Slashdot Mirror
Firefox 6 Ships Next Week, 8 Blocks Sneaky Add-Ons
CWmike writes "Mozilla is on track to release Firefox 6 next week, according to notes posted on the company's website. 'On track with a few bugs still remaining. No concerns for Tuesday,' the notes stated. Firefox 6 includes several noticeable changes, including highlighting domain names in the address bar — both Chrome and Microsoft's Internet Explorer 9 do something similar by boldfacing domain names — and reducing startup time when users rely on Panorama, the browser's multi-tab organizer. Meanwhile, Mozilla said this week that starting with Firefox 8, Mozilla will automatically block browser add-ons until users approve them, which should put an end to sneaky installs."
Quick, invent more 'stuff' to throw at it, so we don't have to fix the bugs we introduced in ff4!
"People who do stupid things with hazardous materials often die." -- Jim Davidson on alt.folklore.urban
I'm really looking forward to this.
I'm planning on running 12 by the end of the year. You guys are waaaay behind.
I would love an auto update to this that did not say "Hey go download here to get the latest version" Or even the current "Downloading patch, please apply later" bit. IM not a chrome user but i like that it auto updates with out intervention.
And by 2014, we'll be preparing to upgrade to Firefox 24. \o/
now they need to do something about helping authors keep their addons up to date or making them work some other way.
example google voice addon - didn't work,wasn't supported on 4 via the official addon site. I ended up going back to 3.6 and finally found someone who updated the addon for 4.*
Screw it now, I'm staying on 3.*
It gets to a point where an addon is part of the functionality of the browser in this case the voice addon was something I relied on daily instead of keeping the tab open.
I'm using FF8 alphas on the Nightly channel, which is part of the Moziila PPA in Ubuntu. It's fantastic. It uses way less memory and is way faster. It's also way stabler than nightlies were when I was running Moziila nightlies in 2001, and they were pretty good even then. The only downside is extensions that haven't caught up. If you're clear for those, I heartily recommend it.
http://rocknerd.co.uk
FireFox7, aka FinalFantasy7, will have a huge step forward dealing with memory. FF6 doesn't have such nice awaited features. I'll skip #6.
-- Rastignac was here.
Great, another rounds of broken add-ons.
First 4, then very quickly Firefox 5 and now version 6? Where is the consistency?
I'd rather they add some easy way to let users install addons that say, "Does not support Firefox x.x". They can put a big disclaimer/warning/alert to make sure the user knows what they are doing, but with the Firefox rapid release schedule I am tired of having my addons break because of version string issues.
One example is the Stylish addon. I am using the Firefox 6 beta in Ubuntu 11.10 alpha and Stylish refuses to install due to the version string. The addon info says it supports Firefox 3.6 - 6.0a2 (key part being "6.0.a2"). That tells me that it should work in later alpha/beta version 6 builds.
Firefox really needs to address the issue of how addons determine whether or not they are out-of-date. The browser version is no longer a useful metric for that.
Exactly how are they going to block that? Anything FireFox has access to, so would an (admin-level) installer.
Unless they're taking a signature from the add-on and some information unique to the user profile and generating a hash/code or that, and keep the hashing algorithm secret somehow?
More 0-days woohoo
Sneaky add-ons are installed by software that has the priviledge of messing with the system, like windows updates. When they have such a priviledge, it's easy to manipulate the user's profile to make it accept the new add-ons. They can't protect themselves against this.
I suppose it's no surprise to anyone who's been paying attention to Firefox development for the past several years, but for fuck's sake, listen to your users and stop with the version number inflation!
Seriously, what makes this a Firefox 6 and not a Firefox 4.2? What new features does it add? Apparently the only really "stand-out" feature is graying out anything that isn't the domain name in the useless-bar. I mean, Awesome Bar.
(Seriously, I like the concept, but I've had quite a few instances this past week where instead of finding "the page I was just on five minutes ago" it does something like "page 3 of this article you read two months ago" with no hint of the URL I'd opened literally ten times already that day. Awesome. Here's an idea, can Firefox try and fix it to make it useful? Like sort based on number of times a page was viewed, counting reloads, so that typing the URL to a forum doesn't find page 2, 3, 4, and 5, but never page 1 because I don't click on the page 1 link enough, I just reload the forum?)
But back to the version number issue - quick, how many people know what version number Chrome is up to off the top of their head? Anyone?
How many people using Firefox 5 here have literally forgotten that they're using Firefox 5, because the last really major update was Firefox 4? I still think of it as "Firefox 4" because it looks identical, and have to be reminded that they've inflated the version number for no useful reason.
Seriously, stop blindly aping Chrome! If you're going to copy something Chrome does, try and understand it! For example, take removing the status bar. Chrome will expand the little URL popup that replaced the status bar if you continue hovering a link. Firefox 4 and 5 don't. And for some reason they randomly switch between left-aligning it and right-aligning the popup. And for fuck's sake, why don't you just expand the popup to fill the entire horizontal width of the window?! I've got the room to display the entire URL! Why doesn't Firefox bother doing so?!
