Unfortunately, looks like the CPU spikes are still a problem on OS X for me.
I check this every time a new version is released as I'd love to make the switch from Chrome, but Firefox with one blank tab and no extensions uses more CPU than Chrome with 15 actual tabs. When I actually throw activity at Firefox it hogs the CPU even more.
I unfortunately notice this behavior even when my MBP (2013, Retina) is in clamshell mode hooked up to a 1440p monitor rendering at native resolution. :(
My hackintosh with a dedicated GPU has no problem at all because it has the rendering power to handle it. But my 2014 MBP Retina with integrated graphics gets destroyed and ends up spiking the CPU as well.
I have the top of the line model from 2013 with 16GB RAM and discrete graphics (forced to always be used in energy saver prefs).
Firefox feels very fast and doesn't lag while being used, but the unnecessary CPU spikes kill my battery. We're talking about one page open at about:blank with no extensions and a completely fresh profile spiking to 20% CPU usage every 10 or so seconds.
I'm unsure if this is actually Firefox fault. I'm running Linux (arch and Ubuntu), Windows and macOS, and Firefox is only slow on macOS, while my macOS machine is the second most powerful machine I have. Running Firefox in a Ubuntu VM on macOS is faster than running it natively on macOS.
The discussion is all specific to that page. The biggest issue is that the animations are being driven from Javascript instead of CSS, so there's little we can do to optimize it.
The rest of the discussion is possible heuristics we can implement to optimize how we layerize the page. It would potentially help on this page, but might make things slower elsewhere.
Well the thing is it seems to only affect every other mac user even with otherwise identical hardware and os version.
If I had to guess I would chalk it up to a qc issue on some Apple specific chipset that just doesn't get hit by the other browsers code paths and just doesn't get noticed in other less used, less intensive apps.
How to diagnose it or work around it is beyond me.
Chiming in here. I would absolutely love to switch to Firefox for a host of reasons, ranging from "I don't want to use Google products" to "I love the tabs". Unfortunately, it's completely unusable for my workflow on a 2015 Macbook Pro with 16G RAM. I keep a variety of tabs pinned (JIRA, Zeplin, Gitlab, Jenkins, Invision, etc) and it seems as if they continue to hog a disproportionate amount of resources even when in the background, which is not an issue on Chrome.
> Unfortunately, it's completely unusable for my workflow on a 2015 Macbook Pro with 16G RAM
I run Arch Linux on this exact hardware and Firefox runs amazingly well there. It only runs like crap under macOS, which makes me think there's something with macOS specifically that's going on there.
I really want to switch to Firefox, so every once in a while I force myself to use it 'properly' for a week or two at least.
Such a two week period has just ended, and while FF does feel snappier than before, I'm running into exactly the issues you describe. When using Firefox, my four-year old MacBook Air pretty regularly slows down to the point where my music starts stuttering and even force-closing FF takes a while. I've had bluetooth devices disconnect too.
Now if it was just that, I could perhaps accept that maybe my laptop just can't handle my browsing behavior anymore. Unfortunately the same thing happens on my 16GB RAM Mac Mini. Not as often, but still.
I don't have these problems with Chrome, even when I clearly have too many tabs open for my memory to handle. Occasionally my music will stutter when I open multiple gmail tabs in Safari (I use Safari for stuff like email and banking), but it's rare, and nothing like what happens with Firefox.
I wonder what's causing that on your computer, because I'm running Firefox on a late-2011 MacBook Pro without any issues, and that's with a ton of tabs opened (including a bunch of pinned tabs running multiple instances of Office 365 in separate containers, Slack, and so forth) and the usual crowd of extensions (UBO et al). It's bouncing between 0 and 5% CPU utilization.
I've tried figuring it out, but haven't been able to pin it down. Is it even possible to check tab resource usage the way I can in Chrome? If not, that's another reason I'd probably keep coming back to the latter.
Thing is, it's not even that I have so many tabs open. It's always been fewer than Chrome, and with fewer active extensions to boot. Furthermore, it also happens on my other computer. The other one has 16 Gb of RAM, so it can handle quite a bit more, but the problem happens often enough that I also switched back to Chrome there.
I had to stop using FF on windows when it tanked because I disabled hardware acceleration (to force google maps to stuff their stupid 3d satellite images in the can). Look at the your rendering settings would be my random guess.
I experience regular "gray outs" on Ubuntu with FF hogging so many resources that eventually the only thing that works is killing it. This happens to me both on a 5 year old desktop as well as on old Lenovo Thinkpad (X2002s). However, I believe flashplayer to be the culprit as I often have quite a few youtube tabs open -- obviously never more than one playing though.
Why are you using Flash Player for Youtube? I've gone nearly 4 years since installing Flash on any machine, its pretty well unnecessary today. If you want to reduce performance though, using Flash is one way of going about it.
