Firefox 59 released | Hacker News Reader

Firefox 59 released

mozilla.org

394 pointsdikiaap231 comments

Firefox 59 is the first Firefox release to support SameSite cookie attribute, joining Chrome.

> Same-site cookies ("First-Party-Only" or "First-Party") allow servers to mitigate the risk of CSRF and information leakage attacks by asserting that a particular cookie should only be sent with requests initiated from the same registrable domain.

I like this one!

How does this play out WRT "Block 3rd party cookies" (which wasn't on by default on Fx and Cr but was on Safari since an eternity?)

It sounds like they serve totally different purposes, where one is a server-side tool to improve security, and the other is a client-side tool to improve privacy.

Note that the definition of "Block 3rd party cookies" in Safari is different from the one in Firefox (not sure about Chrome). Firefox blocks a lot more stuff when that option is enabled in Firefox than Safari does when its option is enabled, which causes more web compat problems.

Blocking 3rd party cookies causes very few problems. I've blocked them for many years and have seen fewer than 10 sites it caused problems with. For those you can whitelist the domain.

My experience has been the opposite. It breaks every Amazon Pay integration and a lot of Paypal integrations as well. It also breaks online banking for the two credit unions I use, both of which seem to use some 3rd party service to run their backends. And then just whitelisting the domain isn't a great solution because then that means they can drop any 3rd party cookies, not just the desirable ones. So suddenly my online banking works, but now I've got google trackers and whatever other crap they want to drop. I believe I was able to whitelist just those cookies but that was quite a few versions ago and the interface is different now so I can't find where I did it. The cookie interface is really terribe; when you go to the site information it tells you whether the site is storing any cookies, and many times it says "No" but if you click the "view cookies" link it shows a ton of them. This happens when the domain doesn't match exactly (for example www.example.com vs example.com) so you can't really trust the Yes|No info box.

> whitelisting the domain isn't a great solution because then that means they can drop any 3rd party cookies, not just the desirable ones

Try something like uMatrix or similar add-ons. They allow you to configure rules similar to application firewalls:

  cookie * * DENY
  cookie * 1stparty ALLOW
  cookie creditunion.org finserv.com ALLOW

Great interface too, at least what I've seen on uMatrix: Most rules are configured with one click in a 'matrix' of hosts and applications.

It really depends. I, too, have been blocking them for years, with few problems. But other people run into problems more often. It really depends on the sites one uses.

The article about "Off-Main-Thread Painting (OMTP)"is an interesting read:

https://mozillagfx.wordpress.com/2017/12/05/off-main-thread-...

Great to see so much effort being put into the fundamentals.

It's great that they're continuing to add back customisation to the new tab page but why am I still restricted to two rows of "top sites"? I'm on a desktop here, I'm not going to run out of pixels. Even on Android I liked having more space for quick links to websites I visit semi-regularly. In fact it was more useful there because there isn't a good way to bring order to the history panel.

Is it just to make room for "highlights"? I switched that off because it's utterly useless for me. I don't find it helpful to be shown a random selection of things I've clicked on. Besides, I am capable of scrolling.

Yes I know, there's an addon for that (probably). But it just feels like such an arbitrary restriction.

Why are only twelve "top sites" allowed?

To reply to myself I've found that about:config has `browser.newtabpage.rows` which I can increase. When I do that and restart the browser it resets to 3 but `browser.newtabpage.activity-stream.topSitesCount` is increased and more sites are shown. Changing `browser.newtabpage.columns` doesn't appear to have an effect.

My question still stands but at least there's a partial workaround, even if it's potentially unstable.

EDIT This setting doesn't appear to be available on Android.

I've changed the `browser.newtabpage.activity-stream.topSitesCount` to 36 which fills my screen with 6x6 boxes filled with my top sites. Except, for the last box which is always empty!

It seems that there is a hard limit in Firefox of 35 top sites.

Did you try restarting your browser? I've just tried setting rows to 12, which sets topSitesCount to 72. This worked for me after restarting the browser.

And why can't I have my home page as my new tab page?

I don't want any of that "top sites" BS on a new tab, I want the local HTML file that I created as my home page. Why not allow it for new tabs?

I'm cursing for about a year about this. I spent days searching for workarounds and stopped short of hacking into firefox. Maybe I should just install firefox 50 and disable updates.

Only thing I've found is that hitting ALT-Home keys in the new tab will load the home page. But it's one of those tiny repeated annoyances that I'd love to avoid.

I can't really comment fairly here as I've gotten used to Chrome for the last 2+ years, but is there really not an extension in Firefox for this, even if it's not natively supported? Seems surprising to me.

Yes there are, but for some reason using a local file is a problem. Something about the ability of WebExtension to access a local file. Probably sensible, but in this case annoying.

Given the existence of things like momentum (https://addons.mozilla.org/en-US/firefox/addon/momentumdash/), I'd suggest maybe looking into packaging your HTML page into a web extension :).

New Tab Override should do the trick for you. I agree that it should be a built in though.

https://addons.mozilla.org/en-US/firefox/addon/new-tab-overr...

"Added settings in about:preferences to stop websites from asking to send notifications"

Been waiting for this one.

Yes! That is really being abused. I don't need notifications from recipe sites!

I’ve yet to find a site from which I do want to receive notifications.

chat apps like discord and slack, mostly

Even discord and slack I don't want browser notifications. Inside their own tab, I can deal with ui elements. There's no need nor no want to have a page alert outside of the web page itself.

It can be useful when you get mentioned, like for example, a user has pinged the moderator role where a quick reaction will be necessary. Desktop notifications can significantly reduce reaction time to important events (though, assuming you do the sane and set Discord to only notify you on mentions that aren't @here or @everyone)

If they need an immediate answer, they are not children and know how to use a phone (I would hope).

I'd imagine the majority of discord users don't share their phone numbers with the channel.