But kudos for aping (poorly) the feature in IE 9 that warns when third party addons have been installed and gives you the option of not using them. It's nice to know that you're going to go ahead and do that after crying about how it's impossible to do, even after IE had launched with that feature.
You are in a maze of twisty little relative jumps, all alike.
I run a PAE kernel and found out the 4GB process limit can come in handy when FF goes apeshit. :)
"This post is an artistic work of fiction and falsehood. Only a fool would take anything posted here as fact."
FF6 has a new Tools > Developer submenu, and they moved Error Console, View Source and Web Console there. Moving View Source there was a big surprise. Any reasonable developer might get totally freaked out searching high and low for View Source if they didn't know about that move.
In case you missed it, Web Console in FF5+ is like the console in Firebug when you have it set to enter JS commands at the bottom of the pane. But the difference is, Web Console is always available. It's not a plug-in like Firebug. So it's something you can count on, even if you upgrade and Firebug breaks in the new version.
No, you dumbass, do try to pay attention. This blocks ALL add-ons until the user explicitly says "yes, I want to install this add-on". Normally when you install an add-on you see that ticking countdown box that warns you, so that's when the user would normally say "yes". But sneaky programs could just slip add-ons into the Firefox directory and it would load them up with no warning at all. Now it'll say "new add-on detected, really let it run?" and if you say no, it will be disabled.
I realize I used a few long words in that but hopefully I got the message across.
I was a long-time Firefox user, I even was part of Spreadfirefox.com and was a "Zealot" that managed to convert my Mom and brother to it and when I was at college I got the IT admins to install it back in the bad old days of IE6. But Firefox has lost its way. Its peak was 1.0 to 3.6. The memory leaks, the obsessive version numbers, the theft of the status bar and ignoring its users and wasting 80 million dollars a year is the last straw. I uninstalled Firefox today, and have switched to Chrome on my main PCs and Safari on my iPad.
Netscape died a horrible death, and Firefox seems to be repeating it. Hopefully enough concerned users fork the Firefox 3.6 code and "re-pheonix" it before it's too late.
They need to add the Firefox 3 UI and the status bar back as an option. Some of the changes they have talked about lately are dumb (like adding facebook functionality to the navigation bar). However, none of them compare to the UI change for no reason. I think the last line of one of the articles prove that.
As of the end of July, 48% of Firefox users were running Firefox 5, while just 11% were still on Firefox 4.
This leaves over 40% running 3.6 and before, not a small number by any stretch. And what was changed from 3.6 to 4? A bunch of things no one would notice immediately and the UI overhaul. I think those demographics are a message. Is Mozilla listening?
I'm still using 3.something.
I bet none of them are standards compliant.
Yes, dumbass, but the last time people suggested that Firefox consider implementing something like that, they claimed it was impossible and couldn't be done.
But the point still stands: providing the ability to block admin-installed addons (as Firefox still doesn't support Group Policy, WTF?!), means that Firefox is a no-go in the enterprise.
Every change they've done recently moves Firefox one step closer to be banned from any computer run by a competent sys-admin.
How about obfuscation like polymorphic code shipped to end users. The installer generates a binary using a set of instructions that vary based on a randomizing algorithm that has a cryptographic random number generator for secure input. When finished it destroys all evidence of what seed created the binary. Then that binary when ran, uses prevention techniques to prevent scanning the binary by malware to determine where the new exploit needs to go. Make it a huge pain to implement these engines by increasing randomness, not software techniques. Keep that algorithm to yourself and ship new code for each customer when they click the download link.
A: Each exploit is now traceable to some extent back to which binary they had to create the exploit.
B: Code can't spread because it can't figure out what to do? (reaching here)
C: Think of memory randomization, applied to the compilation process with regards to unique instruction ordering with the same output and data layouts that have varying patterns. (think of endianness but at the memory/storage areas in the binary, with MANY more types of layouts designed to be confusing and unique per binary. Every other bit, every 3 bits then 2 back with opposite truth parsing(0,1 switched), etc.Instead of BigEndian, use cascading vertical/horizontal binary spreads with randomized loaders also generated.
That would be damn near impossible to crack, but after dabbling with Genetic Programming (programs that generate lots of programs with a sliding wedge towards a certain fitness and mutate those programs similar to evolution) I believe what I described above is possible.
We generated code using such a program that could be fed 255 integers (5 digit limit per integer) of various counting schemes (start at 0 count by 2, count by 12s, 4 forwards, 2 backwards with mutation, Fibonacci! etc.) and it figured out how to write python programs to solve the problem or generate the input from the source code without saving the input inside the program! Even more is every time it's ran it figures out a different program to solve the input. Using a fitness algorithm for complexity trimming (some programs took hours to generate 255 integer output, some took 1 minute, some took 400 milliseconds) we can trim the user experience to an acceptable level.
That is where I see the future going. You heard it here back in measly 2011! And everyone thinks I'm crazy.....
Why can't that work? On a serious level. I know the talent is in here reading this.
