Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Update on Multi-Process Firefox

blog.mozilla.org

Post reply on HN
Print view View original

121–130 of 175 posts

Re: Update on Multi-Process Firefox

#121

Post by nikolay »

ngokevin wrote: post #118
nikolay wrote: post #112

I've been gradually disappointed by Firefox - more and more websites no longer work. Okay, I'm using Firefox Developer Edition, but nonetheless, it's been disappointing compared to, let's say, Google Chrom Canary. Memory and CPU footprint is worse than latest Canary, too!

The compatibility issue isn't to blame on web developers increasingly testing only on or making their apps exclusive Chrome?

I think it's the privacy and other "features" that make Facebook and Google login not work, for example.

Re: Update on Multi-Process Firefox

#122

Post by wsha »

mandelbulb wrote: post #56

Make Vimperator great again - still no support. :'( Two year old github issue but yet any development to be seen or reported :/

Sadly it's not realistic. Even if someone were to put in the time to make it work, Vimperator would just break again within the next year when XUL addons are deprecated in favor of WebExtensions, and unfortunately WebExtensions do not support all the features of Vimperator. You can use VimFx for now, but again it is a XUL addon and would take a massive amount of work to port to a WebExtension. Hopefully someone can get Vimium or cVim ported over as a WebExtension. I have played around some with this but haven't got it to work. It seems some of the needed API's are missing from WebExtensions at the moment.

Re: Update on Multi-Process Firefox

#123

Post by newscracker »

I like how Firefox multiprocess has evolved and continues to evolve taking measurements into account.

When multiprocess was announced several years ago, my main worry was that it would be a Chrome clone with one process per tab and become sluggish like Chrome (my general use of browsers is a minimum of 10 tabs and on certain machines a lot, lot more). Finding a balance between responsiveness, security and stability is what I'd prefer. I personally don't see many stability issues in Firefox even without multiprocess. So I'm happy with this progress even though it may seem slow to dime people or that Firefox isn't keeping up with Chrome (the latter is meaningless for my use since Firefox and it's extensions provide a lot more).

Re: Update on Multi-Process Firefox

#124

Post by mdekkers »

bgwhn wrote: post #12
saberworks wrote: post #8

It would be really nice if the addons/extensions page showed which addons were listed as compatible and which weren't. I might thus make a decision to disable/delete an addon or two that were blocking me from using multiprocess firefox. As it is, I have no idea if my firefox is using multiprocess at all. All I know is that I use a lot of addons and I'm probably not, then, using multiprocess.

You can figure out what add-ons are compatible by cross-referencing http://www.arewee10syet.com/

That site listed a few of my addons as incompatible, but turned out that that site hasn't been updated in a while. Example, it lists Lastpass as incompatible, but it works just fine for me.

Re: Update on Multi-Process Firefox

#125

Post by ronjouch »

paddy_m wrote: post #104
ronjouch wrote: post #5

Earlier quoted context omitted.

And if you want so, the about:config key to fiddle with is called "dom.ipc.processCount" and currently defaults to 1, with plans to raise this soon, as described by the article. EDIT also, to come back to DanBlake's question about whether Firefox now "run[s] each tab in its own process yet" : no it doesn't; the architecture that was chosen is a process pool, available to all tabs. I won't elaborate much as you can fi…

Would you mind linking to some of those blogposts. I used to read the weekly mozilla meeting notes but they seemed to have stopped including content about 2 years ago. https://wiki.mozilla.org/WeeklyUpdates I just looked and now it seems that notes link out to separate pages. What's the best way for a non-contributor to follow firefox development?

> "Would you mind linking to some of those blogposts."

A quick parsing of https://wiki.mozilla.org/Electrolysis yields "Memory Usage of Firefox with e10s Enabled" [1] and its followup "Are they slim yet?" [2], with [1] closing with these words:

"Simply put: the more content processes we use, the more memory we use. On the plus side it’s not a 1:1 factor, with 8 content processes we see roughly a doubling of memory usage on the TabsOpenSettled measurement. It’s a bit worse on Windows, a bit better on OSX, but it’s not 8 times worse. Overall we see a 10-20% increase in memory usage for the 1 content process case (which is what we plan on shipping initially). This seems like a fair tradeoff for potential security and performance benefits, but as we try to grow the number of content processes we’ll need to take another look at where that memory is being used."