Well, if you don't know me well enough to have my phone number, you certainly shouldn't have the privilege of making my computer beep at me, whenever the hell you feel like it. But, that's me. Maybe you like people interrupting you. (Or maybe you like interrupting people because you think you are too important to wait en queue.)

The discord server I moderate has a strict "don't ping moderators unless it's important" rule and in general anyone who just pings for shits and giggles gets quickly muted.

I like it when people interrupt what I do when the server I moderate and therefore are responsible for is atm being swarmed by people spamming senseless shit into the chat or someone going of the deep end and yelling racist slurs, thank you very much.

I'm not suggesting your position is wrong, just that like IRC there's going to be lots of discord users who have a use for some kind of notification but would never want to give their number out.

Fair enough.

Study websites reminding me that it is time for my next study session are great for establishing the habit of doing my studies.

When I was single I would have had time for 3 of them - total in my life. I think systems should automatically block all requests for notifications if you already have 3 since it is not possible for more than that to be useful. (the exact number is debatable, but 3 is close). Note that I said systems - I want firefox, chrome, ie, safari to work together across all the computers I have to keep the number down - for security reasons I understand they cannot but that would be ideal.

I like calendar notifications, reminding me it's time to go do something

Web sites like gmail.com.

Indeed, once upon a time folks used RSS.

But we are in an age where any random website wants to send you realtime push notifications.

Instead of pestering you with an overlay that takes up screen real estate, the browser should indicate what subscription options are available for a page, e.g. in the address bar next to the page info icons.

Instead of obtruding modal, there should be just a little icon in address bar notifying that notifications are available, like some browsers used to notify about available RSS feed.

By now, you can have almost it by clicking on the (i) icon before the URL in the address bar. It lists permissions, if any, including notifications.

Sites have been replacing the actual API request with a HTML5 overlay, precisely because of this - only when clicking "notify me" does the actual permission get requested, with the requisite dialog.

(The rationale is "once the permission is denied, there's no way to bring it up again; a dismissed HTML5 box can be brought up again an infinite number of times")

As for me, I went for WebAPI Manager, which blocks most of the annoyances (Vibrate API? Never saw anyone outside malvertising use that), while still allowing me to manually whitelist sites.

Good thing Firefox is working on that too: https://news.ycombinator.com/item?id=16574228

Right. This is also the current standard approach for asking for iOS push permissions. Without intermediary dialog, the app can't try again.

And as app builders you have to explain people how to enable the permission anyways if people try to e.g. use their current location. Luckily since a few versions now, iOS allows apps to directly open a certain settings page, e.g. location services.

But if user decline native dialog once, it will not be asked again (at least for location services). So intermediary dialog won't help on iOS.

I usually disable JS and white list the sites that actually deserve it.

I actually came up with a decent use for the vibrate function on this control [0]. It's a clickwheel type interface for picking an altitude. When the number changes it triggers the vibrate function making it easier to use the interface without looking.

[0] https://i.imgur.com/aiFFzLP.png

edit:

I've linked a demo below with the following caveat, this demo was set up exclusively as a mobile interface and still needs plenty of work. One of the issues is the clickwheel size is initialized as a percentage of the screen height when the page loads. So if you resize the window significantly or switch to mobile mode in dev tools, you must refresh the page to get the clickwheel size to adjust accordingly.

https://faa-aviation-data-portal.github.io/pirep/pirepMobile...

That's how it was meant to be implemented, you trigger the request in response to user action.

The direct and immediate one is laziness, and I wonder what the success rate of that is.

Same dark pattern as phone apps that ask for permission with their own dialog then bring up real one or ones that ask for a rating but only send you to real app store rating system if you'll rate them highly.

Been trying to decide whether to do this myself, this really a dark pattern?

We're not asking until until they try to use the feature that requires location, but are using the additional pop-up. Rationale being that they probably want to allow it or they wouldnt be using app, and making it so they dont mistakenly deny and have to mess with setting is really more of a convenience.

But ya it does seem to go against the OS defined behavior.

It is still far too hidden for what should be a popular option.

Privacy > Permission > Notification Settings > Block new requests asking to allow notifications

It should be listed under > Pause notifications until Firefox restarts

I've been using the beta 59.0bxx series for quite a while and there are plenty of fun settings in about:preferences and about:config. One very important setting from preference, as I push those I know to adopt password managers, is the "Remember logins and passwords for websites" checkbox (don't save them in the browser). Another important security related setting for those with relatively new 2fa tokens is the "security.webauth.u2f" setting in about:config (still set to false by default with Firefox 59).

I agree, this was bothering me too. But they implemented it all or nothing. If I disable it in preferences it is completely disabled, no way to activate it for a single site.

They should have disabled the popup, but still allow me to activate it on a per site basis by clicking on the ⓘ in the url bar. Same for location, webcam and mic. I don't want sites to nag me with the location sharing notification, but on google maps I would like to enable it manually by pressing the ⓘ and granting permission.

about:preferences is not the place for this as the vast majority of users know nothing about what they cannot see in a menu structure.

about:preferences is the page you end up at when you select the "Preferences" option from the menu.

When apps send me notifications on my phone, it is a reminder for me to uninstall them.

When sites send me annoying ~~adds~~ notifications, it is a reminder for me to turn on adblock for them.

edit: adds to notifications (also I can't spell obviously)

Sadly, still unusable for me since the Quantum update.

A clean install and profile, no add-ons, and google.com/maps (for example) will max out my cpu. There must be some bug related to old hardware and OS (Macbook Pro 2009 + 10.11.x) or more people would be experiencing this.

I have been having problems with Google Maps too, and Google Hangouts has been even worse. I'm pretty much at the point now where I have asked people to contact me on different services because Hangouts has become completely unusable.

I was assuming that this was just Google not caring about Firefox support. But now that you mention it, this seems to have started around the time of the Quantum update and doesn't affect all my machines. So maybe I'm experiencing this bug as well? I never connected the two things.

Google had a NPAPI plugin for non Chrome browsers and proprietary webRTC extensions for Chrome, so when legacy plugin support was removed, it stopped working for voice/video calls in Firefox. Hangouts Meet is even worse, being Chrome only, though officially Google was looking into it.

I noticed that too. Firefox does not work efficiently on macOS, unfortunately.

Additionally, Google doesn't do much to fix their web apps performance on Firefox. I even suspect the contrary.

I suspect Google engineers have more important tickets to work on than "do a bunch of complicated VM analysis to slow down Maps on Firefox."

The test framework, or some live analytics, show your change makes Chrome 1 second faster and Firefox 2 seconds slower (in a 2 minute representative test). Do you push the change?

I have something similar on my MBP late 2012, running latest OS. I've switched to Safari on that machine to avoid running out of battery and maxing the fans.

Same for me on a 3.5 GHz Intel Core i7 13" Macbook pro. I had to revert back to Chrome.

e.g. Animations on stripe.com would auto max out my CPU.

Same for the mobile version. It's unusable on single core devices.

I have a 2017 macbook pro, battery life is atrocious with Firefox now. If anyone from the power optimization team is reading this what logs can I provide you?

Mike Conley has picked up the Quantum performance baton from Ehsan, and has started posting Performance fixes on his blog [1]. Please write to him; he is good with replying, and will provide instructions on how to profile your Mac machines.

1. https://mikeconley.ca/blog/2018/03/01/firefox-performance-up...

I think I must be hitting something like this too— Mid 2014 MBP running 10.11.6. I tried Quantum when it came out and used it exclusively for about a week, but there were just a myriad of small problems around performance. It would eat CPU and RAM, input would lag especially text entry, Hangouts didn't work, etc. And then on top of that a bunch of annoying interop issues like problems with the 1password extension ("the add-on could not be downloaded because of a connection failure"), not being able to right-click selected text and access the Services menu.

I really want to love Firefox, and I'm giving it another try with 59, but at the end of the day I need my computer to work for me. :(

hey, I have a similar issue on PC. It's not that the CPU maxes out but that everything grounds to a halt, especially when typing in text boxes on JS-heavy sites.

I'm on HN, right? Totally fine. Amazing browser. I hit up Twitter, and I could be stuck halfway through a tweet and it freezes up for a good minute or two. Same on FB, same on many other sites.

I LOVE the browser but on my laptop (Thinkpad), I had to switch back to Chrome.

The issues came about with the Quantum update for me as well.

Same. Totally bogs down my maxed out Macbook Pro on some HTML5 video.

Hi! I'm collecting profiles on cases like yours.

Here are instructions if you'd like to help: https://www.reddit.com/r/firefox/comments/7knnn4/firefox_qua...

I haven't actually checked if firefox was eating my CPU but with my pre-quantum profile on post-quantum Firefox (connected to firefox Sync) Firefox would essentially lock up whenever I was on a page with a lot of Canvas stuff. I made a fresh profile and the issue was gone. Don't know if this is at all related to what you're collecting.

Thanks, I will do so.

When Firefox optimised for multi-process, did they test for single-core? I'm on a machine released this decade - a rare instance of a single core x86_64 machine (fanless Atom NUC)

Facebook is pretty much now unusable with they do their asynchronous page loading of stories.

You are not alone. Firefox is unusable for me on my mac book pro 15' 2015. HTML 5 videos are basically a slideshow.

On my PC with windows 10, Firefox eat 25% to 30% of my CPU looking at a twitch stream (1080p60). Chrome and Edge will use about 5% while looking at the same stream.

For me Google maps loads fine, then a second or two later it will freeze for ~5 seconds, then go back to working normally. (i7 MBP 2017, OSX 10.13.3, Firefox 59)

Working perfectly here. Firefox 59.0.1 OS X 10.11.6 MacBook Pro (15-inch, Mid 2009) 2.8 GHz Intel Core 2 Duo 4 GB 1067 MHz DDR3 NVIDIA GeForce 9400M 256 MB

Although its a minor thing, I'm personally delighted with drag & drop for the new tab tiles. These are the kind of little usability improvements that can tip the finely balanced scales between major browsers these days.

I am delighted too. You look at it and you felt as if the Drag and Drop should have been there in the first place.

I could only wish they allow 3 rows of Top Site ( without editing about:preference ), and an option to have Open All Top Sites in New Tab. Which is very useful since I check those site daily.

Tab Tiles?

Top sites tiles on the new tab page in Firefox.

https://www.mozilla.org/media/img/firefox/home/new-tab.a1c9f...

After I switched to Firefox 59 beta on Windows (7 upgraded to 10), I see the following issue all the time. When accessing finance.google.com in one tab and etrade.com (internal page) in another, if I am on finance.google.com and if it refreshes and is slow for some reason I start seeing a blank page with a spinner on that tab and while this is happening, if I switch to the etrade.com tab, I see that page is blank too and has a spinner that doesn't refresh until the finance.google.com tab comes back -- almost like the browser is stuck in a system call somewhere and all tabs using javascript are blocked. Maybe it's likely some interaction with some addons I have (mostly NoScript), but it's been frustrating and for the first time in so many years I am considering switching browsers :-(.

BTW, where is the best way to ask about issues like this regarding Firefox, it seems like Googling for Firefox issues mostly only brings up years old issues and no one is publically posting about more recent issues like this. Also it's hard to describe problems like this in a reasonable way to come up with search terms that Google likes and doesn't hit old reported problems.

Not sure it's related but a lot of Google's sites are terrible on Firefox and Edge. IIRC they are using polyfills for some heavy stuff (if you assume no malice, that's probably the cause of many problems).

https://www.reddit.com/r/firefox https://support.mozilla.org/

Try using safe-mode https://support.mozilla.org/en-US/kb/troubleshoot-firefox-is... or maybe creating a new profile. The beta versions, just like Chromium or Chrome, occasionally have some weird bugs. At least for me most of them are always hardware acceleration related.

