browser crash bugs

[needshaircut]

The NoScript motto is “less traps more web”.

What about browser crash bugs?

Something is crashing Firefox quite badly here every few days lately and I cannot figure out what it is.

Here is a link to a page of crash bugs, some work against Firefox and some don’t.
http://www.computerbytesman.com/securit ... /index.htm
Can anything be done by NoScript about those that do crash Firefox?

[Alan Baxter]

@needshaircut:
Those tests date from the old IE4 and Netscape days. Very old news. Go ahead and try each of those tests with NoScript installed without allowing computerbytesman.com. Please report back which tests crash Firefox with NoScript.

[therube]

Though if common.js were coded within the (index.htm) page itself, then would JavaScript references be necessary? If not, then assuming the browser would be "vulnerable" to the "exploits" therein, then NoScript would not provide protections.

(I guess what I'm saying is, take the same code, but write it in a way that does not require JavaScript.)

[needshaircut]

@ Alan Baxter Old news or not, if they crash Firefox, they crash Firefox. I actually need some JavaScripted content, as most people do, so while disallowing computerbytesman.com may stop those tests from running, it isn’t a useable solution to the problems those tests represent.

In any event, when I next experience a nasty real world Firefox crash, I’ll try to report back. Usually happens directly after a page reloads due to letting NoScript allow some content to run.

[Alan Baxter]

disallowing computerbytesman.com may stop those tests from running, it isn’t a useable solution to the problems those tests represent.

If you want to allow JavaScript on a malicious site, then be my guest. NoScript isn't designed to protect you from malicious scripts hosted by a malicious site. So what's your point? Do you completely misunderstand what NoScript means?

These tests were problematic for Netscape and IE4 on Windows 98 for goodness sake! Run the tests yourself right now, with or without Javascript enabled, as you please. I bet none of them crash Firefox, let alone your computer. Do the tests yourself and report back, if you're so concerned.

In any event, when I next experience a nasty real world Firefox crash, I’ll try to report back. Usually happens directly after a page reloads due to letting NoScript allow some content to run.

Please do, with details about the web site, exactly what you Allowed, and what your other extensions are. Firefox crashes are usually due to a bad extension or third-party plugin. In any event, breakpad automatically reports the crash details to the Mozilla developers. That way they can see if there's a Firefox bug or vulnerability that needs to be fixed. Firefox crashes aren't caused by these nine year old tests you've dug up.

[GµårÐïåñ]

"SOMETHING" has been crashing Fx lately and I am suffering from it too, I mean I can hardly get anything done. I have tried everything and the Fx support people are their usual incompetent selves and won't address anything since they want to push the 3.6 and I have come to use Chrome and IE8 for goodness sake so I can get work done. I can't rely on a browser that crashes every few pages. This happens with NO addon installed, increases in frequency with NS installed and becomes intolerable if you have NS and ABP installed. I have dug up alot of references to it but no response AT ALL from the Fx team, so I guess they are just pulling an ostrich burying the head in the sand and ignoring it but that is hardly an NS issue, even if NS makes it a worse, its a core function issue. The fact is that NS protects against most if not all crash/web bugs as long as the page doesn't have scripting enabled, so what more can one expect? Maybe pressure should be put on the Fx team to improve their product or just dump it, because pushing crap for free isn't doing anyone any good.

[needshaircut]

NoScript isn't designed to protect you from malicious scripts hosted by a malicious site. So what's your point? Do you completely misunderstand what NoScript means?

. . . In any event, breakpad automatically reports the crash details to the Mozilla developers. That way they can see if there's a Firefox bug or vulnerability that needs to be fixed.

If only I knew which sites and which scripts were dangerous in advance, I wouldn't visit or allow them, of course. NoScript does have functionality that goes beyond 'on/off' for scripting. So, I figured scripted crash bugs would be worth pointing out.

When Firefox stops responding and I try to restart it, I do get a small popup about the fact that it has crashed and that it is going to report back to Mozilla if I allow it to. I allow that, but then I get a message that it is unable to report back. This is typically followed by windows error messages about windows failing to save some temporary file to hard drive 2. (The computer is a laptop, with one hard drive, as far as I know.) Rebooting fixes Firefox for awhile, and sometimes causes chkdsk to run. When chkdsk runs, it usually finds some browsing session related content file that is screwed-up.