Some comments at https://billmccloskey.wordpress.com/2013/12/05/multiprocess-... also hint at pre-existing discussion on the one-process-per-tab subject, but sadly without linking to it. You'd probably followup with a mailing lists search.

[1] http://www.erahm.org/2016/02/11/memory-usage-of-firefox-with...

[2] http://www.erahm.org/2016/02/12/are-they-slim-yet/

--

As for "What's the best way for a non-contributor to follow Firefox development?", I follow https://planet.mozilla.org/ , and if that's too much of a firehose to your taste, maybe try https://hacks.mozilla.org/

Re: Update on Multi-Process Firefox

#126

Post by JoshTriplett »

TheRealPomax wrote: post #120
JoshTriplett wrote: post #119

Earlier quoted context omitted.

Apparently the SSL certificate for that site expired a few days ago and nobody renewed it. Certificate expiration seems like one of the most ridiculous aspects of TLS. Its only use case (apart from generating more money for CAs) seems like "we somehow can't revoke the certificate, but if we wait it'll expire".

that's kind of also a really important use case. "There is no reason to trust this certificate in perpetuity, so pretending it hasn't been compromised over a long period of time would be stupidly insecure. Let's expire it after X months". There is no reason to trust a certificate that's older than a year, for instance, there's been more than enough time for someone to reverse engineer the keys if they wanted. The rea…

> There is no reason to trust a certificate that's older than a year, for instance, there's been more than enough time for someone to reverse engineer the keys if they wanted.

Cryptography does not work that way.

> It's pretty trivial to set up cert renewal, so why didn't they?

That holds true today, with Let's Encrypt; their short expiration date seems to exist largely to force people to automate it, and in that regard it seems quite effective. But prior to that, many CAs did not have scriptable automated processes to renew certificates.

Re: Update on Multi-Process Firefox

#127

Post by oblio »

nikolay wrote: post #121
ngokevin wrote: post #118

Earlier quoted context omitted.

The compatibility issue isn't to blame on web developers increasingly testing only on or making their apps exclusive Chrome?

I think it's the privacy and other "features" that make Facebook and Google login not work, for example.

What "privacy"?

Re: Update on Multi-Process Firefox

#128

Post by febed »

satysin wrote: post #50

Even on a clean Firefox 50.1.0 x64 install with just uBlock Origin and multi-process enabled Firefox feels quite a bit slower than Chrome. This really shouldn't still be a thing especially with Chrome also having to deal with a bunch of Flash based content on some sites (looking at you BBC News).

Vanilla Firefox with multiprocess was sluggish for me as well. I had no such issues with the latest Firefox Developer edition - give it a shot.

Re: Update on Multi-Process Firefox

#129

Post by cpeterso »

cpeterso wrote: post #113
pwg wrote: post #61

Has anyone noticed bookmarklets not working when multi-process is turned on? I installed a V 51 developer edition to test out multi-process, and when multi-process was on, my bookmarks that modify the dom did not work. Any one else noticed anything similar?

That's sounds like a bug, as if the bookmarklet is executed in the main browser process and can't see the page's DOM. What DOM bookmarklet water you using? The Pinboard bookmarklet works for me but it is not agreeing the DOM.

I tried the "WhatFont" bookmarklet, which inspect DOM elements and reports the fonts used. It works for me in Firefox Nightly 53 with e10s enabled.

http://www.chengyinliu.com/whatfont.html

Re: Update on Multi-Process Firefox

#130

Post by yuubi »

JoshTriplett wrote: post #119
bgwhn wrote: post #12

Earlier quoted context omitted.

You can figure out what add-ons are compatible by cross-referencing http://www.arewee10syet.com/

Apparently the SSL certificate for that site expired a few days ago and nobody renewed it. Certificate expiration seems like one of the most ridiculous aspects of TLS. Its only use case (apart from generating more money for CAs) seems like "we somehow can't revoke the certificate, but if we wait it'll expire".

If the cert is revoked, the revocation notice only needs to be published until the expiration date instead of forever.
Post reply on HN