This is now the third time in a row they've pushed these new "update or die" version updates incompatible with the automatic update features. I may be on top of things and know when to go click "Help -> About" to trigger the update process, but my parents and such aren't, and I'll be damned I'm going to individually walk them through the process every few weeks like this. I'm done with Firefox.
At this point, my family is more likely to have an up-to-date and properly-patched browser with IE than with Firefox.
And by 2014, we'll be preparing to upgrade to Firefox 24. \o/
It's doubtful that Firefox will be relevant much beyond mid-2012. It's already sliding into irrelevancy. It is becoming the XFree86 of the browser world.
There's just no pressing reason to use Firefox these days. It's no longer 2002, where the only other browser was IE6. With Chrome and Safari and Opera and even IE9 available to us, Firefox offers no benefit.
All of its competitors today are faster, they use less memory, they have better developer tools, they are more extensible, they have smaller installers, and don't go changing their UI every fucking release.
The UI bug that allows light colored system fonts to screw up input boxes is still there. Why are they still phoning in the Linux version?
What sites are you going to where your browser is soaking 7GB... I've never seen FF run over 800MB, though that is too high imho. However, never really had an issue with it, though I've been using Chrome as my primary for about a year.
Michael J. Ryan - tracker1.info
Can't tell if trolling, or just stupid.
They didn't claim it was impossible forever no matter what.
They claimed it was impossible with the current architecture of Firefox.
Obviously they've re-written the way the browser handles Addons, and can now push this feature.
Be patient humans.... version 42 will answer all your questions....
".....to report whether they still work or are having some issues with alpha and beta releases. Note: Recommended for alpha and beta users only! in other words, not for anyone who is interested in maintaining a production stable system.
Firefox has jumped the shark, "upgrade or die". fuck you, Mozilla Corporation.
will they fix the memory leaks ?
I can answer that. Go to video sites that are NOT Youtube, like most of the 'Ow my balls!" kind of dumb vid sites? watch FF suck down memory like a wino sucking down MD 20/20. The Chormium based and IE seem to give mem back when you close tabs, FF? Not so much.
I have also seen cases where I had left FF running and went off to do something else and forgot about it and several hours later come back and memory had JUMPED a good 30-40%! Again it seems to be tied to whether or not you had played any videos that day.
This is why after years of having FF in my standard customer build I have replaced it with Comodo Dragon which IMHO has some nice extra features as far as security and no calling home to Google so it is the best Chromium based for me and my customers. I've always loved the FF UI but IMHO after the 3.6.xx branch FF has just gone to shit. Its memory usage is nuts, it spikes the living hell out of CPUs, especially if the tab you are launching contains video, its just a mess. I have to support everything from netbooks and late P4 office machines to multicore and I need something that will give a consistent experience. Dragon does, FF don't.
Personally I think it is the Gecko engine. I just don't think it has been able to take all the extra crap they have bolted on like plugin separation. Where FF once was a nice lean solid browser ever since Chromium came out it has been "Me too!" to the 50th power instead of just making FF the best FF it can be and I think it shows. I haven't seen memory leakage this bad since the 2.x.x branch. I hope they fix it so I can have another browser in the toolbox but right now the responsiveness and resource usage just isn't there. When it takes a good 25+ seconds to launch FF on a 2.8GHz quad with 8Gb of RAM? That is fucked up. The Dragon takes less than 5 seconds from click to typing.
ACs don't waste your time replying, your posts are never seen by me.
I'm probably pissing in the wind responding to an AC but here goes anyway. I've used Firefox since it was Phoenix 0.6. I've run Phoenix/Firebird/Firefox on Windows 2000 Pro, Red Hat/Gnome, on Pentium III and Pentium 4 systems. Firefox has always had an appetite for memory but I've never observed the kind of consumption you describe. I've loaded the high res HST image of the Helix Nebula with the intent of breaking Firefox and it didn't break. It just used seven hundred meg and wow, none of the memory leaked, it was all released when I closed the tab. I've opened two dozen posts on /. in separate tabs. I've opened a dozen tabs on a dozen windows and every time I closed them most all of the memory was released. So prove it. Download Sysinternals Process Explorer or something comparable for your OS and show us a screen shot that documents this claim.
for Linux .. how about a QT version ? GTK being old and broken it seems wasteful to code on top of it. .. the FlashVideoReplacer plugin does this, but its as crashy as anything... .. ..
how about an inbuilt ability to open flash videos with say, mplayer for a start
and in default settings - not to allow 3rd party cookies. or flash cookies. i know there are extensions for this
basic cross site scripting protection, and oh my god, the memory leaks
The only downside is extensions
I've been loving Firefox for years, but this fast release schedule is driving me nuts. Every time a new "major" version comes out now, at least one or two of my extensions break. The first one to go (on FF4) was Ubiquity, which still isn't fixed, and the stupid thing about that is Ubiquity is a Mozilla Labs extension. It's pretty sad when their own damn extensions can't even keep up, let alone 3rd party stuff.
