Firefox 49 will ship with multi-process architecture
Mozilla plans to make Firefox 49 the first stable version of the web browser in which the browser's new multi-process architecture Electrolysis (E10S) is enabled by default for a major part of the browser's userbase.
It seems that the never ending story of making Firefox a multi-process browser will soon come to an end, as plans are underway to deliver the feature to the stable version of the browser with the coming two releases.
Firefox E10S is Mozilla's largest undertaking up until now, if you leave out the creation of Firefox in first place, and that's probably one of the major reasons why the implementation was plagued with delays.
It all started in 2013 when Mozilla announced that it would revive the multi-process project for Firefox, but it took about a year before it made the project a priority.
Mozilla wanted to deliver E10s as early as Firefox 36, but things got delayed along the way and it was clear that it would take longer than expected to deliver the new feature to the stable version of the browser.
The Firefox E10S Story starts with Firefox 48
Firefox E10s met all release criteria in the last six weeks during beta tests including stability, page loading times, startup and shutdown times, memory use, and more.
The current -- and likely final -- schedule for multi-process Firefox is therefore:
- Firefox 48 Stable: about 1% of qualifying users will have the multi-process architecture enabled by default. Mozilla plans to increase the number during the release cycle.
- Firefox 49 Stable: 100% of qualifying users will be able to use E10S.
Qualifying users are all users that don't use add-ons and have not activated accessibility support over 30 days.
The groups that will have to wait a bit for E10S account for about half of our release users and include Windows XP users, users with screen readers, RTL users, and the largest group, extension users.)
Firefox 48 will be released on August 1, 2016, and Firefox 49, on September 12, 2016.
Mozilla revealed in January that 40% of Firefox users don't use add-ons. If you add Windows XP users to that, and those who have used accessibility tools, you'll probably come close to the 50% of users that Asa Dotzler mentioned recently when he revealed the schedule.
This means that the major part of the "close-to" 50% will get E10S with Firefox 48, or at the latest with Firefox 49, and that the remaining percent won't get it with Firefox 49 but at a later point in time.
Not all add-ons are compatible with E10S, and while major ones are already, others are not right now and some will never be compatible (for example abandoned add-ons).
Launching E10S only for users without add-ons ensures that add-on compatibility issues won't play a role when it comes to the release.
Mozilla announced that it will monitor the launch carefully, and there is still a chance that the final release will be postponed if major bugs are discovered in the process.
Firefox Electrolysis is already on by default on the Nightly and Developer channel. 50% of qualifying beta users got E10S with Firefox 47 Beta, and 100% will get it with today's update to Firefox 48 Beta.
Now You: Have you tried Electrolysis already? What's your experience so far?
I’m reluctant to leave ESR but I am going to have to try this.
By the way, pro-tip: Using 64-bit Firefox has the helpful side-effect of ignoring all 32bit plug-ins. Very nice.
I don’t care about Electrolysis. As long as Classic Theme Restorer(and some other addons not exist in Chrome) still work, I’ll keep using Firefox. If those won’t work anymore, what’s the point using Firefox? It’s just a Chrome clone.
one of my most used addons died today :(
I’m with you, I don’t want to use a chrome clone. PaleMoon is not the best alternative, because a lot of addons don’t compatible with it. :(
The latest versions of the addons often won’t work with Pale Moon, because they’ve been updated for the Australis interface, or to maintain compatibility with other aspects of Firefox. However for most of them there is either an earlier compatible version, a PaleMoon forked version, or another alternative.
I’ve just tested Element Properties 10 and seems to work OK (though not sure how much information it normally provides).
Some more info at https://addons.palemoon.org/resources/incompatible/
FWIW, I was running FireFox with over 80 addons, moved to Pale Moon a year ago, and it only took a couple of hours to get the addons (or equivalent alternatives) running in Pale Moon.
Scary times ahead. Will RAM usage skyrocket (nightmare for people that use more than few tabs open at the same time)? Add-ons, how many of them will die (though the “end” of add-ons is coming anyways in the future)? Which other browser one could move to? I hate constant change now that I’ve “grown up”.
According to Mozilla, about 20% more RAM at first, but this will get optimized: https://www.ghacks.net/2016/02/13/multi-process-firefox-will-use-up-to-20-more-ram-initially/
The only kinda alternative is Vivaldi. It has some good things, but it has the chrome-problem: It takes huge amounts of memory and worse, you cannot have more than a couple of tabs because the tab bar is simply shit.