It’s usually the profile (some tweet years ago, some bad setting ) would be great if Mozilla had some solution for this other than zap your profile. Alternatively it’s an addon they have.
I'm in the exact same boat. Trying since a year to switch (back) to FF, but my testing period always end in frustration. Chrome is much more "snappy" in all regards.
I just don't understand this. Furthermore, I don't know if I've ever seen a "scientific" test of FF vs. Chrome using real world apps.
At work, I have several GUIs in browsers that I access through Chrome, and they cause instant 3+ GB of usage. I can't test FF due to group policy, but anyway, Chrome can be a hog, too.
Are you saying that those same half dozen apps, when pinned in Chrome, do not cause any RAM issues for you?
Reading so many comments like this confuse me. I run an older mac, and I find firefox very stable. It's with chrome that I have all the problems you're describing. Every time I'm forced to use Chrome, I'm reminded how unusable it is. But I'm in the minority it seems.
Strangely, none of my coworkers seem to have problems. But every release seems to get worse; first it was terrible performance, then fonts in pages and the UI became blurry. This happens even on newly-installed MBPs, so it’s doesn’t seem like it could be related to anything I’ve done. And yet… coworkers don’t have problems.
Why do you think a piece of software running on your laptop would behave differently from millions running on the same hardware.
Profile, addon, custom install location, weird env variables
Yep, regardless of whose "fault" it is — and I totally buy that it's macOS's fault, considering how much effort my startup Cyph has to spend on working around Safari/iOS bugs — at the end of the day Firefox is unusable for me if my battery won't last for any reasonable amount of time and I have to deal with a kernel panic every other day.
Really hoping this gets sorted out soon, because on most other counts I prefer the current Firefox to Chrome. (The other thing keeping me on Chrome is pinch-to-zoom support.)
Maybe related: Does anyone have the issue of a few number of FF tabs showing up as several in task manager? Say I have 2 or 3 tabs open, it'll show something like 7-10 in task manager. And yes, they are hogging a ton of RAM, upwards of half of a 16GB setup
At least good to hear I'm not the only one. I want to like it for all of its speed and sleekness, but it regularly runs hot and gets the fans going full blast. Odd that it's still an issue.
I have the same issue on my MBP Late 2013, but on my MBP 2017 everything works smooth. Not as smooth as chrome, but nothing really noticeable. I always have 40+ tabs open.
I reported it and they know what the problem is but it got marked as no-priority. Everytime there’s a new release I check if it’s solved and move back to Safari. Every release there are people complaining about it. I know I have been for 2 years now, but it falls on deaf ears.
The enterprise part is great to see (though it was something that seemed long overdue). I recently noticed on a Firefox mailing list that Mike Kaply (of Kaply Consulting, a company that helped with enterprise deployment and configuration of Firefox) had officially joined the Firefox team. Mike's presence is going to make Firefox in the enterprise much better.
On Firefox 60, I'm still not on board with the newest versions after the support for legacy extensions was removed. SessionManager, an awesome (legacy, XUL) extension, still doesn't have a perfect equivalent in the Web Extension world. Tab Session Manager, which has similar functionality, seems to be lagging behind and struggling with issues in Firefox that prevent it from becoming a good session manager.
If there's one thing I could ask the Firefox team, it would be to focus on enabling web extensions to do almost everything that legacy extensions were able to. Without the power of feature rich and stable extensions, Firefox is currently inadequate for me (though I still use it as my primary browser).
Thanks, I did see the Session Sync extension before (and again now after the challenge winners were announced). Unfortunately, it seems to be quite thin on features, with the man advantage being syncing the sessions easily since they're stored as bookmarks. I also found the UI unintuitive (for example, I have to click on the extension icon, then go to the history tab, and then click on the current session to see the Save button).
Tab Session Manager is comparatively a lot richer in features, and also allows importing from the (legacy) SessionManager sessions.
On the one hand, the FF webextension API is under continual development [0]. On the other hand, "almost everything" previously possible does not seem not a realistic expectation since it'd mean the ability to change almost every single atom of browser behavior.
Considering that Firefox pioneered the extensions concept among the masses and had a wide variety of terrific extensions (and still has) as a key differentiator, and since many users have appreciated Firefox for this reason (even if they didn't care much about FOSS), Firefox must continue to push heavily on this front.
Of course, everything that was possible before wouldn't make sense in the web extensions platform (for security and other reasons), which is why I said "almost everything". But things like tab management (even Tab Mix Plus on web extensions is only in beta now), session management, etc., are highly important to the browser UX for a lot of Firefox users. They're highly sticky factors as well that prevent users from switching to another browser.
I've also seen some people who deploy and use Firefox in enterprise state that they're going to be on ESR 52 for quite sometime, with one of the reasons being some extensions not being available (or ported) on web extensions.
I'm not going to blame Mozilla or the Firefox team for the current state of affairs. But knowing what we know (with the advance notice on web extensions), the main advantage of Firefox on the extensions front shouldn't be lost. If anything, this piece needs to be accelerated further and also have the Mozilla Firefox team drive new features and abilities in web extensions that make other browsers look outdated.
I'm a die hard Firefox fan, and would always be cheering it on to thrive and grow.
>Bookmarks no longer support multiple keywords for the same URL unless the request has different POST data
Can someone explain this one to me? Does this mean I can't use "gi", "i", "image", and "is" as my Google Image Search keywords and need to pick one? Reading the issue [0] isn't helping me. I only recently updated to FF59 from FF39, having to abandon most of my workflow. I really hope this update isn't another nail on the stairs for me, seems every single update since FF40 has broken a significant piece of my workflow. I'd appreciate knowing before updating and then having to roll-back.
> But people expect that it's the specific bookmarks that have keywords, not the urls+postdata, because it always worked like this.
> The UI also shows this. There's a keyword field for a bookmark. And it's not labeled keyword for url/post data. This is a discrepancy, and results in unexpectedly overwritten keywords.
I'm guessing the title is supposed to be read "(multiple bookmark) keywords", not "multiple (bookmark keywords)", and the release note restated it wrongly.
Like, if you have two separate bookmarks with different names, but they go to the same URL, they share whatever's in the keywords field.
I had someone on IRC confirm for me before I updated, multiple keywords for the same URL (eg: youtube.com/?q=%s) still works. I'm still not 100% sure what is meant by this line though, which is what caused confusion for me:
>It is no longer possible to have multiple bookmark keywords for the same url, unless the request has different POST data.
I guess I'm confused what "POST data" is - as it apparently isn't %s and I don't know what else you can change for a keyword bookmark so I'm not sure how you'd even change what the POST data is.
"y" and "yt" both work to navigate to "youtube.com" and "y search" and "yt search" both did a youtube search for "search".
I'm so excited about the future of Firefox. In addition to the above, OpenBSD is currently working on pledge support in FF. In a couple years FF could be simultaneously faster, more private and more secure than Chrome.
I use FF on OpenBSD as my daily browser. I have at least three different profiles running simultaneously with multiple tabs open in each (though I don't go nuts with tabs like some do -- I never tend to have more than 10 or so open).
It does occasionally seem to get stuck in a state where it just starts chewing CPU cycles and RAM. I've seen this mostly with Google sites like gmail or docs, so I assume it's JS-related. But it's fairly rare, and seems to be improving with each release.
The other problem with container tabs, which isn't really a Mozilla problem as much as a Google problem, is that if you have enough segregated google profiles going you seem to hit some google anti-abuse codepaths or aomething. I have 3 separate Google accounts separated with container tabs (on purpose, I know Google supports multiple accounts), and a separate regular browsing container, and occassionally just following a link from Gmail or hangouts where Google bounces you through their redirect hangs for 30 seconds or so. When it's doing that it generally does it consistently though.
I'm unreasonably happy that Native Module support is now enabled by default in Firefox, Safari, Chrome, and Edge!
My side-projects and experiments (like the WebGL2 "minecraft-y" demo - https://mrspeaker.github.io/webgl2-voxels/) now work out of the box: no build system, no transpilers, no dot-files. Happy day!
Native modules have brought the fun and immediacy back to development for me. I never quite got used to the delay and boilerplate of build systems, despite their power.
But performance seems a little inconsistent in production. For iOS Safari in particular, I've had apps take over a minute to load via modules, versus seconds when bundled. So I stick to rollup for deploy (maybe http2 is supposed to fix this?)
If anyone from lastpass is reading this thread, please fix your plugin. Everytime I try to edit the password from the prompt on the page, it opens a blank page for the password vault page.
I love lastpass, I'm actually a paying user, but it's getting more and more buggy. Sometime you generate passwords to change an existing one, and it doesn't update, so you only have the old one. Sometime autocomplete doesn't work: either the menu doesn't do anything, or the first click on the input icon doesn't work.
It's not too bad, and honestly the mobile version is a god send (it fills passwords in apps !), but it's irritating.
Agreed. But LogMeIn purchased them, so I'm unsurprised that price has increased and quality has dipped.
One issue I've been seeing more of (both on Android and on desktop) is the cache not re-populating automatically. You have to go Advanced -> Clear Local Cache, and only then will it download a fresh copy.
Or worse, sometimes after adding new credentials it never sync's it to the cloud. It remains on that device, but that's it.
Anyone else get the bug where it doesn't auto-fill (shows grey box), you right click and it shows that you have no stored logins, and then once you close that (right click) menu you are prompted like normal? This happens to me frequently, though seems to be specific sites.
Mine's worse. Literally every other login fails. I don't get a rejection or anything--the box will light up and turn red--but the vault will appear empty and none of the sites will be found. I have to log out of it and login a second time (meaning put in my giant master password), and only the second time will it find the stuff in my vault correctly.
uh. i thought i was the only one. sometimes i have lastpass tell me that i have no internet connection -- while i'm trying to login into lastpass to login into another site.
Seconded, the difference in quality between the Chrome extension and the Firefox add-on is palpable. I hadn't noticed a difference until after Quantum as released, and it seems to be getting worse.
I think they've removed some app permissions on access to the clipboard in Firefox. LastPass seems to be notoriously bad with copy-pasting and empty-page loads.
>Added a policy engine that allows customized Firefox deployments in enterprise environments, using Windows Group Policy or a cross-platform JSON file
Hopefully this means the two gigantic companies I work for will phase out Firefox ESR for Quantum. I also hope they move from 32-bit Firefox to 64-bit, but I've learned to keep my expectations low when it comes to these places progressing with tech.
>Added support for Web Authentication API, which allows USB tokens for website authentication.
That's awesome. Ok. Maybe I'll let myself get a little more hopeful. :')
Firefox ESR 60 has been released at the same time, and is more or less identical to the non-ESR version. ESR makes a lot of sense for corporate networks, even if it can be a drag being a bit behind the curve.
Sure, ESR does make a lot of sense. According to Mozilla's timeline[0], enterprises will have to move to Firefox 60 ESR after August 2018, because the 52 ESR will be EOL after that time.
Pocket doesn't collect any browsing history. From the linked docs:
> If the “Recommended by Pocket” feature is enabled, Pocket will send a list of the best stories on the web to Firefox every day. With each story, Pocket also sends a list of related websites that, when visited, signal likely interest in the story. Your Firefox browser compares your browsing history with the list of related websites to sort and filter through each day’s stories and recommend the ones that are most likely to interest you.
> Important Note: Neither Mozilla nor Pocket receives a copy of your browser history. The entire process of sorting and filtering which stories you should see happens locally in your copy of Firefox.
The feature isn't "EU-excepted", it's "US-only", probably because they only have recommendations for US audiences.
I have no idea if they have a bias on what they send to Firefox, but their approach (send stories with a list of other sites that signal you might be interested) is obviously going to facilitate a filter bubble. Is your usual filter bubble right wing? If it is that suggests a bias on their part, otherwise maybe they're just amplifying your left wing browser history?
I have that setting as well. The new config specifically for this is "browser.newtabpage.activity-stream.showSponsored", you might want to flip it too just in case. The config name doesn't specifically reference pocket - so perhaps it could be used in other ways as well.
Does Pocket Sponsored Stories violate privacy? It doesn't seem like they track the user or anything like that. And, Mozilla claims their approach is "Privacy-Conscious".
Not sure why this comment is grey, FF is now adware by objective definition of the word adware.
Wiki >Adware, or advertising-supported software, is software that generates revenue for its developer by automatically generating online advertisements in the user interface of the software or on a screen presented to the user during the installation process
Because by that definition Firefox has been adware since Firefox 1.5, released in 2005, due to the ubiquitous "Google" logo in the search box that every user sees by default upon installing the browser, and for which Firefox receives the bulk of its revenue.
That's significantly more defensible (it has to search somewhere by default, and most engines are known by their logos, and Google is the only search engine worth a crap) than imposing paid sites on you by default.
Search has been a basic part of using the web for years now. All browsers support the OpenSearch standard. Now, you may feel it would be better for a user to have to manually add Google to their search bar, but relegating search entirely to an extension is an odd suggestion.
Can't speak to the greying as I have no part in that. That said, while I agree to a point, to me, adware is where ads are being served that are impacting the UI proper. Because of where it happens, apparently in a new tab/window, it isn't adding advertisements when I go to a web page that weren't already there. I also don't have to do a specific action to get away from the ad: I don't have to close the ad directly, it just disappears as a result of me doing normal behavior. So while perhaps "adware" it isn't the annoying kind at the very least.
I don't use Firefox so I'm not sure what it is showing, but if it is like the new tab in Chrome that pocket's extension generates, it is basically the top 3 in the recommended section of Pocket which seems to be popular articles of no particular importance or targeted. Earlier I saw links to an article for a Ryan Reynolds interview where he talks about his anxiety (apparently, didn't read it) and another article about all cultures appropriating (again, didn't read, so going on title). If it were targeted, I'd expect articles of a more technology/development slant.
Mozilla owns Pocket. They're not calling them Pocket Sponsored Stories to be cute. You, too, can now buy screen space in the premiere open-source browser: https://getpocket.com/sponsor
I wish someone with the time would start producing a "De-Mozilla'ed" fork a la Chromium. I think Firefox has gotten to that point, things like this and the Mr. Robot fiasco indicate the project needs to be saved from itself.
The basically idea is not to "fork" Firefox per se, but to maintain a set of patches that are appropriate for a user-respecting privacy friendly browser. If a true fork one day becomes necessary, this would be a nice starting point. For now, I'm just trying to completely remove all bad code.