So, back to your point about extensions being the only downside, honestly, do we use Firefox for any other reason? I could have ditched FF for Chrome or even IE9 (shudder) but it's the extensions that make Firefox so awesome, and that's what's suffering the most with this bullshit release schedule.
I still haven't gotten Firefox 5 completely back to the old 3.6 look and feel, which was more workable and required fewer button clicks. The last nagging issue is the one that Firefox no longer displays in the drop-down the history of links in the current tab, so you can't quickly go back to the top of a rabbit trail that you started down. Sometimes that was my only way out of stupid sites that disable the back button.
Oh, and the Federal Student Aid site (FAFSA.gov) only supports Firefox 3.5 and 3.6, one of which is no longer supported by FF and the other of which will also soon be not supported.
If you are not allowed to question your government then the government has answered your question.
As another person posted we really need an option to stop user installed add ons in schools etc...
I'm also using Comodo and I'm using it all the time now for the last 6 months.
I'd been a FF user since beta 1.0 days. I've grown sick and tired of memory problems and bells and whistles with FF.
It took me a few weeks to really get used to Chrome but its a much better experience.
All my extensions will break, and I'll have to manually edit them to work. They would have of course worked if the version number went up by a decimal point.
It's really worrisome, because it seems to be part of a much larger trend, that of FOSS software in general going downhill, and basically self-destructing.
Just look at the situation with Linux desktops, where both Ubuntu (by far the most popular distro) and Gnome have decided to abandon their users in an attempt to woo ADHD teenagers who do nothing but play games and browse Facebook. The other big desktop, KDE, hasn't gone the dumb-down route like those, but it hasn't really improved much in 10 years either, pursuing a radical rewrite which doesn't seem to have really improved the user experience or improved reliability or performance, but instead offers memory-hungry features of questionable value, namely desktop search and indexing.
The mainstay for Linux/Unix graphics, the X Window System, is also being abandoned in favor of a system that doesn't have network transparency the way X does, eliminating one of the giant features that makes X so great. X is full of legacy cruft and really did need a rewrite to get good performance on modern hardware, but the Wayland people have thrown the baby out with the bathwater in dropping one of the most useful features of X, so pretty soon Linux users won't be able to run applications remotely any more, they'll have to do it like Windows users, using RDP, VNC, etc. where an entire desktop session has to be started up and logged into on the remote computer and opened in a new window. No longer will sysadmins be able to open multiple apps from multiple servers and have them all display on one screen together.
Now Firefox seems to be driving off a cliff too. Before long, we're going to have MS dominant on the desktop again, even though the "desktop" may be decreasingly popular in the home and mainly used in workplaces and for a small number of power users, developers, etc.; and Apple and other consumer device makers dominant for consumer/home markets, making tablets and smartphones that "the masses" use to access the internet.
It's quite sad, because Linux and FOSS had a lot of momentum there for a while, and seemed to be making great progress. But instead of just being happy with that, and trying to get all the important applications and infrastructure to a certain level of maturity and then just going into maintenance mode and encouraging the devs work on other projects to fill in other gaps that exist in the FOSS landscape, the developers just couldn't leave well enough alone, and had to keep reinventing the wheel over and over again, much like their proprietary counterparts where companies want to keep adding more and more features (bloat) so they can convince users to keep paying for regular "upgrades". Obviously, it's not like this with all FOSS projects; the kernel just keeps evolving and adding more drivers (which is a never-ending task with new hardware constantly coming out), openssh hasn't changed significantly in ages, nor has the bash shell, my favorite monitoring program gkrellm doesn't seem to have anyone trying to revamp it over and over, etc. But the big projects just can't seem to help themselves.
What a mistake! Firefox 5 hasn't improved at all. The memory leaks are still there, and they're far worse now. I used it for a couple of hours, and its memory use was over 7 GB. Luckily, I've got 16 GB, so 7 GB wasn't that painful. Still, it was totally unacceptable for a web browser to ever use that much memory. The performance pales in comparison to Chrome, Safari, and even IE. It felt slow, while the others feel fast and responsive.
I notice you didn't mention konqueror. Does that mean you use (64bit) Windows? There is no 64bit windows version of FF5. Maybe you saw 7MB of RAM used and were confused? IIRC, a 32bit app can't address 7GB of RAM no matter how many bits the OS supports.
I'm always suprised when I see reports like this because on my computer (4GB RAM, Ubuntu 10.04 or Windows 7 Home Premium), Firefox and Chrome are pretty comparable memory-wise and Chrome is slightly faster, but only by a little. I very rarely have more than 10 tabs open, mostly documentation type stuff, not to many images or flash, and I've never seen a browser take up more than a couple hundred megs of memory. I'm definitly not calling bullshit on anyone who says this, but what I am wondering is if Firefox is just holding on to the memory to speed itself up and will it give the memory back if another program needs it. It's like in Linux, where if you just look at a simple graph on the memory your system is using it will always be around 90%, but if you investigate a little deeper in to the issue you'll see that most of that is being used as cache. It's very possible Firefox could be doing something similar.
Even with only 4GB of memory, I find I really have to go out of my way to start hitting the swap. If you have 16GB of memory, and the amout Firefox is taking isn't really hurting anything else, why not just let it have it? All it's going to do is make for a faster experience. The real question is what happens when Firefox's allocated memory starts getting in the way of other programs, it would be interesting to see some experiments testing that out because according to Mozilla themselves say they have most of the major memory leaks fixed.
Other issues could be poorly written plug-ins and bulky websites. I know there are a few plug-ins that allow you to manually clear out the memory Firefox is using, and can provide some more data on what exactly is being used.
Firefox 3.6.19 forever! I am now treating Firefox like an abandoned application. Google developers have now taken over. It may still be the best current browser due to its useful extensions, but it is like a bad copy of Chrome and imho inferior to Firefox 3.6.19 in most ways.
If I had to choose between Chrome and Firefox 4+, I really don't know what I would choose. Despite the horrible interface and all the badly implemented Chrome-ness Firefox 4+ still has unique functionality in the form of extensions like NoScript, Adblock Plus, and Scrapbook. They contain functionality that I just cannot live without and I haven't seen 100% replicated in any other browser. So I would probably be forced to stick with Firefox 4+ even though I prefer Chrome, Opera, and even MSIE in terms of the interface and usability etc.
Sure Chrome has NotScript, but it just doesn't work very well compared to NoScript. It's not a viable replacement. I ended up using the built in javascript whitelisting functionality which was a huge PITA. It was like going back to IE4 when I had to manually add sites to security zones by actually typing in the URLs.
If it some point a critical security flaw is found in Firefox 3.6.19 complete with exploits in the wild I may reluctantly migrate to Opera. Or maybe by that time someone will have forked Firefox 3.6.19 to at least apply security fixes as needed.
As of today Firefox 3.6.19 is still downloadable for Windows and Mac OS X and is available as a binary in the repositories of both of the Linux distros I use: ArchLinux and TinyCore.
Quite an experience to live in fear, isn't it? That's what it is to be a slave.
It was mainly Slashdot, Wikipedia, and some local news sites. It was some of the most basic browsing possible. That's why I was so surprised that it used that much memory.
I didn't even have Flash or Java installed, and I didn't bother installing any Firefox extensions, so I'm pretty sure that the problems are at the very core of Firefox itself.
Even 800 MB isn't reasonable at all. I've never seen the sum of Chrome's worker processes exceed 200 MB, even with over 50 tabs open. There's no reason to justify Firefox needing so much more memory than other browsers.
I came to your article from another article and am really interested in this learning about this. , I feel strongly about information and love learning more on this. If possible, as you gain expertise, It is extremely helpful for me. would you mind updating your blog with more information?
Moncler Shirt
No concerns for Tuesday
Well, guess I know what I'm doing on my next day off (Wednesday- don't ask). I'll be fixing all the backwards ass bullshit they seem to love including in these "rapid fire" releases. Disable compatibility checking, test everything to see it still works, yadda yadda.
This is getting old.
Sex. Drugs, and Unix.
Greetings fellow Dragon user! May I suggest a few extensions? Readability makes those stupid "hey lets split the content up over a bunch of pages!" crap DIAF by rendering it all as a single page, you can even adjust how big you want the text which is great for my older customers. ABP of course to get rid of the annoying ad crap, and ForecastFox (I know, but hey they kept the name) gives you a single icon which gives you the radar and 5 day forecast with a single click and will warn you if something nasty is headed your way.
I too had been on FF forever, and on the Mozilla Suite before that, but IMHO they are really going the wrong direction at Mozilla. Instead of concentrating on making FF a great solid fast light browser like what the original intent was they have been aping the hell out of Chrome instead of caring about under the hood and it shows. It is slow, buggy, sucks CPU and memory, all in all it just isn't as nice an experience as the Dragon which you've found out.
I love how the latest Dragon lets you use the Comodo secure DNS for ONLY the Dragon, as it has kept a few of my customers away from phishing sites and is faster than most ISP DNS services to boot, at least around here. It is nice little features like that that IMHO FF should be working on, instead of mearly going "me too!" when it comes to Chrome.
ACs don't waste your time replying, your posts are never seen by me.
For businesses and users worried about the fast update cycle of Firefox, perhaps they could consider using SeaMonkey. And, if they don't want SeaMonkey's chat and e-mail, perhaps they could make a fork of it, strip the unwanted features and... oh, wait!
I'm not totally opposed to the fast development cycle but I think Mozilla could step down the version numbering pace a few notches. e.g. release a new update every 6 weeks, but call it Firefox N.x+1 or N.x.y+1 instead of Firefox N+1 (Firefox 5 would be 4.1 and 6 would be 4.2). Push up the "big version" once each year or so.
Some fellow commenters have said it above: Mozilla don't listen to their users and ignore the obvious (e.g. huge memory consumption/memory leaks). Memory usage is better in Fx 7 and 8 thanks to the Memshrink project by Nicholas Nethercorte. However, they keep finding more and more leaks/inefficiencies. Some of them are caused by new features, some of them have been present for a long time, even with bugs filed about them. I hope they get their act together soon.
2019 is going to be the year of Linux on the desktop.
Weird, right now I have the same sites open in Chrome and it's using 14 Gb of ram. Firefox is using 150 MB for those sites. It has to be you that is the source of the problem
I'll stay on the stable 3.6 forever it seems.
I dont see the point of switching to the unstable branch that keeps braking things.
WTF. I'm still on FF3.x, yo!
***What's with the rash of major updates?!***
Within my 90 days let-the-suckers-go-first-buffer period FF4.x was eol'ed, right. It was said that FF5.x was the "security update to 4.x. And now, again, 5.x is to be supplanted by 6.x? Why bother updating, concordantlly I'll upgrade when shit breaks. Nice going Mozilla. *sigh* Or get on a periodic release schedule and stick to it.
I just ran leak test of my own on Firefox 6.0.0.4240 (as reported by Process Explorer). Here's a link to the image I mentioned above. The file is nearly nine meg in size and renders as 16,000 by 16,000 pixels. http://imgsrc.hubblesite.org/hu/db/images/hs-2003-11-a-full_jpg.jpg. Irfanview reports its rendered size as 732.42 MB (768,000,040 Bytes). I opened the image in Firefox, opened fifteen tabs on /. which drove Firefox to a peak working set of 2,404,128K. After closing all but this tab the working set went down to 234,000K. (System, HP G72 Intel I3 processor with 4GB ram, Win7 HPE 64bit.)
Why do new versions come out so often?
You seem to have some of your facts wrong. While it's true that Wayland itself doesn't have support for network transparency in the core, it's certainly possible to do network transparency with it. Protocols like SPICE and RDP will make it possible to send individual windows over the network as well as the full root window. There's nothing preventing this, and it will be possible in the future.
Nobody's saying it *leaks* memory. It just uses tons of it, especially on javascript- and video-heavy sites, and hangs inexplicably for seconds at a time. Loading up some giant images or only running 15 tabs isn't going to do much of anything in any browser.
So am I... This one change could make the web twice as safe for most users (and I'm tired of explaining to them which part of the URL is the domain name).
There's a lot of people using the word "leak" probably without understanding the contextual meaning. I've never denied that Firefox uses a lot of memory but even back when I was running nightly builds during the v3 development cycle I didn't see the kind of memory consumption often claimed here at /. or Mozillazine. Your post made me realize that my test was tainted. Sometime in the recent past I began using Noscript to keep those pesky scripts from shoveling malware onto my system. Noscript can be an annoyance but it is better than the alternatives such. Now; dol we blame Firefox for memory problems caused by Java Script or do we blame Larry?
I have to say ... IT SUCKS!!! It still has the annoying problem of not opening links until you click it multiple times.
The quality of Firefox is going to the garbage dump release after release. Instead of trying to mimic Google's quality less release schedule, how about actually doing some SQA and not releasing untested garbage?
Some plugins and addons use/leak lots of memory. I notice Firebug makes firefox bloat up a lot, so I generally don't have firebug or disable it.
Anyway I run chrome, firefox (and IE sometimes), and chrome uses a lot of memory too - the difference is chrome tends to have multiple processes, so you don't see one big fat process, but you get multiple processes which add up to about the same thing. But it's easier to free up memory with chrome - since if you close one chrome process, the memory is freed up by the OS, none of that "I hope Firefox starts freeing up some memory before I have to kill the entire browser".
Who is Larry and why would you blame him for javascript memory problems?
I hate updates when opening programs. I shut down the FF update service at FF 5.
It seems to me very much like Wayland is simple by virtue of making all the hard parts Someone Else's Problem. Want network transparency? Someone else will write an application to send individual windows over the network, and figure out a way of dealing with the fact that Wayland apps require OpenGL ES which doesn't run very fast in software and isn't designed to run over RDP. Of course, every application will have to add seperate code paths for it because Wayland uses the DRI device interface directly with no abstraction layer over the top - the developers don't care about network transparency and don't want to have to do any work to support it. Want to run your Wayland apps on anything except Linux with a modern KMS driver? Someone else will write Wayland replacements for the other platforms (because we can't be bothered dealing with bloat like portability and abstraction). Want to run them on graphics hardware that doesn't support OpenGL ES 2.0? Someone else will write fast software emulation.
Could it be that Moz is using the Intel compiler? Because my exp is the opposite of yours but I am using AMD CPUs. If one doesn't set very specific flags on the Intel compiler it puts out crippled code for all but certain Intel approved CPUs. I have also noticed FF seems to behave a little better on X64 than X32 which may also explain why you aren't seeing it.
I can say on the AMD nettop I'm typing this on FF is unsuitable for purpose and has been since the 3.6.x branch due to CPU spikes and memory suckage.
ACs don't waste your time replying, your posts are never seen by me.
When a program can install an addon, it can enable it, too.
So has anybody created a fork of Firefox yet that includes all the browser optimisations, faster JavaScript, etc, but doesn't bugger up the UI every month for no good reason (and ideally has a sane version numbering system)?
System Requirements: 32-bit Windows 7 / Vista / XP 128 MB RAM 40 MB Hard Disk Space
No luck for other systems' users.
... from the forgotten corner in europe
Firefox 6 has been already released unofficially (i.e. files are available, the announcement hasn't been yet made).
What, are they doing a major rev every couple of months now to catch up to IE's 9?
Anyone else here having memories of the Slackware/RedHat/Debian version wars of the 1990s?
Why did it take the developers this long to realize that people may not want a bunch a crap added to Firefox without their permission? Blocking the addition of features without express consent of the user should have been the default setting from the very beginning.
I have to close Firebug anytime a heavy webpage (a huge table with sorting, like in a Drupal admin area, etc.). If I don't it locks up for several seconds anytime I try to leave that page. About 25% of those times it just hangs until I have to restart it. It has made FF 3.6 look like the last version of FF for me... and I've been using it (or Mozilla before FF even existed) since the Netscape 6 debacle. Sorry guys - I hate the idea of Chrome as much as any privacy lover would, and I love certain addons (Adblock/NoScript/Firebug/etc.), but as a web dev FF4+ just isn't cutting it. Unless one of these new version fix those issues, I have to leave FF behind. It's now wasting my time on the job. BTW - Isn't it ironic that it's the same version number (4.x) where the Netscape code became too buggy/broken and had to be scrapped? Hopefully that's not an omen for Firefox...
What other systems? That covers every version of Windows still supported! If you mean Linux well duh, who is gonna do the extra work to support a 1% share OS that is a royal PITA to support? Do you support Apt or RPM? Will the version you just released work with the kernel that came out yesterday, or will you have to start over again because Linus Goatse'd the kernel for fun?
It is based on Chromium so it is FOSS, I'm sure if you look around their website you'll find the code. if you are on Linux I'm sure you know the compiler dance, hell I bet you have a compiler installed as SOP don't ya? So feel free to make a Debian or whatever Linux flavor you use version, just don't expect support.
Frankly though you really shouldn't be surprised as Linux is anything but stable ATM with all the changes from the kernel to the windowing system to the DEs all getting major overhauls. Comodo is pretty conservative with their releases and Linux is anything but conservative ATM. But the nice thing about FOSS is you can DIY right? So why don't you? That's what I got told when I asked for a feature or bug fix, so it must be SOP in Linux. Have fun!
BTW if you get tired of that kind of "fun" deals.woot! often has sales on Win 7 HP, usually $75 for a single or $100 for the triple pack. Quite an awesome OS BTW, and supported until 2020 so no worries there. Frankly even LTS can't hold a candle to the MSFT support cycles. And if you are on Mac I'm sure you have bootcamp like all my Mac customers do, just install it there. Oh and if you meant Windows X64? runs like a champ, I'm typing this on Win 7 HP X64 in Comodo Dragon, purrs like a big old kitteh.
ACs don't waste your time replying, your posts are never seen by me.
I don't know about you, but the last time I remember using the home button was in Netscape 3. Nowadays I just resume to whatever state I left the browser in.
The concept of a "home page" is obsolete since at least ten years.
Who is General Failure and why is he reading my hard disk?
The Wayland people seems to be focusing on making sure the core of the compositor is as good as it can be. Delivering what it has promised, a smooth desktop experience without tearing, lag, redrawing or flicker.
I believe we'll see network transparency in Wayland just after it has been adopted. The reason we don't see network transparency support in the Wayland core at this time it's because they are busy working on making sure that Wayland will deliver the smooth desktop experience and graphics.
Sorry guys - I hate the idea of Chrome as much as any privacy lover would, and I love certain addons (Adblock/NoScript/Firebug/etc.), but as a web dev FF4+ just isn't cutting it. Unless one of these new version fix those issues, I have to leave FF behind. It's now wasting my time on the job.
Fortunately...I found many of the same addons for Chrome that I was using in FF. Don't get me wrong...have been using and am used to FF...but when I have to wait for my quad-core to catch up to whatever I'm doing in FF...something had to give...so went with Chrome and the problems went away.
Whenever I read that FF gets their act together and it starts acting like Chrome/Opera/IE...I will start using it again.
Don't worry about the world coming to an end today. It's already tomorrow in Australia. - Charles M. Schulz
Kthx for the info and your kind comments, now please take a deep breath and relax.
... from the forgotten corner in europe
I'm very skeptical. With things like this, major features such as this generally need to be designed it at the beginning, they can't just be bolted on after-the-fact and expected to work well. Even if it's not implemented until later, it needs to be designed into the specification from the start. It's like designing a human launch system like STS or Ares and not bothering to design in the return trip until later.
Maybe adding some definitions in the protocol for network transparency would be nice, and make that plug-able, in the sense that we could use different back-ends for the network transparency, like X11, NX, SPICE, RDP, VNC, whatever.
So we are never stuck with a single and obsolete protocol like X11 that doesn't even perform that well. Wouldn't that be a good idea?
Sounds good in theory. Anyone here an expert on this sort of thing who could chime in?
Hi, I'm working with the creator of wayland on a google summer of code project to enable remote clients with wayland. This Is fairly conceptually simple, I just forward the information that goes over the main wayland socket, preserving the wire protocol, and on another socket, I use a RFB like protocol to pass on creation and changes to the framebuffer. Later on, I'll need a third socket to pass Drag and Drop data, but that is still a ways off.
If you have any questions come ask in #wayland on freenode. I'll answer what I can, and krh comes by frequently.
Thanks for the reply.
I think the big question that regular Linux users have is: how is my current ability to do "ssh -X" going to change when Wayland takes over? Right now, with any X application (which is pretty much every Unix graphical app), I can run applications remotely, these days usually over ssh for security. I don't have to worry about whether the application was written to do this or not, as every X application has this ability.
When Wayland becomes the default, how will this change? Will I be able to ssh -X into 10 machines and run 10 apps simultaneously, all showing their windows on my current desktop screen? From what I'm reading, the answer is either "no", "maybe it'll be done later", "it'll be up to each application" (which means it won't be done), etc. What's the real answer? There's a lot of serious Linux users, especially sysadmins and developers, who rely on this ability every day.
Now, if you can use some other more-efficient protocol like NX or whatever to improve network performance, that'd certainly be great, esp. if you could make the system support multiple protocols (perhaps negotiating which protocol to use, depending on which ones are installed on each end).
Is it going to be possible to choose the back-end to use for the network transparency in Wayland? like NX, SPICE, or any other protocol that performs better than others, I'd like to see some competition in then network transparent protocols. And if Wayland makes that plug-able it would be even better.
Thanks and keep up the great work. Wayland rocks.
Minor comment; but the Windows RemoteApp feature should do the same thing as X for running programs on mixed servers. Programs are run strictly in their own window; in a fashion identical to Windows XP mode - if you've used that.
I've only been working on this since May, and my understanding of the inner working of X is hazy, so I don't know how the finished product will work(right now I am working locally manually starting everything) or how ssh will play into it. But the effect is going to be the same; you will be able to run a client program running on a remote machine, while viewing it in the compositor running on your desktop.
This will not be built into the core of wayland. The instructions passed from client to compositor contain no contents of the frame buffer, since the FB is shared in memory. A separate connection is needed to mimic this shared framebuffer. Wayland offers a big advantage here, the intermediate software (that I am working on) knows when the screen is updated, and where. This means the least possible amount of data is sent.
I'll keep the code modular enough that, if needed, I can switch in another protocol, once I have a working prototype.
Right now, all I know is that the software I am writing knows what the compositor knows, and I just plan on sending the parts of the framebuffer matching the damage requsts. Any plugins for remote framebuffer sharing will have to work alongside the data going over the main wayland connection, so I don't know how that will work.
Your loss... can't say I'll miss you.
My Firefox has 37 tabs open, and a lot of extensions loaded(which will be disabled next time I start Firefox). It has 477 MB allocated to it. This does not seem excessive. I've had no problems with Firefox, except one or two sites with bad JavaScript, this wasn't much of a problem. It used to be a much bigger problem before 3.?, but still very rare in my experience.
Nice, doesn't network transparency belong to Qt/GTK though?
With your contributions, Wayland itself will be able to do network transparency?
I haven't implemented everything, but as long as a client doesn't send any references to objects the compositor doesn't have, it should work fine. I don't know how realistic this is, or how well it can be made to handle custom additions to the protocol. My contributions will be separate from wayland, more like wayland-demos, but it should be usable with other compositors; I don't know how varied the other compositors will be.
Wayland is just the protocol defined in a xml file. There is also a wire format specified in the docs, and there is no framebuffer information in there, that is done through shared memory. I am pretty sure wayland will always need something like RFB to mimic this shared framebuffer, in addition to the wayland socket, and another connection for drag and drop data.
Once you understand the internals of wayland, I don't think it would be a big step to make this remote client, so it is entirely feasible that Qt/Gtk could come up with their own.
I am really tired of them putting out major versions at this speed. Since these are obviously minor versions, why can't they just call them minor versions and stop breaking people's add-ons? I considered Firefox my browser of choice for a long time even though I disliked the large memory footprint, for one reason- Add-ons. ADD-ONS. The same ADD-ONS which they are now breaking as fast as they can. I've now switched to a webkit browser with far fewer features but less suck- Midori. You might want to check it out if you can't see the end of this Firefox crap. http://www.twotoasts.de/index.php?/pages/midori_summary.html
I have that problem with Opera, though to be fair, I have all of my email and RSS feeds going through it, and tend to average around 60 open tabs (I got lazy the last few years and stopped using bookmarks with sessions that save on exit); so, when it uses over a gig of RAM, I really only have myself to blame. Would be nice if it used separate processes for the tabs, though, and named them so it would be easy to find out which tabs are taking up all the memory.
Guess 4 Gigs just isn't enough anymore, time to upgrade to 8 or 16.
RTFA is Known to the State of California to cause cancer.