Hacker News Reader

Two of the main advantages of Firefox for many are its memory usage and its extensibility. Those two advantages would be lost if Firefox added sandboxing. I'm glad that I can choose between sandboxing and better extensions and I would be sad to see that choice disappear.

>Two of the main advantages of Firefox for many are its memory usage and its extensibility. Those two advantages would be lost if Firefox added sandboxing. I'm glad that I can choose between sandboxing and better extensions and I would be sad to see that choice disappear.

Totally orthogonal issues.

Not to mention Firefox ate memory like a pig until recently, despite not having sandboxing.

Now it's somewhat better with post 16 stuff, but still it's slow. Chrome is much more responsive and fast.

>Totally orthogonal issues.

http://blog.ffextensionguru.com/2011/11/20/electrolysis-proj... Add-on were one of the major issues blocking Multi-process Firefox.

>Not to mention Firefox ate memory like a pig until recently, despite not having sandboxing.

Agreed. Sandboxed Chrome was like a gift from the IT gods back in the bloated FF3 days.

>Now it's somewhat better with post 16 stuff, but still it's slow. Chrome is much more responsive and fast.

Responsive and fast until it starts thrashing when your physical RAM gets all used up.

>Add-on were one of the major issues blocking Multi-process Firefox.

Well, they will need adaptation to communicate cross pages, but it's not like you can't have sandboxing and plugins (a la Chrome/Safari).

>Responsive and fast until it starts thrashing when your physical RAM gets all used up.

Haven't really seen that. But I rarely have more than 20 tabs open (usually around 10-15). OS X, and using Canary.

>Now it's somewhat better with post 16 stuff, but still it's slow. Chrome is much more responsive and fast.

This has not been my experience. Chrome used to be faster, but for the past year, Firefox's responsiveness have blown Chrome's out of the water for me personally on the PCs I use.

> Totally orthogonal issues.

They're really not. Many add-ons would function without issues with sandboxing (once updated), but there are many cases where requiring IPC between tabs/windows would be prohibitive (particularly for individualized add-ons).

Two of the main advantages of Firefox for many are its memory usage

Really? I've heard that said before, but right now Firefox is grabbing almost half a gigabyte of my RAM just to have about a dozen tabs open on simple forum sites like HN. Some of the extensions are significant offenders at the moment, particularly things like the AdBlock ones, but Firefox itself is still greedy even with no pages open and running in safe mode with all the plug-ins disabled, and of course Firefox without any of the major plug-ins is much less useful as a default browser.

Please file a bug with the contents of about:memory if you find situations in which Firefox uses much more memory than alternative browsers; Firefox is very serious about memory usage these days. You can cc me on the bug if you like to make sure it gets to the right people.

Thanks for the tip; I hadn't heard of about:memory before.

Is there a sane way to report bugs these days? I'm happy to do things like describing a useful test case or attaching a screenshot of about:memory, but I've spent too many half-hours trying and ultimately failed to report useful information via Bugzilla to ever go near it again. It's like a case study of how to do every possible thing wrong in usability terms. :-(

I know the interface got upgraded recently, and I've heard good things about the new upgrade—although I don't actually work with Bugzilla on a daily basis, so I don't have first hand experience. Anyhow, you might want to try the new version and see whether it works better for you.

OK, if I run into a reproducible version of the problem I'll see if I can file something helpful.

Reset Firefox fixes many problems like this -- see https://support.mozilla.org/en-US/kb/reset-firefox-easily-fi...

> The second issue is the lack of tab sand-boxing. A fairly large part of my job is testing for edge cases in our products. Our products are JavaScript heavy. I often find issues which crash the browser... well.. they crash Firefox. In Chrome the tab shows an error in the console and I have the option to close the tab. In Firefox I never get an error in the console. The whole browser freezes and stops responding.

Why not simply run two FF instances in parallel (e.g., via the profile manager) - one for actual browsing, and another for testing, that can blow up as often as it "needs" to?

My experience* is that Firefox has many more bugs than Chrome, Firefox bugs take much longer to be fixed, and users are more likely to be using older versions of Firefox so bugs affect us for longer.

In 3 years our webapp has been affected by very few Chrome specific bugs none of which caused us serious problems, or hard to work-around. Chrome has fixed bugs that affected us within a release or two (not years).

Currently our app has about 4 Firefox specific work-arounds. The worst was opened in 2004 and last month took me over 20 hours work to diagnose due to it only affecting us under specific circumstances: https://bugzilla.mozilla.org/show_bug.cgi?id=263945

* we have a complex pure JavaScript front-end, 95% written by me, with no library dependencies (because we started with scriptaculous many years ago, and outgrew it).

Sandboxed tabs was planned with Electrolysis, which was dropped, and will hopefully be delivered with Servo. But it seems a while off production.

I agree. One partial workaround I use is to test sketchy pages in a second Firefox process. If things go bad, it takes out all of those tabs, but not my primary browser tabs.