You do have to build the source yourself, however, since binaries can't be distributed without violating Mozilla's trademark.
Well, it's trying to be. It's still millions of lines of code and tens of thousands of design decisions made by Google.
The same would be true for your wish of a Firefox-fork, though I'm not sure there's that much sense in it, since what Mozilla has been doing has not infringed on privacy. You'd be patching out things that are harmless.
An extension popped up in the extension list, which did nothing at all, but had a weird name that made people panic they had somehow gotten malware.
And it was pushed as part of an unpaid marketing campaign for Mr. Robot. Which people thought was Mozilla cashing in big time by installing an extension they didn't want. It was done on a friendship basis. Mr. Robot has been pushing Firefox and so Mozilla wanted to give something back.
Which was a fun little easter egg that Mr. Robot fans could have activated. It would have flipped some words on webpages upside down as reference to Mr. Robot. Which, again, it didn't do, it did nothing unless you specifically went into about:config and activated it there (which is what Mr. Robot fans should have figured out themselves).
This also resulted in a discussion of how it's unthinkable that Mozilla can install an extension without asking you.
...in their browser that people are ok with them auto-updating, through which they can push even less restricted code than in an extension, and that completely invisible.
So, it was a fuck-up on Mozilla's part (in that the extension was visible in the extension list; if that wouldn't have been the case, it wouldn't have been a problem at all), but man was it blown out of proportion by journalists and know-it-alls, who just have to show that the innocent-thought Mozilla is evil.
>This also resulted in a discussion of how it's unthinkable that Mozilla can install an extension without asking you.
You snark, but at the end of the day, they really shouldn't be installing stuff silently and without affirmative consent. The character of the stuff is irrelevant to this principle.
This is the first post I've seen that defends CDPR wholeheartedly. From everything I've heard, it's an overly broad, draconian code that is written to destroy any company the EU wishes.
I'm glad that they have released support for the Web Authentication API. Hopefully I won't need to use Chrome for websites I choose to be more secure with.
Although, with a quick look it seems like I still can't use U2F with Google on Firefox.
But just yesterday people tried to "educate" me that it works with 60 and for 59 I had to toggle an about:config switch.
It's great that they support the proper web standard. It would be useful if all those web sites supported it. This is a long-known issue, and nobody cares about non-Chrome browsers.
I’m probably the one that will implement the Web Authentication API for FastMail and Topicbox, though it’s a some way down my list of things to do at present. I looked into it a couple of months back (I would have liked us to get it out before browsers enabled it by default), but documentation was very scarce, and so it wasn’t particularly clear what we’d need to do to migrate from u2f.js to webauthn (especially while still supporting both), and then other things came up. Since then, https://www.imperialviolet.org/2018/03/27/webauthn.html has been written, which will help, but it’d still be nice to have a concise “here’s what to do, backend and frontend, to migrate from u2f.js to webauthn” guide. If no one has by the time we support webauthn, I’ll probably write such a guide.
For now, it’s not as high a priority as it could be, because the functionality it provides is already available in Chrome (for that matter, I don’t believe Chrome’s webauthn implementation hits stable until the next release), Firefox can get it by enabling security.webauth.u2f which is good enough as a short-term measure when it’s always been required in the past anyway, and Edge doesn’t have many users (and few of them currently do 2FA). It’s pragmatism, sadly.
Alas, I'm completely unable to get the soft token to work (after enabling it in config:about). It even fail on https://u2f.bin.coffee/ which seems to claim it should work.
It's a usability trade off; I have a physical key, but I'm asked to authenticate 20+ times a day which make it a pain on my port-limited MacBook Pro.
Switched from Chrome for privacy reasons a couple of months ago. As a webdev (and hence heavy DevTools user), it feels excellent, the inspector's a little bit faster, and having containers for Facebook and Google tracking services is excellent. Spell check in text boxes doesn't work though - apparently it should, I have no idea why it doesn't.
Yeah, I've made some big reports already (CSP errors from undefined source when using Stripe with LastPass enabled) but I want to check it out more before bothering anyone - maybe it's my Windows language config
For spell check, I've noticed I have to right-click and check spelling. I actually kind of like not having the browser second guess me, but not sure if this is an intentional default on FF's part.
Firefox overall performance starts to be pretty good and it's
pretty pleasant to use. It feels smoother than Chrome on 2015 Macbook Pro.
But I guess there's a long tail of issues that keep some people from (still) switching to Firefox.
Mine is lack of smooth pinch-zoom on macOS. Once that is fixed, I think I'll be using Firefox most of the time.
On 2018 Android flagship phone, scrolling speed seems to be maybe 58/59 fps. Occasional jerky motion sticks out like a sore thumb. I think it might feel smoother steady 30 fps, because at least then all frames take equally long to render.
I'm psyched to see the fix to https://bugzilla.mozilla.org/show_bug.cgi?id=1193394 released, which (amongst other things) mean that you can finally write code using both IndexedDB and promises without inadvertently committing the active transaction on promise resolution. Now I just need to wait until the market share of other Firefox versions decreases before I can actually take advantage in production :)
Firefox 52 may be around for a while - it's the last LTS before extensions were broken by Quantum, and it's also the last release that runs on XP and Vista.
That's a minimum, but I haven't heard any noise from Mozilla about pushing it any further. Either way I'm sure some third party will pick it up and it will live forever. The only question is the % of people who switch.
In case anyone else is getting a "What's new" page after the restart that is just a spammy and broken Firefox Accounts advert, you can get to the more normal release notes via the About Firefox dialog's "What's new" link, or just go here:
It's off by default. To enable it you have to go into about:config. I'm saying it should be on by default, or at least exposed in the Settings for people to find. The concern about doing that was that some sites would break, but their definition of break included things which should break because they were previously abusing the lack of FPI. I have been browsing with FPI on for a while and nothing has broken. Protecting our privacy should be the default state of affairs. Mozilla did say Firefox would be opinionated...
Mozilla did a study with user testing. Breaks lots of federated logins / Single-sign on solutions. Too much breakage to enable it by default. If you confuse users with this, they switch to GChrome and all you did for them is a privacy disservice.
There's a First Party Isolation add-on though if you want a reachable toggle button.
I have it enabled and I can attest that it breaks a fair amount of sites. I think it’s fair to expose that as an expert level setting. It could benefit from a bit of advertising though.
This is extremely easy to do in uMatrix. If you don't want to use the other features, just set a default policy of allow all, then block third party cookies. You can then selectively allow access to cookies on a site by site basis.
Good point. So the technology is there, and an "expert" UI.
It would be neat to see Mozilla create an intuitive UI for this to enable first party isolation by default, like a permissions model for other things. A modal popup "Unbury.me wants to grant access to google.com", for example.
I've been using Firefox Nightly as my daily browser for a while and I am beginning to notice more and more websites breaking on it - mostly Google services, although Facebook as well, recently, too. I know it's complicated by the fact that I'm on the Nightly channel, but has anyone else noticed an increase in sites that are incompatible with FF?
has anyone else noticed an increase in sites that are incompatible with FF?
Yes, but I'm not sure it's just Firefox. I suspect a significant number of sites are now being developed and tested only with Chrome, using Google's developer tools and recommended practices and browser feature set. Unfortunately, those things may not be supported or work well on other browsers, and if you don't choose carefully and test properly...
I've noticed this while using the last few stable releases of FF (on Debian 9.4, on the odd chance it's relevant). Off the top of my head, the (popular) sites I regularly have issues with are: ebay (creating/editing listings), LinkedIn (have to refresh the homepage multiple times before all of the components appear and are usable), GitHub (WYSIWYG).
I've just assumed these issue are caused by add-ons behaving badly, but I haven't bothered to get scientific about it yet.
Nightly has been my daily driver for years, with a couple of brief interludes. There’s the very occasional possibly serious breakage, but I haven’t noticed any compatibility troubles for a long time. (But I don’t use Facebook, and of Google I only ever use Maps.)
Actually, I suppose there are a couple of sites which Tracking Protection completely broke, but that’s not Nightly-specific.
I'm using Firefox on Android - not sure what friction you're referring to here, but the lack of 'reload on pull down when scrolled to top' feature is seriously annoying for me.
Despite this, I stick with Firefox because I want to use it - I can imagine that the vast, vast majority of users who have no opinion on which browser they use simply don't/won't put up with this and just go back to the easier to use Chrome. It's little things like this that make all the difference.
In every other Android app there's no bound on how much scroll momentum you can build up; if you want to zoom to the bottom of a long page, you can keep flicking to go as fast as you want. Firefox mobile adds an obnoxious amount of extra friction that prevents you from going much faster than your finger moves, making scrolling feel like wading through water.
As an aside, I think iOS might have the same behavior, which could be the reason they're doing it, but it goes against the system standard and is completely unbearable for me.
I use Brave browser, which on desktop is a little slow, but on mobile is just a Chromium fork with some extra privacy features (including ad blocking). I'd certainly like to support Firefox in principle, but Brave is a good substitute.
Oh, good, I thought you were trading a bit of janky scrolling for an ad infested web :) I've never noticed the scrolling issue specific to Firefox but I feel like everything on Android has a few hiccups like that. I've always chalked it up to JVM garbage collection. I'd be surprised if Brave didn't have a few UI quirks too.
Edit: Just read your other reply, I thought you meant frame skips or something. I don't hit many long pages on mobile so I've never realized it did what you were describing. I wonder if they even know about it.
uMatrix is excellent; more capabilities than RequestPolicy with a much better UI. Written by gorhill (Raymond Hill) who also made uBlock (and still maintains uBlock Origin); rock solid, low resource.
It's not for non-technical users, but if you are looking for RequestPolicy or NoScript I'm sure you'll manage.
> Firefox's new parallel CSS engine — also known as Quantum CSS or Stylo — which was first enabled by default in Firefox 57 for desktop, has now been enabled in Firefox for Android.
That's exciting, but on my phone the Play Store still doesn't provide the new version.
One killer feature from Safari missing still is 'Show tab overview' with thumbnails of all your tabs ... anyone know of a well maintained plugin that does something like this? I think it should be built into the browser.
It's in the Preferences under General -> Tabs -> "Ctrl+Tab cycles through tabs in recently used order". Right in the middle of the screen, when you open the Preferences-page...
For such a privacy minded browser, it still preserves closed tabs in Private Mode. If you do (in OSX) cmd+shift+T, you can reopen the closed tabs from earlier. This does not happen in Chrome nor Safari.
Aside from continuous memory issues which others have mentioned, my biggest issue is that we still cannot have access to core FF pages and elements with webextensions (ex: mouse gestures on blank/new tabpages, userscripts on RSS feeds, scroll through tabs with mouse wheel, etc...)
Or better, opt in to not pause; I hacked up a -STOP for webkit (surf/glide) when the tab/window was not in the foreground. The X plumbing was tricky and never quite right. Still on my TODO to finish.
Running 60, installed and configured the group policy templates. Still can't get Firefox to pull trusted root certificates from the Windows certificate store. #headdesk
My coworker likes to say "None of this [tech] shit is cool" and that's kind of how I feel about browsers.
Firefox Quantum is quite nice but in the past, I've had a constant resizing bug that makes it unusable on my home laptop plus the numerous issues listed below.
Alternatively, Chrome is neat but I find it runs sluggish after a while on my older Macbook. I wish it had container tabs a la Firefox and the upcoming "disabled autoplay by default" change is really annoying.
I feel the same about OSs anyway but it's easy to complain rather than contribute bug fixes. Mind you, I don't know eg; C++ so
Ah, that’s what autoscrolling is, is it? I wish it had some explanation of what non-obvious preferences are, at least a link to a document explaining it.
It will never be ported fully. It's one of those extensions that dived so deep into Firefox's source code that Mozilla changing things in the source code caused it to break a lot of the times.
Mozilla categorically doesn't want that anymore, as it will inevitably cause some Mozilla devs to try to not break those extensions and therefore not do things such as code refactoring. The sort of stuff that you can ignore in the short-term in order to not break extensions, but which causes huge problems in the long run.
Firefox for Android will be copying Chrome's scrolling physics for faster scrolling, since that's what Android users are used to. Chrome doesn't use Android's standard scrolling physics, so custom scrolling code is needed to emulate its behavior.
You can test this in Firefox 61 Beta by setting the "apz.android.chrome_fling_physics.enabled" pref to true in about:config. The new behavior will hopefully be enabled by default in Firefox 62.
Ah, I was wondering why my scrolling was getting "fling"ed so much on Nightly. It's kinda unusable because 2/3 of the time, when I lift my thumb off the screen, it gets interpreted as a fling. Maybe I'll get used to being very careful, but for now I'm glad to know there's a toggle.
There are still fling bugs to be fixed. :) Feel free to file a bug in Bugzilla that blocks bug 1448439. Perhaps you are seeing one of these known bugs:
From the comparison of the lists, looks like desktop is still their focus, which points to a sad level of navel gazing for an internet company in 2018. Makes me question whether they are focused on competing with Chrome or whether they just want to make cool tech.
The desktop (and laptop) is still my focus, too. It's where I do all my work. I'm glad someone still gives a damn. I'm glad not everyone is chasing the crowdsourced social-mobile-AI-blockchain. Not all tech has to point in the same direction.
That said I use Firefox on Android every day and it's fantastic, so what are you talking about specifically?
Firefox has major issues with scrolling and also with a lot of plugins that claim to be available on mobile but simply don't work. They need to do a marketing push because nobody wins when one browser has monopoly share on an OS, especially if it's the default browser.
Desktop is where nearly all Firefox users are, all of whom immediately get the benefits of improvements to Firefox, and whose use of Firefox funds all Mozilla.
I haven't tried Firefox since version 47 and haven't used it as my main since 37. I gave 61 (dev edition) a spin yesterday. All the extensions that make the browser usable for me were like 1990s geocities pages: "Under Construction" if they existed at all. They were just placeholders that lacked most features with links to bugzilla pages to vote for features to be added back into Firefox.
Garbage collection was pretty terrible too. I managed to use up 4 GB of ram with ~10 tabs. In my normal (firefox fork) browser I run 300 active/loaded tabs and 700 'suspended' tabs for ~1k total and never get that high.
