to confirm that yet. If > anyone knows of a way for me to collect more data, or find logs to view, > please let me know and I will gather anything you need. - Note: this is happening on Windows 7 x86_64 Build 7601
I was having this exact same problem; but I have resolved it. Click on the Firefox button at the top left of your window. Click addons. Disable ALL of the extensions. Close Firefox (you'll have to end process again) To test if fixed: Open Firefox Browse til hearts content Close Firefox Attempt to reopen Firefox If this hasn't fixed the problem: Disable ALL of the plugins that are not up-to-date or are unrecognized. Retest... If this still hasn't fixed the issue, keep troubleshooting because mine is fixed. I isolated my problem to the extention: Skype Click to Call 5.11.0.9874 To isolate yours (assuming you've already fixed the problem): Open Firefox Enable a single extention or addon Close Firefox Open Firefox Close Firefox Open Firefox (If you get an error, you've found the problem ext or add) If successful repeat above steps for each add/ext until you get an error
Guys please calm down. This is known, yes you can reproduce it on any version of firefox and any os. Why? Because right now it is the expected behavior - firefox closes the window first and then proceeds to shut down the process. It's done this way to give a sense of "speed". Like say, if you were to shut down windows and it would first show a black screen - no image and only the shut down the PC. And yeah it's obvious that if you have more add-ons and plugins firefox will take more time to close. With less addon firefox will close before you can relaunch it. Anyway as i said this is how it is expected to work - they know about it. They are currently fixing bugs to make firefox close faster so, be patient and don't spam bugs with facts, that are already known - it only slows the developers down.
This has started happening again with version 14.0.1.
@ PROPHET: It might be reproduceable on any version but it doesn't happen for the average user on every version. When it starts to become prevalent on a particular release then people have a right to post and complain. And this has been happening for the last 8 years so "please calm down" ain't gonna calm anyone down so to speak so maybe save your breath there matey And I can confirm this is happening again with 14.0.1
(In reply to Jesse Ruderman from comment #1) > If Firefox sticks around after you close all windows, that's a bug. Does > Firefox always stick around after you close the last window or have you only > seen it happen once? I've been experiencing this same issue. It never happened before. Started happening the moment I installed FF ver. 25. -- This is DEFINITELY a bug. Quite annoying to have to sit with my task manager open to ensure that FF closes completely every time. I'm not a coder / programmer. I'm a computer builder and trouble-shooter.. so I have no grasp whatsoever on debugging. All I can say is that the bug exists, and it needs to be remedied. Thanks!
Almost forgot to mention -- The problem seems to be somewhat "random", but definitely repeatable. Unfortunately due to said randomness of the bug, there really is no discernible pattern to this problem. I first experienced the issue when using it with Trillian. Initially, I thought that trillian was somehow giving the newly-upgraded FF troubles. I just got to work and updated FF here, too. Not 5 minutes into my day with FF 25, I experienced the same random bug.
I'm not sure this is the right place for this discussion, seeing as this is a meta-bug (it might be better if you filed a bug report specifically for Firefox 25 and set it as blocking this), but could you try the utility from https://ftp.mozilla.org/pub/mozilla.org/utilities/crashfirefox-intentionally/ to crash the Firefox process intentionally and submit a crash report? (which you can then find in about:crashes) Might help to get a stack and figure out what it's doing.
(In reply to Keith Duncan from comment #0) > User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.6) > Gecko/20040206 Firefox/0.8 > Build Identifier: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.6) > Gecko/20040206 Firefox/0.8 > > Hidden window can store username/password called via .htaccess > Exiting the application apparently completely does not "reset" the > application. > Upon further examination (CTR+ALT+DEL) it appeared there was a "hidden" > Firefox > window/component running. > Suspect this may be an intermittent fault - difficult to debug :-( > Problem with the OS or a problem with Firefox? > Perhaps future releases could include some sort of check on exit/shutdown? > > > > Reproducible: Didn't try > Steps to Reproduce: > 1. > 2. > 3. > > > > > Sincere apologies if this is already known about - somehow I suspect not. > To the end user it *appears* as if everything is ok... he/she closes the > application - normally the system would then "forget" all about > usernames/password. > > However with a hidden window running in the background (difficult to > replicate) > it perhaps leads the user into a false sense of security. > > I discovered this problem when I was setting up a secure directory... I > couldnt > understand what was going on.. I checked the server directory > permissions..etc.. > One directory was asking me for userID - whilst the other was letting me in > no > problem. > > It was only when I looked at background processes/applications I saw a > suspicious "Firefox" entry - I closed this down and checked the website and > sure > enough htaccess was performing as I would expect. > > For the stuff I am developing it is maybe not so important - but in some > instances security could be compromised - the only real workaround I can > think > of in the meantime is to reset your PC after entering "secure" data ;-) I seen it happen in former versions as well - I happens every time I double click a html-file, but only if FF is not already running
This happens in Linux constantly. This is Firefox 32 that has lots of new features and 10.5 year old bugs. I kill -11 everytime.
This may be another issue entirely, or it may be related. I'm unsure. I always have "Show my windows and tabs from last time" selected. I'm using Firefox 34.0 now. However, before I upgraded from version 29.0, each time I would close a browser session, my tabs for .htaccess password-protected directory documents would always re-open in the next subsequently launched browser session as blank white pages with only the .htaccess login dialog box showing. The behavior seemed to be by design. Now, after upgrading to Firefox 34.0, the .htaccess password-protected directory document tabs always open showing all of the data from the password-protected directory documents which were visible on those open tabs prior to my closing the browser session. This seems like a security issue. But if it's unrelated to the login credentials living on in a persistently running instance of firefox.exe, then I apologize for commenting about it here.

Given the lack of new dupes or comments the past 7 years, and the lack of open dependent bugs, let's close this out.

Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: