Steps to reproduce: 1. Enable quicklaunch from the advanced section of preferences. 2. Go to http://www.hypererotica.com/easytour.html. 3. Select "quit" from the file menu. Actual result: the hypererotica site closes, but a pop-up advertisement takes its place. Expected result: browser exits. This is pretty important for when my parents or girlfriend walks into my room. (Is there any other reason for the exit command to exist?)

note that there is a bug to disable loading of popups from onunload handlers

bz: bug 33448 would fix this bug in most cases, but might not fix the case where as site does onunload="otherwindow.setTimeout(makeNewPopup, 0)". I'm not sure what the solution is -- disable all untrusted javascript when the user hits Ctrl+Q?

From bug 65121: ------- Additional Comments From Joseph Elwell 2001-04-25 17:47 ------- Netscape has long been the browser of choice for porn surfers. The one reason for this is when the popups become unbearable and numerous, "File... Quit" will end it *all*.

Confirming bug. (Although I thought this was already reported, but couldn't find it anywhere.) Netscape has ALWAYS acted in a way that "Exit" would immediately kill the entire browser, and that NOTHING would operate after its use. When we added turbo, we lost this functionality. Now "Exit" doesn't really exit, it just closes the window and has an imaginary instance running in the background. Removing Exit isn't a choice, because Exit destroys all windows and anything running within them. What needs to happen is that on File | Exit or on File | Close of the last window, ALL Javascript is killed, ALL applets are killed, Everything is killed (including mail biff unless we get a pref to check mail while iconified.) Effectively, mozilla in turbo only mode should not act as if it had Any windows open.

Yeah, Quit/Exit should close all windows and end all processing, even when QuickLaunch keeps some of the program loaded. ->blaker

Cc'ing jst for ideas on fixing this. Basically we need to stop processing scripts/firing pending timers.

What if a script calls a function to save data in a cookie in its onunload function? It would be nice if we could avoid breaking that kind of script.

Didn't danm fix this already?

Yup. This works for me using a recent build. To answer Jesse's comment 7: no problem, unless you have the "warn me before storing a cookie" option active. In that case, no alert for you, and no cookie, I imagine. If only there were a flag to indicate you were trying to open a window for legitimate purposes, and if only we could keep people from misusing it... Bah. I said "exit," damnit. Not "second guess me." So exit. I like the new face of Mozilla. *** This bug has been marked as a duplicate of 33448 ***