There's seemingly still a lot of work to be done before it's usable as a main browser.
I know everyone has there own work flows, but how can you even keep track of that many tabs?
After about 8 or 9, I find myself tabbing through half of a dozen tabs to get where I'm going, and at that point I realize I've added like 5 new tasks to my original stack of todos or research or whatever I was originally doing. At that point it's time to shut some tabs and concentrate on the original item.
Also, once I've logged into, for example, GMail or Facebook, I do not want to be browsing other sites while under their eye. I instead usually open multiple Chromiums / Chrome, each app for a specific task.
> I know everyone has there own work flows, but how can you even keep track of that many tabs?
You don't.
You just push new tabs onto the stack. When you're done with a topic, you start popping those tabs until you're back at the previous topic.
Then that previous topic is either something unrelated which you wanted to read later, so you just leave it there. Or it's whatever you were researching before you had to look up something that's mentioned in the text, meaning that you now know this thing and can continue working on the previous topic.
You never have to know what your tabs to the left contain. You only care for the last handful of tabs to the right, which you can use as if those tabs to the left wouldn't exist (well, on Firefox at least, on Chrome they become unreadable).
I've been doing large sessions ever since Opera introduced tabs in my teen years (2000). It's how I grew up and second nature now.
Tabs are great because they provide context both as a span of tabs showing how I got from point A to point B and as individual tab due to tab history backwards. I've been regularly saving my session files since ~2001 although I only have data from 2003 onward. It's great to load up an old one and see what I was doing on some old date.
Nowdays the first couple hundred are almost static over a year or two. They are things I always want to have open. The most recent ~300 or so are in flux from week to week. There is a gradient. The older tabs and their favicons/etc form a landscape I know and can navigate just by a glance. The newer ones I can see the general topic of a group/span of tabs by clicking in and using tab search. I also use minimal browser skins/layouts to pack the max amount of tabs in.
Every couple months I'll go through and "clean" up a session by closing some and bookmarking others. The tabs that make it through this are added to the static landscape of useful things for current projects. Kind of like how sedimentary rock is layed down in layers.
Every year or two I'll start a new session. It always feels so directionless.
I do not use 'web app' sites like gmail or Facebook.
> I find myself tabbing through half of a dozen tabs to get where I'm going
Why are you tabbing at all? If you type "* foo" in the url bar (without hitting enter), you will get a list of tabs whose url or title includes "foo". You can select one to jump to it. Much faster than linearly going through tabs.
That ram usage seems suspicious. I never left firefox, have hundreds of tabs open and never really go over 1,5GB Ram usage. I think you should investigate that further. Maybe one of the pages you open really leaks memory or an extension is at fault here?
What fork are you using and do you use any extensions to manage that many tabs? I have a similar workflow on current Firefox but I find managing that many tabs difficult. I used to use tab groups for this.
Do you ever use any of them or get back to them? I used to do that too on a smaller scale but realized after a while that I don't really need them to find things as the browsers address bar is usually enough and worst case I can fall back to a web search.
I never really had a case where I couldn't find something again that I would've found through my bookmarks as these are usually not tagged well or there's not enough information in the URL / title of the bookmark to have a big surface area for a fuzzy search.
It took me a while to realize that but I noticed that it's very refreshing to keep less things around and try to hoard them (bookmarks, browser tabs, downloaded movies, shows, music) like I did in the past. After observing it for a while I realized that it never happened that I couldn't find something again or re-download if necessary.
I use my bookmarks constantly and I have a well structured folder hierarchy so I don't run into the problems you describe. There is no downside to having lots of bookmarks. It isn't the same as a physical space in which the 'clutter' or 'less is more' ideologies may have merit.
I respect that. It doesn't work for me but it's an interesting thought. Props on being able to arrange it successfully. Says some positive things about you by extension.
I only keep bookmarks that are really important or used often, the latter which doesn't really make sense but when I want food I open up my Food folder and looking at places around me helps me pick something.
It was usable as a main browser at some point between 4.0 and the current quantum (read - when 90% of the useful extensions were axed and the only advantage over chrome was thrown out)
I can't tell whether it's mostly web 2.1's fault or Firefox's gross mismanagement and stubbornness
Firefox still has an edge over Chrome here; Mozilla continues to add new Web Extension APIs for frequently-requested functionality. For example, this release added tab hiding, making it possible to build extensions that add Panorama tab groups.
Anyway, maintaining the old extension API wasn't sustainable. Many older extensions broke repeatedly due to refactoring efforts like Electrolysis. These refactors ultimately resulted in some of Firefox's current speed gains.
I check this every time a new version is released as I'd love to make the switch from Chrome, but Firefox with one blank tab and no extensions uses more CPU than Chrome with 15 actual tabs. When I actually throw activity at Firefox it hogs the CPU even more.
Here's hoping v61 addresses it.