Again, even if NoScript is not responsible, this mess starts every few days, when I give NoScript permission to run scripts, which I have to do everyday to get content I need.

[therube]

"SOMETHING" has been crashing Fx lately

Repeatable? URLs where this happens?
Does it happen in Safe Mode? With a new Profile?

Run a CHKDSK & your drive manufacturers diagnostic.
Run a memory diagnostic.

[Alan Baxter]

So, I figured scripted crash bugs would be worth pointing out.

Thank you for doing so. My point was that the issues on the test site you linked were addressed years ago with browser and OS improvements. Also, I had checked all of them out previously and was irritated that you apparently hadn't. Sorry I took my irritation out on you, especially when you're having computer problems.

I see therube has already responded to your problems. Please answer his questions and follow his advice.

[needshaircut]

@therube When it happens again, I’ll report back if it is repeatable, URLs, etc. If it isn’t repeatable, etc, I’ll move on to diagnostics, log analysis, etc.

@Alan Baxter, at least some of those ‘old’ javascripted crash bugs are still effective here, which is why I pointed them out. If none of the nine tests force you to reboot and/or crash Firefox for you, congratulations. While I regret that I didn’t document specifically which of the nine were effective against Firefox currently, I do not plan on running them again at present because I do not want to take a chance at muddying the water with demos when I am trying to track down a real world browser crash issue.

[Alan Baxter]

While I regret that I didn’t document specifically which of the nine were effective against Firefox currently, I do not plan on running them again at present because I do not want to take a chance at muddying the water with demos when I am trying to track down a real world browser crash issue.

I don't blame you a bit! That sounds very prudent. I doubt any of these old crash tests are the cause of your serious computer problems.

[GµårÐïåñ]

"SOMETHING" has been crashing Fx lately

Repeatable? URLs where this happens?
Does it happen in Safe Mode? With a new Profile?

Run a CHKDSK & your drive manufacturers diagnostic.
Run a memory diagnostic.

Happens all the time, at unpredictable intervals and on any page, no specific page. Sometimes opening the options will crash it or going to about:config so nothing that can be pinpointed or URLs. Yes it happens in safe mode as well and basically the 3.5.2 update has been nothing but crap. They rushed to push out something and instead of fixing problems it has become the problem.

Its not my drive and its not my memory. Although Fx would love to think its the only thing that runs on a machine, everything else is fine, just Fx is broken, so I am confident in saying the problem is Fx not anything else. It has been reported by alot of people but all the mods at Mozilla keep doing is closing the threads and ignoring it, no one and I mean NO ONE has even acknowledged the problem or done anything to address it. One mod was actually stupid enough to admit on the record that they want to push 3.6 and so they are not going to support any existing issues, that's just crap.

<note: it crashed TWICE while posting this reply.>

[therube]

SeaMonkey or SeaMonkey 2.0 Beta 1

[GµårÐïåñ]

I would try it but the fact is that I don't need all the stuff that is shoved in SeaMonkey, I just want a browser, I have email client and calendar and will never use it for that. Plus I figure if they can't get Fx with just a browser right, I can't trust any project that is trying to shove even more on the buggy Mozilla framework. I am seriously considering ditching Fx for Chrome (assuming they get their head out of their rear and improve it a bit) or god forbid even IE8 is an improvement over Fx (but that's last resort) but in the meantime the only worthy browser seems to be Opera frankly, maybe charging for their product was the right way to go, they take more care to make sure it works right since they stand to lose something. Honestly the only thing I will miss not using Fx is NoScript or I would have ditched it starting with release of crap v3.0 but held out for the promised perfection of 3.5 only to find its worse than 3.x and now they are pushing 3.6 and putting support on the back burner. Sad and pathetic really.

[needshaircut]

Bringing this thread back to topic. . . There have been lots of crashes and hangs under Firefox 3.5.x. If any of the bugs, conflicts, and vulnerabilities that cause this have characteristics which NoScript could recognize and as a result counter, that would be nice.

Some scripted content needs to be allowed to run in order to get needed information and therefore it isn’t always possible to forbid in advance all troublesome scripted content. To the extent possible, in addition to protection against XSS, CSRF and ClickJacking, what about putting a stop to at least some subset of content which causes browser hangs and crashes?