News for hackers. Stuff that matters.
Re:Firefox 27 Released by lingben on Tuesday February 04, 2014 @12:06PM (#7178681)
Right now firefox beats the pants off chrome for memory usage - that is why I'm using it.
Re:Firefox 27 Released by clarry on Tuesday February 04, 2014 @12:19PM (#7178784)
Still it can't deal with memory pressure, and will just eventually oom & crash.
Re:Firefox 27 Released by bad_user on Tuesday February 04, 2014 @12:48PM (#7179021)
The problems I've been having in Firefox are with badly behaved apps - with memory leaks and stuff. Firefox manages memory better than Chrome, but when a tab leaks it can freeze the whole browser. Still, this happens rarely enough that it isn't an issue for me, plus Firefox is much better at keeping many tabs open. I did disable Flash and now I have to "click to play" - but I'm doing the same thing in Chrome too, because fuck Flash.

I feel like Chrome is better when rendering stuff - like we've got this plotted graph in our app's admin, rendered with Javascript and I believe HTML5 Canvas or something, with too many points on it and Firefox can't handle the rendering, it chokes, whereas Chrome can do it. On the other hand, I feel like Firefox is snappier in general.

Re:Firefox 27 Released by ndesaulniers on Tuesday February 04, 2014 @09:27PM (#7181544)
> tab leaks it can freeze the whole browser

Process per tab is _almost_ ready: http://billmccloskey.wordpress.com/2013/12/05/multiprocess-f...

You can play with it by setting `browser.tabs.remote` to true in about:config.

Re:Firefox 27 Released by ZoFreX on Wednesday February 05, 2014 @05:58AM (#7183043)
I don't even want process per tab, I'd be happy with something much simpler: Tell me which tab is causing Firefox to use 100% CPU! This happens fairly frequently with lots of tabs open (I guess I use a lot of websites with really bad javascript) and the only recourse is to either quit Firefox entirely and re-open it, or play a guessing game.
Re:Firefox 27 Released by ndesaulniers on Wednesday February 05, 2014 @05:46PM (#7187416)
Firefox will stop any client side JavaScript after 5-6s and ask whether you want to keep running the slow running script or not. If will even change to that tab before prompting you. Isn't that what you just described?
Re:Firefox 27 Released by ZoFreX on Thursday February 06, 2014 @10:56AM (#7190955)
Only if it totally locks up the browser though, not if the browser remains responsive but eats a lot of cycles.
Re:Firefox 27 Released by ndesaulniers on Tuesday February 04, 2014 @12:30PM (#7178879)
What's your host environment and browsing habits like (lots of tabs open, many pinned tabs, lots of video, flash video)?
Re:Firefox 27 Released by clarry on Tuesday February 04, 2014 @01:08PM (#7179158)
Having just one gigabyte of physical RAM, I run Firefox with a 512M rlimit (the OpenBSD default), as I have to reserve some space for other software too (and swap is painfully slow on the spinning disk).

As far as browsing habits go, I do usually have quite a few tabs open (currently I've got 38, though they're not all loaded; I don't think I often go much over this often). None are pinned. Video is disabled, and I've got no flash or plugins.

I've got three extensions (noscript, https-everywhere, duckduckgo plus). I tend to run with scripts disabled and this helps a lot with memory pressure. Sometimes I make the mistake of thinking I need to enable javascript for something (hi Google), and that's usually when memory usage skyrockets and the thing dumps core :-) But eventually it'll happen even if I don't allow any scripts.

Re:Firefox 27 Released by ndesaulniers on Tuesday February 04, 2014 @09:46PM (#7181600)
This is getting close to the limit of my expertise. What resource is bumping into it's system limit? Can you back up that answer with a stack trace or core dump? If so, would you mind filing a bug in our bug tracker? [0] It looks like if you put [MemShrink] in the Whiteboard field, the memshrink team should pick it up. If not, you could jump in #memshrink in our IRC server, irc.mozilla.org, and let someone there know. [1]

You can ping me directly on irc, too. My handle is a literal new line character (which _is_ a valid nick), and I'd be happy to help you get the bug filed if you run into any issues. Once you have a bugzilla account, you'll see it's generally pretty easy to submit bugs and provide feedback in channels that allow us to more easily find and work on issues. We're happy to help, too!

I'd file the issue myself, but I'd prefer to enable you to see how easy it is, and for you to provide the stack trace. Even providing steps to reproduce issues makes you a part of our Mozilla community. Welcome!

[0] https://bugzilla.mozilla.org/ [1] https://wiki.mozilla.org/Performance/MemShrink

Re:Firefox 27 Released by slyall on Tuesday February 04, 2014 @03:33PM (#7180069)
Re:Firefox 27 Released by nnethercote on Wednesday February 05, 2014 @03:35AM (#7182593)
512 MiB rlimit and you usually have 10s of tabs open? I'm not surprised you're having problems; that's a pretty unusual and challenging configuration. Do other browsers behave any better?
Re:Firefox 27 Released by jobigoud on Tuesday February 04, 2014 @12:54PM (#7179067)
You are talking about Chrome right ? We have a memory intensive app which works perfectly on FF26 (and IE11), and crashes the tab on Chrome 32 with memory hitting 3GB+ and up. Their garbage collector must be leaking like crazy. (Works in Chrome 22).