I've had trouble accessing the Google Analytics dashboard(s) in FF as well. Nothing gets rendered except the background gradient.

In the interest of not using Chrome though, when I run into such issues I use Brave as it's Chromium-based and thus allows me to side-step those problems.

The times I had problems with GA was because of some uBlock rule or weird A/B tests (I assume) or caching, clearing cookies and site data fixed the problem.

Google web apps and sites don't work well in Firefox, generally speaking.

Examples where Chrome performance and functionality is superior include:

- Inbox (try downloading attachments) - Maps - Analytics - ...

> Inbox (try downloading attachments)

The only times I got issues with this on Firefox was due to some add-on (Disconnect for instance). On Firefox out of the box I haven't had issues with downloading attachments lately, even with the add-on on.

Maps work fine for me (Dev Edition here though).

I've also noticed the same blank page with spinner (which seems to hang and never actually refresh in my case) on non Google sites since the quantum upgrade. Figured it was a bug that would get squashed in a subsequent version, but no luck yet.

I have this exact issue, and I really don't think it's google's fault. Both those sites are using canvas to draw charts. I experienced it on crypto trading sites mostly. Try a fresh profile, copy your bookmarks and history over to the new one. I did this and the problem was gone. (See my other comment in the thread)

> where is the best way to ask about issues like this regarding Firefox?

The #firefox channel on irc.mozilla.org is a good place to start:

https://wiki.mozilla.org/IRC#Project_Channels

And the Firefox subreddit is pretty active:

https://www.reddit.com/r/firefox/

Firefox uses a limited number of content processes (I think the default is 4? Might depend on your hardware). So if both tabs get assigned to the same process, one tab freezing up means you can't use the other tab at the same time. Check under Options -> Performance to see if yours is set to use only one, or not to use separate content processes at all.

Still on 56, sadly. The addons crucial for my workflow have stopped working. Shame these concerns were not properly addressed, even after months of clamour about breaking changes like these.

What add-ons?

For me it's LiveClick. http://projects.protej.com/liveclick/

'Display preview in live bookmark item tooltip' and 'keep menu open after middle clicking items' are worth me staying on FF 55 for the time being.

Sure I have tried upgrading and using other addons but nothing really does as well for me as this addon.

Same here. But I use the ESR version which is still mantained with security updates, right now its at 52.7

For me, it's: save MAFF format, foxyproxy, downthemall, "Image and flash blocker", tab mix plus, self exploding cookies, Torrent Status Tool.

It's sad.I guess I'm getting old and grumpy.

> Added settings in about:preferences to stop websites from asking to send notifications or access your device’s camera, microphone, and location, while still allowing trusted websites to use these features

Good. I get so tired of having to click that little [x] on the location prompt.

However it's really just covering up a bigger issue, which is that the permissions prompts in Firefox and Chrome should be redesigned. It would be good if they were designed so that they were still noticeable, but they:

* don't cover part of the webpage

* don't cover other parts of the browser's UI

* and don't take focus away from the webpage

That leaves no room for them to appear at all. I'm all for it.

Yeah, why isn’t it one of those thin drop down bars at the top that can be closed? Have it go away after a certain number of seconds.

More good news: expect to see Fx60 on Fdroid.

https://forum.f-droid.org/t/making-it-easier-for-f-droid-to-...

I switched to Firefox from Chrome at 58. I've been very happy with the experience all around. Keep up the good work, Mozilla!

I too switched for several months, but am already back to using Chrome. Even with the performance improvements, it's still noticeably slower than Chrome.

I'm curious: what OS do you use and what sites are you visiting?

I've had a stellar experience on Arch Linux. The RAM usage is down. I can't say I really ever found Chrome slow, but nor is Firefox for me. I think they're both plenty fast, but I prefer sites without lots of garbage flying around on the page anyway.

macOS High Sierra. I generally have a lot of tabs open: two Inbox tabs, Facebook, StackOverflow chat + question pages for what I'm working on, many JIRA tickets, HN, sometimes Twitch, etc.

I think it has something to do with video rendering since opening Twitch/Youtube often causes the problems to start.

Try https://addons.mozilla.org/ca/firefox/addon/auto-tab-discard... I had been using Tab Suspender from the same author but recently noticed that it was deprecated for FF 57 or newer.

Pretty happy with it on long-running tabs I check once or twice a day.

Still no proper Yukikey 2FA support. Github kinda works, Gmail still completely broken. I would love to switch from Chrome for the privacy benefits, but this is an absolute deal-breaker.

Edit: There is hope for Firefox 60 or 61 -> https://bugzilla.mozilla.org/show_bug.cgi?id=1409573

IIRC it's the problem of Gmail specifically, and how they implemented 2FA. Spec-compliant sites, like Giithub, work.

Yup, Fastmail works too. Actually, with FF58 and the U2F flag switched on, Google is the only site I use where my Yubikey doesn't work.

It's the old Microsoft strategy.

Break the standard behaviour slightly to make competitors product unusable in your products.

> Firefox Private Browsing Mode will remove path information from referrers to prevent cross-site tracking

Surprised this wasn't done yet.

Well, politics. As a browser vendor, you need to provide webpage owners ways to make (more) money, especially when other browsers that users are happily voting for, do provide these ways.

Sadly no news on the Web-extensions front. Several of my extensions continue broken.

What news are you looking for? I'm fairly sure extensions targeting the older APIs will not be supported in any future release.

We're waiting for the webextensions API to be extended with a bunch of missing functionality so that extension authors can reimplement things that were made impossible when the older APIs were deprecated. This is something that ideally should have been done before or shortly after the quantum release but was not seen as a priority.

Old extensions won't ever really work again, but part of me is still kind of holding out hope for WebExtensions API features for things like "allow extension on 'system' pages (new tab, settings, etc)", "focus addressbar", "hide/replace address/tab bars", etc.

Some of those are coming or pretty much ready (tab hiding, addressbar focus), and others I've almost given up on, given how Mozilla seems to have changed their philosophy.

That's unfortunate. I finally gave up on Firefox as my main browser when 57 came out, and one of the big reasons was that so many of the useful extensions that had set Firefox above other browsers for me were broken and didn't have good replacements. In some of those cases, my understanding is that you can't currently write a direct replacement because the new API doesn't support the required functionality. I was really hoping that by two releases later these kinds of issues would have been addressed and new Firefox would at least not be worse than old Firefox in this respect. :-(

> That's unfortunate. I finally gave up on Firefox as my main browser when 57 came out, and one of the big reasons was that so many of the useful extensions that had set Firefox above other browsers for me were broken and didn't have good replacements

I never really understood this argument, you gave up Firefox since you could no longer have your XUL/XPCOM extensions in favor of serious browsers who have less WebExt APIs than Firefox and don't have those same extensions?

Mozilla took away perfectly good functionality and gave existing fans a big middle finger while doing so. Although I am a fan of Mozilla, it is entirely understandable to me why someone wouldn't want to keep using their products.

And lets not ignore the fact that most sites are designed with Chrome as their intended target. FF simply doesn't work as well on many sites.

This is - in my opinion - one of the best reasons to use Firefox.

I don't want to go back to one browser dominating the web.

Not really. I gave up Firefox (as my main browser; I'm a web dev so use all of the major ones regularly anyway) because the main thing keeping me there was its extensibility, and that advantage went away. Firefox was and remains slower, less reliable, and less capable than Chrome, and IME has actually become worse in those respects as well since Quantum.

> Firefox was and remains slower

Did you try out the Firefox Nightly with WebRender enabled?

No, but somehow I get the feeling we're talking about entirely different levels of "slower" here.

You're talking about GPU-accelerating complicated page renders.

I'm talking about things like regularly seeing flashes of unstyled content on page load, which I thought we'd left behind somewhere in the last millennium.

> I'm talking about things like regularly seeing flashes of unstyled content on page load

Never seen on my end for quiet some time, maybe worth a bug report: https://bugzilla.mozilla.org

I had a look around for other people with this problem, and all the examples I found were caused by extentions. Ghostery just fixed theirs a few days ago, for example. https://bugzilla.mozilla.org/show_bug.cgi?id=1404468#c36

Based on the differences among different machines I use with Firefox installed, it is certainly possible that extensions are causing a lot of the degradation since Quantum, but I'm not sure assigning blame is interesting or useful here. The fact is that my experience as a Firefox user is now much worse than it was before. The new extension model was supposed to make things faster, more secure, more reliable, but sadly the result seems to have been very much the opposite.

The new extension model makes better extensions possible, but it isn't going to make good extensions appear out of thin air. It's not going to stop a buggy extension from causing FOUCs or slowing down the rendering of a page.

I don't want to make excuses if you're having a bad experience though. But I'm hopeful about the future of Firefox.

OK, but surely it should at least isolate buggy extensions so they can't compromise the browser itself? Wasn't better security and stability in the presence of bad extensions a key selling point of the new extension model? In practice, Firefox was failing to shut down and restart cleanly for me perhaps one time in three since Quantum, having been solid as a rock until that point. (I say "was" because I haven't yet seen this problem since updating to 59 the other day, so it's possible that that particular bug has now been fixed.)

Likewise, if it makes better extensions possible, how come the only observable difference so far seems to be that various things that extensions used to be able to do are no longer possible? It's looking like a classic case of going all-in on a big software rewrite, but finding that a lot of little details have been overlooked and useful functionality has been lost as a result.

The main feature of the new model is that it has less functionality than the old model. Firefox is betting that it can enable close to 100% of the functionality that users actually want via clean APIs instead of opening up literally every internal aspect of the browser to extensions. There will be bugs, but specifying APIs makes them easier to fix.

The advantages are speed, security, ease of installation, ease of development since more thought was put into the APIs, compatibility with Chrome extensions, and better compatibility across platforms since extensions now use less native code. Those benefits are "observable". The downside is that you have to wait for Firefox devs to think though the security, performance, and usability aspects of every bit of functionality an extension dev wants, and then wait for the new or updated extension to exist.

Those benefits are "observable".

Not yet, they aren't. In fact, the exact opposite has been observed here, many times already. This is my point.

As you say, I'm sure we all hope that the situation will improve in the future, and then maybe the sales pitch for Quantum will start to look a bit more realistic. Until then, sad as it makes me, Firefox will remain relegated to primarily a testing tool on my system, because today other browsers are simply better in every way that matters.

There are many reports of unusual CPU usage in some scenarios on Mac OS, you can help Mozilla debug the issue following this: https://www.reddit.com/r/firefox/comments/7knnn4/firefox_qua...

What's difficult to understand? I've been using Firefox as my main browser because of the extensions while choosing to miss out on features that other browsers had and Firefox didn't. Simply because the pros outweighed the cons.

I still use it as my main browser at home, but with the extensions that used to set Firefox apart gone and forced to reset my workflow and expectations to the minimum common functionality, I find myself using Safari more and more at work simply because the battery usage is better and has out of the box support for SPNEGO. As an added bonus, I can watch HD videos smoothly without killing my processor and draining my battery.

> Sadly no news on the Web-extensions front

Did you actually read the release notes? Quoting:

> Enhanced WebExtensions API including better support for decentralized protocols and the ability to dynamically register content scripts.

https://blog.mozilla.org/addons/2018/01/26/extensions-firefo...

Yeah, so, nothing really useful when what you’re waiting for is proper keyboard customization.

> Improved graphics rendering using Off-Main-Thread Painting (OMTP) for Mac users

Does this update improve its power consumption? Currently, Firefox has a significantly larger effect on my MacBook's battery life compared to Safari with similar usage.

Ya, I'm really disappointed with the battery consumption on FF, almost wanting me to move back to Chrome (shudder).

Switched to FF Nightly from Chrome about 6 months ago because Chrome's memory usage was awful. Firefox has been much better, especially since I modified the following in about:config:

    browser.sessionhistory.max_total_viewers => 0
    dom.ipc.processCount                     => 1

Isn't limiting the ipc processes severely capping (CPU) performance?

It's a tradeoff, IMO.

Either you spare a few CPU cycles or a bit of memory. Though chrome doesn't seem to be doing good on either front.

I haven't really noticed much of a difference in performance but memory usage is much lower.

It will cause all your tabs to run in one process, so only one CPU core can do work there.

This will especially be noticeable when multiple tabs are actively doing something, for example if you open up multiple tabs at once or background tracking script are stealing your CPU cycles.

Blocking tracking scripts and ads or just JavaScript in general can alleviate this to the point where you won't really notice.

https://bugzilla.mozilla.org/show_bug.cgi?id=1353319 - RESOLVED FIXED in Firefox 59, super happy about that.

I'll save you a click:

> Render the HTML preview within the Response side-panel

> Component: Developer Tools: Netmonitor

> Reported: 11 months ago

Firefox gets raving feedback here, but for me it is unusably slow on a Linux machine.

While websites do render fast, the UI is not very responsive, whether its laggy typing into the address bar [1] or the crawling devtools.

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=1408699

I switched back to Firefox for dev work on my system (Arch Linux), because Chrome would slow to a crawl with its devtools open. Firefox (since quantum) seems to fare better.

I also use Arch and Firefox Beta daily at home. Don't see any issue.

59 is probably a bit of a less spectacula release than Quantum itself but the recent updates nonetheless bring me joy.

The "stop notifying me about permissions and just block it" is something I genuinely didn't know I wanted. Though I do hope (but will have to test later) if this also works for HTML Canvas since I use the additional anti fingerprinting.

Won't be updating beyond FF 55.0.3 because 'Status-4-Evar' add-on doesn't work with Quantum. Apparently it can't be updated to work with Quantum for reasons I don't know, but was mentioned on the support forum.

Until such time as I can prevent the annoying floating URL pop-up in bottom corner when hovering over links, I won't be updating. I don't see why I can't have a harmless fixed status bar extending 100% at bottom of window. It creates a nice gap between Windows taskbar and the browser, with the URL hover text contained neatly within and doesn't appear over the page as an irritating floating layer.

Is it law at Mozilla that FF must imitate Chrome on interface style? Perhaps my choice of wanting choice is the wrong choice. Perhaps it's my fault as user, for wanting something that was taken away.

Another add-on I liked, now kaput, is Mozilla's own Lightbeam. The important list view and domain blocking function of that extension has been broken for some time. The march towards the "best browser" seems to involve breaking things and pissing people off.

Used to prefer the status bar but am sort of used to not having it now. But I think the new way is broken also. Don't see why I have to look down to see the link target. Why can't it be a tooltip at the mouse pointer, perhaps below any alt text?

I would not choose to have a tooltip link hover, but you should be allowed to have it if you want. Via options, about:config, or possible via an extension.

In UI terms, the concept of a fixed footer is well established. We see it often on sites and applications. FF allows adjustment of the millisecond delay of the link text, but no control over other properties of that element. I don't get the logic. When vendors move the furniture around in their "security updates" it makes me distrust the updates, particularly when the furniture can't be moved back to how it looked before.

Is allow client side decorations available yet in about:config? It was in a fedora release about 2 months ago, and it was removed because it wasn't in mainline. I have the title bar disabled for now, but that's broken in it's own way(dragging doesn't work right).

No, it wasn't ready: https://bugzilla.mozilla.org/show_bug.cgi?id=1440461

Should be part of Firefox 60.

That's a shame. Client side decoration seemed ready when it was enabled on Fedora.

A little OT, but how can I enable/use the Screenshots feature? I cannot find it.

Right click -> Take a Screenshot

Also, there are three dots next to the address bar, where it can be found.

If you open that menu with the three dots, then control-click on the Screenshots menu item, you can add a Screenshot button to the address bar.

https://news.ycombinator.com/item?id=16576601

Is there a way to trigger a full screen screenshot via Selenium/WebDriver? (not a partial/window screenshot - that I can do)

If I'm understanding correctly, could you just select the body tag and screenshot that? In Selenium Python that'd be:

    # driver is a Firefox driver
    body = driver.find_element_by_tag_name('body')
    screenshot = body.screenshot_as_base64 # png bytes

Just tried this - doesn't seem to work:

- a lot of (visible) elements are missing

- the screenshot is still cropped to whatever viewport is used

Definitely just speculating here, but there is a `driver.save_screenshot(path)` method that would probably be worth trying.

This worked for me:

    from selenium.webdriver import Firefox
    from selenium.webdriver.firefox.options import Options

    geckodriver_path = '/usr/local/bin/geckodriver'
    options = Options()
    options.add_argument('-headless')
    driver = Firefox(executable_path=geckodriver_path, firefox_options=options)

    url='https://news.ycombinator.com/item?id=16576015'
    driver.get(url)
    driver.save_screenshot('./hn.png')

Firefox has a screenshot feature that I'd like to trigger from Selenium, which is by going through the "..." icon in the URL bar which opens up a menu with its last item being "Take a screenshot" (and then you can select a full page screenshot).

Is there a way to programmatically trigger this feature or sequence?

It's the lowest item in the context menu when you rightclick anywhere on a website.

I think that's been around for a while, right? It's under my right-click menu.

I cannot see it there either.

https://postimg.org/image/dexsmbvp9/

On the right-hand side of the location bar, there's a Page actions menu (the 3 dots) which includes taking a screenshot.

I cannot see it there.

https://postimg.org/image/59fqo4pfx/

Are you sure you've updated your browser? I have the option in that menu.

Sure, I have. This feature is there for a long time as of my understanding (since Quantum, maybe?).

