Live data from Hacker News

Firefox 15 plugs the add-on leaks

blog.mozilla.org

101–110 of 125 posts

Re: Firefox 15 plugs the add-on leaks

#101

This is awesome news. The single biggest reason that I initially moved to Chrome, from Firefox, was the memory consumption situation. But lately, Firefox has gotten a lot better, and I've basically switched back to FF now as my primary browser. Hopefully this will make the situation even better. And that's good for everybody. Competition is a Good Thing and it would be a shame if Chrome become so dominant as to compl…

Chrome and Firefox been using memory at equal proportion. I use exactly two add-ons and comparing FF/Chrome to IE to check on memory consumption, I see IE doing much better.

Re: Firefox 15 plugs the add-on leaks

#102

Earlier quoted context omitted.

Wow, that's actually very nice in terms of RAM usage. Right now in chrome, with only 6 tabs open I am using (according to about:memory)800MB of RAM[1]. I may very well switch back to Firefox if I can get performance like that. [1] http://puu.sh/JXL9

Is it normal for a gmail tab to use 250+ MB?

Only 250MB? For me it's 600-850MB... it's ridiculous.

Re: Firefox 15 plugs the add-on leaks

#103

I'm on Aurora right now (15a2), and after YEARS of bitching am cautiously optimistic this is really it! Just a few weeks ago I was closing FF once a twice a day as it approached almost a GB of RAM. Now it's holding steady at under 230mb! I have few extensions, but use firebug. I knew it was part of the problem but it's unfortunately essential for me. With this fixed perhaps they can now work on tab isolation, another…

Wow, that's actually very nice in terms of RAM usage. Right now in chrome, with only 6 tabs open I am using (according to about:memory)800MB of RAM[1]. I may very well switch back to Firefox if I can get performance like that. [1] http://puu.sh/JXL9

Weird. My version of about:memory shows ~600MB used, but the Task Manager in Chrome (Shift-ESC) shows only ~160MB.

Not sure what the discrepancy is there.

Re: Firefox 15 plugs the add-on leaks

#104

Earlier quoted context omitted.

>Firefox lost market share because Chrome was a leaner, faster, and better browser. Not to mention secure. Sandboxed PDF reader, java whitelisting, sandboxed and auto-updating flash, etc while FF and IE were arrogantly telling users "Fuck you, update your own add ons, they're not our problem" even when all the exploits were from these add ons. Now IE and FF ape some of these features. I can't wait to hear the next an…

I always wonder why Firefox gives the option to update Java and Flash et al when it never works. Has anyone ever gotten plugins to update from within Firefox?

Totally. Irritates me to the core. First give an option to update it 'automatically', then tell them "Sorry, please do a manual install'!

Re: Firefox 15 plugs the add-on leaks

#105

> Leaky add-ons usually cause zombie compartments by holding references to DOM structures within a web page even after the page has been closed or navigated away from. This prevents Firefox from garbage-collecting the page’s compartment. If an add-on relies on persistent references to closed pages by design (such as a form data history plugin), is there a way to adapt it to this new feature to not let Firefox sever i…

Copy them. I can think of no good reason that one would ever legitimately want or need to keep references to a page that is no longer loaded. It's like returning a pointer to a local variable from a C function; you should expect it to be invalid once it's scope is gone! Copying is likely to use less total memory anyway, because it will rarely result in having to allocate new compartments for the plugins, but it will allow an entire compartment for a page to be collected, that is otherwise kept around for the sake comparatively small objects.

This actually seems to me to have a lot of implications for the design of garbage collectors generally. In theory, correct garbage collectors are never supposed to collect objects that may still be used, which in practice usually means "that are reachable". But, we've already got such a thing as "weak references" to fudge that a little. This seems like kind of the opposite idea- having a sort of "superstrong" owner reference such that once it goes away, you know that the object will never be used again (or shouldn't be, anyway, with future uses constituting bugs), and can be collected despite what other references might exist.

It also seems like one of those obvious-after-the-fact genius ideas to have garbage collection running independently per compartment. Do Web Workers, e.g., get their own separate garbage collection, too? Any kind of shared-nothing concurrency model ought to be able to do the same thing to improve GC pauses; are there language implementations (I'm thinking probably Erlang here) that actually do that?

Re: Firefox 15 plugs the add-on leaks

#107
post #98

Earlier quoted context omitted.

Architecture choices or business model?

I'm not close enough to divine the details, but I'm talking about Chrome plugins not being able to block requests. This might appear to be a facility to protect advertising vectors, but I don't know for sure. http://superuser.com/questions/257102/why-doesnt-chromes-adb...

This used to be true, not thing have changed since then. Chrome didn't have any WebRequest api, so Adblock Plus for Chrome loaded some JS/CSS in the page and tried to hide/block as much ads as possible. [1]

Since Chrome 14 (i'm not totally sure about this version) WebRequest api where implemented, but only as experimental api (you would have to enable them from chrome://flags), and Chrome >17 enable them by default.

So Adblock Plus >1.2 beta [2] and Chrome >17 should block ads just like Firefox. It's possible that there are some small difference, but that's likely a bug and should be reported.

[1] http://adblockplus.org/development-builds/experimental-adblo...

[2] http://adblockplus.org/releases/adblock-plus-12-beta-for-goo...

I'm not sure what's the situation with other adblocker for chrome.

Re: Firefox 15 plugs the add-on leaks

#108

> Leaky add-ons usually cause zombie compartments by holding references to DOM structures within a web page even after the page has been closed or navigated away from. This prevents Firefox from garbage-collecting the page’s compartment. If an add-on relies on persistent references to closed pages by design (such as a form data history plugin), is there a way to adapt it to this new feature to not let Firefox sever i…

Addons have to copy out the data they want before the page closes. It isn't acceptable to keep around all data on a page just to keep a few little bits of information.

Re: Firefox 15 plugs the add-on leaks

#109

Earlier quoted context omitted.

>Firefox lost market share because Chrome was a leaner, faster, and better browser. Not to mention secure. Sandboxed PDF reader, java whitelisting, sandboxed and auto-updating flash, etc while FF and IE were arrogantly telling users "Fuck you, update your own add ons, they're not our problem" even when all the exploits were from these add ons. Now IE and FF ape some of these features. I can't wait to hear the next an…

I do not prefer chrome's PDF reader. I use XPDF, mostly like how it works (I generally dislike PDFs, but that's another issue). Auto-updating is a misfeature on a package-based Linux distro (e.g.: any sane one). The extensions management is better in Chrome. But not the permissions handling. I'd prefer to tell my browser what it is/isn't allowed to say about me, and for extensions to sort out the damage (is it worth…

How is the permission handling in Firefox better?

Re: Firefox 15 plugs the add-on leaks

#110

Earlier quoted context omitted.

I do not prefer chrome's PDF reader. I use XPDF, mostly like how it works (I generally dislike PDFs, but that's another issue). Auto-updating is a misfeature on a package-based Linux distro (e.g.: any sane one). The extensions management is better in Chrome. But not the permissions handling. I'd prefer to tell my browser what it is/isn't allowed to say about me, and for extensions to sort out the damage (is it worth…

How is the permission handling in Firefox better?

Meh, I probably overstepped on that one. I'm not sure how the FF permissions work, and frankly, the extensions infrastructure is one that leaves me a bit queazy. That said, I'd prefer both (or rather, all) browsers acted as I indicated: allow the user to state what data are shared, with what granularity and retention policies, and report on what site(s) request/receive it. Apps can work out whether or not they care to play.
Post reply on HN