@Pete: Long-term Pale Moon will probably remain a single process browser. I can really recommend it, as it runs quite stable here. Its development path is quite conservative (no fiddling with the UI). Mid-term you could also switch to Firefox 45 ESR, which is single process and will be supported until the first half of 2017. Don’t listen to the advice of Ben though, Vivaldi is a Multiprocess Browser just like Chrome or the Firefox of the future. However, featurewise Vivaldi is great of course. Conclusion: Your only long-term option will most likely be Pale Moon.
K-Meleon is the most real alternative: mature, lightweight, oldschooled ;-) Only few FF-addons can be used with it but it has its own ecosystem with more than 700 native extensions and comparatively easy macrolanguage to create new extensions.
They can be pure macros, or that macros may integrate 3rd-party apps, JS-userscripts or system commands into K-Meleon.
I really dont understand the argument “x plugin wont work anymore its all mozillas fault!” How about instead of getting angry at mozilla you get angry at the developer of the plugin. mozilla is trying new things and trying to improve things but because some plugins break people do nothing but complain. im sorry but plugin developers have to learn to update and evolve there software just like everyone else does.
The old motto if it aint broke dont fix it does not work as often as people want it to. if it did we would still be solely reliant on flash for video. i know change can be a pain but sometimes you have to suck it up and deal. sorry for the rant but people whining about plugins being broken and its all mozillas fault is really really tiring to hear over and over again.
i honestly am excited for mozillas servo project and hope that they really push that, but because plugins will break thats is all people will see and care about. not the many other benefits :/
Firstly, it’s “Add-Ons”, not “Plug-ins”. Plug-ins are things like Flash or Silverlight. And secondly I don’t know what you want to tell us. Most Add-On developers are competent and have supported their Add-Ons for many years. Most popular Add-Ons like Tab Mix Plus or AdBlock Plus have been around for more than one decade and counting. The developers have swallowed every BS Mozilla has thrown at them, but in recent times Mozilla has gone full-blown mad. Electrolysis (e10s), the multiprocess architecture of Firefox will require major updates. This is not the most serious problem, though. In fact, Mozilla plans to abandon their core technologies XUL and XPCOM which many powerful Add-Ons rely on. This is to say that the Add-On has to be rewritten COMPLETELY, if possible at all. WebExtensions are API-based and don’t allow full access to the browser core. For people like me who only use Firefox because of things like Classic Theme Restorer or Tab Mix Plus, this is the final nail in the coffin. The future of those Add-Ons is currently unknown. I really wish some people wouldn’t be so ignorant and reduce the importance of Add-Ons so that it fits into their horizon (“whiners”). And Servo is nothing but a tech-demo as of now.
I concur in that an E10S Firefox without essential add-ons like TMP, CTR, Status-4-Evar, NoScript, Stylish, uBlock Origin effectively makes it not Firefox but a new product that isn’t going to be acceptable to me.
The future of Firefox has been a showdown b/w Mozilla’s direction since Australis (not necessarily or e.ntirely Australis itself) and a direction pre-Australis loyal users would like to see the browser go in. That is: stop compromising the interface and flexibility by continuosly cutting back existing APIs in the name of competing with (copying?) Chrome. Instead, focus on under-the-hood development to bring Firefox up to parity with other browsers.
e10s has taken a ludicrously long time thanks to stupid management decisions as much as any time delays in the coding required. Firefox OS was a bad idea. Indecision about the importance and priority of implementing e10s compromised it’s delivery date as badly as any other factor. If time in the tech world is more like the ratio of dog years to human years, the years since 2009 when Chrome *debuted* as multi-process until e10s is bedded in with full add-on support and reliable running ( probably in late 2017 at the earliest) will be equivalent to about 50 years. It’s only in that sort of perspective that we get to understand how badly Mozilla has managed our beloved browser in the second half of it’s lifespan. They’ve been more concerned with crucifying their own royalty and murdering free speech whilst whinging about their relatively small resources compared to their competitors than actually improving the product.
A certain number of loyal users would tolerate delayed implementation of the right changes if resources were the main issue. That’s the underdog support mentality that created Firefox’s original momentum and market share. But to have the product we have loved so much held back by bad management is just like giving yourself to charity only to find the administrators of said charity have flown the coop with the cash to the Cayman Islands or wherever.
I have just 2 GB of RAM and in the earlier days i can’t even open two windows of chrome or it will crash. Now though, I can open two windows each having 3 tabs open and their share of user scripts and extensions. That’s how optimizable the chromium platform is with well development. But I still miss the old Ff
I know how that feels
I got a PC with 4GB [till couple of weeks ago 3.25GB recognized by XP will I upgraded to 8.1], and I love having lots of tabs open, even if I don’t use em…
for example- now I have…67 tabs open [on 4 windows, it used to be 90 :D ] and finding the lowest-ram-monster browser is definitly important!!
Now I wonder would this new FF be worth trying? or should it be older version of it? or maybe pale-moon which I’ve never tried and was mentioned here by author and commentator….^
Not sure what to chose..
Any suggestion shall be welcomed, GL&HF =]
Try K-Meleon – most lightweight Gecko where I have now 100 tabs open under Win 7 with 2Gb RAM.
Needs some time and efforts to get used with it but it gives your more resource-freedom than any modern browser.
What to say on this? Copied from Chrome, years to late, now trying to get to the right course again to not lose peoples but if that really helps? I doubt. Mozilla has absolute no concept or anything in the pipe to compete with Chrome. Especially there exist a lot of popular forks, so the no spying or ‘against’ google argument not really counts.
But if anyone is interested:
An interesting analysis of what BetterFirefox should do – your list looks virtually identical to how Pale Moon has already been built.
Have you considered collaborating on Pale Moon instead of starting from scratch? I know they’re keen for more developers to participate…
I doubt that moving to E10S will help Mozilla in the short run when it comes to market share. It will be a good feature eventually, when the security stuff is added but right now, it will certainly irritate users especially when the add-on crowd gets it as well.
Why should full implementation of “security stuff” (E10S-related) help Mozilla on market? It will make FF even more chrommish but all this shit really doesn’t help Chrome to be secure and safe browser!
I see only downsides (in terms of usability) without any advantages for normal user.
Cheers CHEF-KOCH … added to Watch [will probably join in :) ]
Just to clarify something: from http://arstechnica.com/information-technology/2016/06/firefox-48-electrolysis-download-details/
“The initial version of e10s will split Firefox in twain, dividing the browser between a UI process and a content process. In short, this means that gnarly websites that hog your CPU won’t cause the entire browser to become unresponsive. Dotzler says the next step is per-tab processes (so one website running out of control doesn’t affect the rest of your tabs), followed by sandboxing for security and then isolating add-ons into their own processes. Mozilla’s goal is for this to all happen before the end of 2016.”
So the initial e10s is two processes, where they are at the stage where it’s about 8 tabs double the memory usage (compared to 8 tabs on a single process). No idea if its linear, probably not. Personally I at most never go past 20 tabs, so I’m not too worried. God help the people who like 200 tabs open :) Although I’m sure there are built in mechanisms for that. Do any users here use chrome with 200 tabs open?
Also: FYI: this actually works in 47, and I think the pref has been around for a while, but I would wait until 48 stable.
// 2631: start the browser in e10s mode (48+)
// After restarting the browser, you can check whether it’s enabled by visiting
// about:support and checking that “Multiprocess Windows” = 1
// use force-enable (hidden pref) if you have add-ons
// user_pref(“browser.tabs.remote.autostart”, true);
// user_pref(“browser.tabs.remote.force-enable”, true);
also there’s this:
// user_pref(“extensions.e10sBlocksEnabling”, false);
Also: I believe chrome has a soft cap on the number of processes, and combines same domain. Which makes sense.
And its still a long long long way to go guys:
Is there a way to only get security updates with Firefox? I’m farly certain there isn’t, but just thought I’d check because there’s only one feature in 47 IMHO and that’s not even a feature. The long-overdue, plugin-replacing h264 support is seemingly almost there with the fix to make Flash-wrapped video play using HTML5 in an iframe. Presumably this bug:
Otherwise I’m only really interested in security and stability updates. In particular security of course. Firefox users should get security updates in the background, without any user interface notifications, as par for the course in continuous integration style. Security updates are probably the only true need for continuous integration deployment methodology.
In addition, Mozilla has to up it’s game to try and be seen as the most security conscious browser as a competitive edge. One of the holes fixed in 47 closes a hole whereby Java applet content could be loaded when they should be blocked. This hole took almost THREE YEARS to fix. It was first reported in August 2013! Chrome had the same issue at the time. If fixed faster, Firefox could have claimed improved security on this topic at least. If Mozilla made security patches the highest priority and fixed them faster, surely this would start to get influential institutions like banks and government recommending Firefox over other browsers again.
No. It’s a compiled executable. Its not like you can totally separate all the components, and it would add just more overhead (note that some updates are small, so they certainly have some modular stuff going on, but most if not all security would be baked into the exe). That’s just my take on it. The good news, for you, is that a lot of FF releases are rather minor (like this one), and that a lot of stuff can be controlled with preferences. Mozilla generally want to strip everything out of FF in my opinion, and make it all “use an addon”. I think/hope the days of dumping shit like pocket and reader and hello are over. I prefer this, as I want them to focus on stability, security, and performance. As for making security updates mandatory, silent, and uncontrolled – I disagree – that is taking away from the end user (although not quite the same, look at MS enforced updates in Win10) – but its a moot point. You either upgrade the browser or not, you get all the changes or not – and then you tweak if needed :)
Just because it’s compiled doesn’t mean it can’t be frequently updated stealth style. Since I reluctantly installed Chrome for development testing, I don’t remember many significant moments bothering with updates. I’d say I’ve never touched it and it’s still up to date but I suspect that would be rose coloured glasses. Updates can be done better than Firefox does them. The Firefox Maintenance Service lives on every Windows installation. Correct me if I’m wrong but that enables Mozilla a substantial degree of flexibility in updating. Sure it’s meant to make Firefox start faster but it’s a service, it’s always there, surely Mozilla can use it to periodically check for updates, background download those updates, and only install them the next time Firefox runs. It should be a lot more transparent though it has become smoother over the last few years. It had to, of course, with these stupid rolling releases. But there’s got to be room for improvement.
Thanks guys for the feedback and sorry for the self-promotion, but it was/is related to Firefox. Can someone link me the mentioned Palemoon list? So that I can contact them. THX.
Just ask moonchild if you get no reply, he’s very contactable, friendly and helpful. He replied within minutes when I emailed him about palemoon entries for the ghacks user.js
Is there anything that Firefox can and Cyberfox fails to do?
Cyberfox is a 64 Bit browser, this means you won’t be able to run any 32 Bit Plug-In (Mind you: Plug-Ins are not the same as Add-Ons – you will be able to run every Add-On perfectly fine). However, this is also true for Firefox 64 Bit and any other 64 Bit variant of Firefox, like Waterfox. Apart from that rather minor issue I can only see advantages in Cyberfox.
I had a look at Cyberfox and I was very close to switching to it. However to switch one’s browser, no matter how similar it might be to your existing browser, one would want to have some faith the company behind the new choice is making a long term commitment. Not only does “8pecxstudios” have an inexplicable, unpronounceably stupid name, it doesn’t inspire all that confidence in me in for other reasons. Not yet, anyway.
It’s also quite aggravating that a random unheard of company can achieve two significant optimizations of Firefox that Mozilla could also easily achieve, but choose not to. IIRC 8pecxstudios are doing nothing new to get Cyberfox running in 64 bit and optimized for particular processor architectures. Processor optimizations have been a compilation option Mozilla coders built in to the browser’s code base for years. Probably since day one. But they choose not to use them. The same seems to apply to 64 bit. Mozilla has a 64 bit version but due to ancient NPAPI plugin issues, they’ve never prioritized their 64 bit version!
Shooting themselves – and all their users – in the foot!
Using forks is respectable, but if one cares about the Firefox codebase which is a necessity to those forks, perhaps one should consider using Firefox directly to support it and help one of the two Chrome competitors fighting against monopoly. Monopoly is well proven to kill innovation and lead to abuse against users.
Chances are, some customisation and maybe a theme will make Firefox look like any of its forks and function similarly under the hood.
Privacy wise it’s also indisputably more interesting for users to use a privacy-oriented mainstream browser like Firefox rather than a privacy-oriented rare breed that stands out. (This is an unfortunate but inescapable truth)
Disable signing renforcement:
Working like a charm on FF48.0.2 with many not signed plugin :)