I have tried this feature when it was only in test pilot mode, but then I removed it (that was a long time ago). Can this be related?

https://postimg.org/image/gw4g2vk3h/

Did screenshots somehow get disabled? In about:config, check the value of the extensions.screenshots.disabled pref.

No it was never disabled. Also, I have tried to disable it and enable it again and restart the browser and similar things...

Weird. Maybe try filing a bug to see if someone who works on that feature has any other ideas.

https://bugzilla.mozilla.org/enter_bug.cgi?product=Firefox&c...

Hey, Screenshots person here...this is a new one to me.

My only thoughts would be the pref flip that seems to be ruled out at this point, or that you're trying to shoot an about:[SOMETHING] page for which we disable screenshots.

Otherwise, file an issue: https://github.com/mozilla-services/screenshots/issues

We're triaging in an hour or so :)

Thanks and done: https://github.com/mozilla-services/screenshots/issues/4220

> Added support for W3C specs for pointer events and improved platform integration with added device support for mouse, pen, and touch screen pointer input

ohh yeah finally! I somehow missed that even though I'm following Nightly.

Can't find in the release notes: Is the Android version now based on quantum as the desktop version? As it was promised a few months back:

https://www.androidpolice.com/2017/09/26/project-quantum-enh...

One of the major Quantum improvements, Stylo, isn't coming to Android until Firefox 60: https://bugzilla.mozilla.org/show_bug.cgi?id=1366049

Unfortunately, it'll still be pretty slow until then.

Can someone tell me what's behind "Faster load times for content on the Firefox Home page"?

I'm sad to say that Firefox Quantum will remain useless to me until proper session management is restored. But without even the necessary API to enable features that were provided by the "Session Manager" add-on for many years, this still seems to be a long way off.

Looks like 'client side decorations' on Linux didn't make this release?

With the tree-style tabs extension, tabs on the top are obsolete :)

CSDs were already in 58 (Customize, Title Bar). What didn't make this release is Wayland support.

I think that was only backported into the Fedora version?

Looks like you are right, I cannot find it in Ubuntu version.

Speaking of linux features. Is it me or did it stop using native desktop notifications?

Nope: https://bugzilla.mozilla.org/show_bug.cgi?id=1440461

Does anyone know why Alt+D doesn't focus into the Address Bar anymore? (I thought the behavior was suppose to be like Ctrl+L) I'm on Linux.

I use Firefox Focus on my phone (android). Anyone know if it tracks the full Firefox releases or is it something completely separate?

It's completely separate.

As in, it doesn't even use Gecko, it uses the native web engine of iOS/Android, so that's WebKit / Android Webview. With how many trackers and ads they block in Focus by default, they wouldn't exactly motivate webpage owners to support Gecko anyways, if they'd use Gecko in it.

This is the current release notes of Firefox Focus on Android: https://support.mozilla.org/en-US/kb/whats-new-focus-android...

Damn I'm excited about this (specifically those damn notification popups!).

Slightly off topic: has anyone else been getting frequent crashes lately? It's as if it's running out of memory then freezing a moment. I thought it was because I'm on an older phone at first (note 3 here) but my wife said the same thing to me and she is in an edge 7.

(Kinda hoping someone knows a config fix to it haha)

Off-Main-Thread Painting (OMTP) for Mac users is some good stuff. FF is now noticeably smoother. Nice!

Has the issue with HSTS supercookies leaking into Private Browsing mode been fixed yet?

Relevant Mozilla and Chromium bugs:

https://bugzilla.mozilla.org/show_bug.cgi?id=930638

https://bugzilla.mozilla.org/show_bug.cgi?id=1232961

https://code.google.com/p/chromium/issues/detail?id=104935

Does this include the forced collection of the user's usage data?

Better stay away from the evil Mozilla corporation, and instead use one of those other, more privacy-aware, browsers like Google Chrome or Microsoft Edge.

That did not answer the question, and in light of discussions just a few weeks ago on this subject, the question was fair.

It's okay to do bad things so long as your main competition is worse?

wrt to satisfying every single extant ideological constraint that determines 'good' or 'bad', yes. a pragmatic assessment is that, due to it not being possible to satisfy contradicting constraints, it is thusly 'okay' to not do the impossible.

Telemetry is opt-in, and can be turned off at about:preferences#privacy

cmd+f,"u2f" or "webauth"

Nothing.. I wonder when they'll ever get to fully supporting u2f (probably via webauth) so I don't have to use Chrome to log into certain websites.

Mozilla will ship the Webauthn API in Firefox 60 (May 9):

https://hacks.mozilla.org/2018/01/using-hardware-token-based...

Congratulations!

What's the memory usage like compare to earlier versions?

Looking at that bug, one problem with implementing the nonstandard "mousewheel" event is that the browsers that implement it have different behavior in terms of what the delta value means, so sites using it tend to do browser sniffing to decide how to handle it. Which means implementing it is likely to break sites unless you exactly match the UA string and behavior of some other browser.

Given that, and given that Firefox does support the per-standard "wheel" event, what is the argument for implementing "mousewheel"?

(I think this comment was meant to be posted somewhere else…)

No, I posted it in the right place. Tepix subsequently edited his/her comment, as far as I can tell. The comment I was replying to, about "mousewheel" events, is certainly no longer in evidence.

Sorry about that, i changed my comment because it was obsolete and i could no longer delete it.

Next time it might make sense to make it clear that the comment was edited and how... ;)

Did they fix the bug introduced with multi-process where select lists in their popped-up state don't respect css styles?

Mozilla should also be focusing on RAM utilization. If they are doing that on the back end they should explicitly mention it in the update details.

They are doing that, with the whole Firefox Quantum thing. I see no need to mention less RAM utilization now given that they reference Quantum. These aren't meant as an advertisement, they are giving the changes since the last update, so there's nothing to say about it. These release notes aren't meant as advertisement.

There were thousands of changes from Firefox 58 to Firefox 59 (I count > 8900 changesets, though some are merges).

The release notes only list the most visible and most unexpected changes. Ongoing memory and performance work rarely clears that bar. It's happening all the time.

Firefox is dead to me ever since they killed their addons and marketed it as a step forward. If I wanted to use Chrome I'd use fucking Chrome.

I don't know why you are being downvoted. I have the same problem. I'm still stuck with Firefox 56 and no upgrade path in sight. I wish Mozilla declared FF 56 a LTS release for those of us who value their addons more than performance benefits of Quantum. Right now I'm still using an out-of-date release that doesn't receive critical security updates, and I have to choose between either downgrading 4 releases down to FF 52 (bad) or upgrading beyond FF 57 and losing addons that I came to rely on (unacceptable).

Which are the addons you're missing?

Do they have a GitHub site?

I miss Tab Groups, FWIW.

Why does it matter if they're on GitHub? Firefox no longer has the APIs for many of these extensions to be possible.

Why does it matter if they're on GitHub?

I'm not OP, but usually the conversation goes like this:

Person 1: <describes problem with open source software>

Person 2: It's open source. Have you submitted a patch yet?

I don't have an example handy, but when FF 57 was first released, I actually took the effort to check what are the prospects of extensions I use in future FF versions. I found at least one extension which required a feature that FF developers refused to fix because it would expose APIs they didn't want to expose, end of story. So it's not just a matter of "scratching your own itch", a lot of it is politics.

So it's not just a matter of "scratching your own itch", a lot of it is politics.

Absolutely. I don't personally believe that open source projects are above criticism - just explaining why the question was likely asked.

Everybody has their thing, but Tile Tabs was a killer app for me, and had no real equivalent in any other browser platform.

Bug 1318532 had a patch submitted which provided a WebExtensions-compatible API which would allow the original implementation to work, but this was spiked permanently because of :reasons: The developer, DW-dev has been gamely trying to provide similar functionality via implementing a browser-managed tiling window manager, but this seems doomed to failure since it is at the mercy of the display rendering on the target OS. For example, on MacOS, a massive shadow is rendered around an active window, which thus renders over the top of a neighboring tiled window.

Doubt this one will get fixed ... unless Chrome implements it. Right now the objective seems to be full Chrome "compatibility" and Chrome doesn't have it.

After I gave up on Classic Theme Restorer, 4 most useful addons remain: EdgeWise, Session Manager, DownThemAll, Status-4-Evar.

The thing that kept me from upgrading was TreeStyleTabs, but there is a non-XUL version of it which works fairly well.

While it is a pain, it's not as if Mozilla removed XUL support with no warning. What addons are you missing? Perhaps we can find updates or replacements?

Session manager is the main one that's missing for me. Nothing else can reliably selectively restore a session with full history and unloaded tabs on startup after a crash for me.

Interesting. I've never had an issue with the built-in session restore functionality. I'm not dismissing you, but asking what issues you've had with it (so that I can watch out for them.)

the add on can make multiple dated session stores and you can pick which to restore from.

Sometimes, usually after a crash, firefox will restore an empty session. Session manager has a working backup in those cases.

Likewise, I will not go further with firefox until there is a working session manager. The one supposed alternative that is compatible has poor reviews and appears unreliable.

I might add that I was a nightly user for over 30 prior versions.

For me it's EdgeWise, Session Manager, DownThemAll, Status-4-Evar.

Classic Theme Restorer, Tab Mix Plus, Session Manager, Status-4-Evar.

I've looked for replacements but haven't found anything that comes close. From what I can gather, much of the functionality that allowed these extensions is now removed from Firefox. I thought equivalent APIs were going to be made available so they could be ported but that didn't happen.

Maintaining a LTS release while focusing on an entirely new technology with a limited number of devs available is an engineering nightmare. They have to choose their battles. The performance of the old Firefox was really bad compared to competitors and this technical debt wasn't sustainable on the long term. I'm happy that they've made this bold move (on the big risk of frustrating some users (rightfully so)), and happy that they rebuild a better browser for all of us (and I'm sure they will come back in the future with the same flexibility for addons).

>and I'm sure they will come back in the future with the same flexibility for addons

No they won't, they have been very clear on this matter. Web Extensions are crippled compared to NPAPI. You can't even disable javascript with them.

They are already maintaining FF 52 LTS (actually, in Mozilla terminology it's called ESR) They should've calculated this in advance and made FF 56 LTS. They could still do it now if they wanted...

He's being downvoted because of his choice of wording. It's perfectly legitimate to prefer Firefox prior to Firefox 57. It's not legitimate to:

- act like Mozilla just killed their extensions lightly. They gave lots of warning in advance and took many extra steps to ease the transition. They also had lots of really good reasons for doing it.

- blame Mozilla for marketing things based on what their market looks like. Most users were not affected at all by the extension ecosystem change, but were positively affected by the newfound performance, security and stability. They also would not have done it in the first place, if they did not think it was a necessary and good step. I really don't know what you'd expect them to market it as.

- in general, speak like you're the only user whose interests are worth considering.

Well, he's upset. After all, when you support a project for years, you convince your friends to switch, participate in the community, maybe write some extensions, and then it removes one of the main reasons you have ever been using it, you get upset. It's understandable... I used to have strong feelings about it as well. You say that Mozilla didn't kill their extensions lightly, etc. But still they didn't think to make 56 a LTS release to allow us a breathing space of at least a year of security updates until we decide what to do next. Going 4 releases back for someone who's used to be on the bleeding edge is quite daunting.

The killer add-on for Firefox is Containers. This lets me run multiple instances of Gmail without them knowing about one another. I run Facebook and Twitter in another container and they don't know about anything Google.

My default profile is not logged into Google or Facebook or anything similar and uses DuckDuckGo for search.

You can set URLs to always open in a container. That gives me the benefit of cookies without having every website being able to detect them.

Firefox is fast, at least as fast as Chrome, for me. I did lose some extensions but the trend is very positive.

Reply on news.ycombinator.com