Hacker News new | past | comments | ask | show | jobs | submit login
Firefox 70 released for Firefox Quantum: Developer Edition (mozilla.org)
362 points by romelsongalia 51 days ago | hide | past | web | favorite | 112 comments

Worth noting, this landed a patch that moves some of the UI work to Core Animation on MacOS[1]. More work will happen that builds on top of this, but some reports in that ticket are promising already.

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=1429522

Moving from OpenGL to Core Animation has required a huge amount of work, but even at this early stage (before many of the changes in progress land) the power savings and performance gains are already significant.

Good work, and much appreciated!


Has there been a significant update to the devtools recently? I've been a devedition user for years, odd to see it on the front page of HN just linking to the homepage..

This is why I dislike these software posts without context. It's never clear if there's been an update or if the submitter just discovered them for the first time.

I really wish posts like this, and posts that are just random Wikipedia articles, would have a paragraph (or at least a sentence) from the submitter.

Why is this interesting? For Wikipedia articles about people this is doubly important because the title is inevitably just the person's name - unless you already know who that person is there's no hint as to why they are significant.

Disagree. I like the way this system stops submitters giving their potentially half-baked opinions on articles, and adds an element of mystery to things.

I'm inclined to agree, tossing in the poster's spin tends to drive the topic, often to odd places rather than the actual thing.

At the same time links to pages without any context as to why leaves me wondering why I should be looking at this page. Particularly product pages that are not easy to parse what the point is sometimes.

The selection of the article itself is sufficient to add spin.

I don't think I understand.

Say Google makes a significant change to their ToS. If you link to the ToS, the title will be "Google ToS". So instead you are left with the option of linking to a Guardian/buzzfeed/wired blog post that is mostly fluff, doesn't include the actual ToS text in context, and has a hyperbolic headline. Which article you chose to post determines the spin.

I don't think just listing a product page or a ToS is "spin". Product pages my be there to talk up their product but we know that.

While I'm in favour of restricting editorializing, simply posting a link called "Firefox Developer Edition" (the title tag on the linked page) seems to only cause confusion[0][1].

The title of this has since been changed, which causes even more confusion since the comments I linked now make no sense to someone who never saw the original title. I'm not sure if the new title would violate the title guideline or not, but it's certainly a better choice than what was there before.

[0] https://news.ycombinator.com/item?id=20819179

[1] https://news.ycombinator.com/item?id=20818856

The title rule encourages blogspam and bad articles. For example if Google were to make a big change their ToS and you link to it, the title would be "Google ToS". This is confusing. So in instead you link to some crappy article with a better title about the same subject. Or not at all.

That’s why I come to the comments first for these posts. I believe the forum is working as designed.

It's the FF 70 Beta Dev Edition. It was released yesterday.

A link to the release notes would be preferable to the homepage then?

Can anyone explain exactly what's different about this?

Are all these features not in normal Firefox? Is this like a beta of normal firefox, set of extensions, or a significant fork?

I think it's mainly a difference in configuration. The developer version, IIRC, allows you to install extensions on your own machine without seeking approval from Mozilla, but on the other hand it apparently subjects you to additional telemetry:

> Firefox Developer Edition automatically sends feedback to Mozilla ... In addition to the data collection described in this Privacy Notice, these versions by default may send certain types of web activity and crash data to Mozilla and in some cases to our partners.

If you just want to be able to install unsigned extensions with neither the extra bugs from running a beta nor the telemetry, you can use the far less advertised "unbranded builds" of Firefox [1] instead.

[1]: https://wiki.mozilla.org/Add-ons/Extension_Signing#Latest_Bu...

Don't the unbranded builds also lack automatic updating?

It's possible. Last I heard was that the unbranded builds were too eager to update [1] so it could well have been disabled totally. The Mozilla wiki doesn't mention anything about updating being disabled which to me implies it is not. If you find out for certain, please let me know!

I don't personally use the builds as Mozilla has not yet blocked signing of our extensions and `web-ext run` serves me well for development. Automated signing is not brilliantly documented but works well enough [2-3] to ensure that our builds are tested by real users before we release a stable version.

[1]: https://www.ghacks.net/?p=123637

[2]: https://github.com/tridactyl/tridactyl/blob/f45c990eac074df8...

[3]: https://github.com/tridactyl/buildbot/blob/master/autobuild....

When was the last time devs took time out of their day to create detailed bug reports with crash logs and context? Telemetry done right is just a way to surface bugs properly.

I use nightly as my main browser because I want Mozilla to see my bugs. And I want Mozilla to not do telemetry in production. No such luck.

> allows you to install extensions on your own machine without seeking approval from Mozilla

Unless it changed quite recently, you can do this with normal Firefox builds too AFAIK

It changed, but not really that recently.


Note specifically the step where the addon must be signed before it can be installed in this fashion. "Developer edition" does not require you have Mozilla's permission to install addons.

Special features: kind of like an IDE on top of Firefox, meant to be used by web developers. Comes with its own config profile, so you can fuck around and not break stuff in stable/beta that you are supposed to have concurrently installed.

Kind of like beta: all of the new APIs and features related to web development are there three months before they hit stable.

The retail edition of Firefox already includes dev tools.

...and Firefox Developer comes with dev tool features that have not yet reached Firefox Stable.

Which ones are not in the stable version yet?

It is based on the beta release channel and has some configuration tweaks. See https://developer.mozilla.org/en-US/docs/Mozilla/Firefox/Dev...

They are not, apparently the reasoning was not to make Firefox bigger with needless features not requires by non-techie users.

At least this is how I recall it, please anyone correct me if I am mistaken.

It's a beta with special features for developers. right now dev edition for me is 70.0b1 regular is firefox 68.0.2

It also has the new logo and I suppose there's new stuff on the dev tools. "What's new" page is broken so not absolutely sure.

Seems very much a beta for developers. Some cool dev features promised, but didn't always work as demoed. Font parameter manipulation for example. UI's there and pretty...it just doesn't affect the look/styling of fonts rendered on a real site like it does their demo site. Also, a number of UI controls slow to respond (e.g. cursors for dragging, clicking, etc.) β β β so very β

Definitely a few developer friendly plugins like one that parses and renders JSON responses in a new tab instead of showing unformatted plain-text.

I can't seem to make the switch away from Chromium devtools. I don't know whether it's because I am habituated to the Chromium ones or whether there is something klunky about the Firefox devtools that I can't put my finger on.

As others have pointed out, it may be worth looking into whether or not your reasons for using Chromium for the devtools are simply habit.

Speaking for myself, I'm in the other camp. I never really ditched Firefox for Chrome except for a brief period[0]. I returned to Firefox back when it was still called "Firefox (Not Responding)". Other than the renaming (/s), it -- generally -- flies. The devtools do what I need them to do, however, I used to hop back and forth from Chromium to FF a lot more.

So I'll be re-evaluating things, again, myself -- just in the other direction. :)

[0] I used Chrome for about two years after its release because Firefox wouldn't work with an internal application. It was just easier to use a single browser.

I tried for the last several weeks to switch to Firefox ("normal" edition) after Google introduced new mechanics pushing account login, but ultimately, and disappointingly, found myself drifting back to Chrome.

Main reasons were speed (Firefox opened slightly slower, and there seemed to feel like subtly more per-character latency typing in URL's) and minor anti-native behavior (the default-disabled-until-you-move-your-mouse Save button was particularly annoying, even though I understand the rationale).

I got the dreaded "Firefox process is already running" dialog a few times and had to close all my windows (ugh). I really missed Session Buddy; tried alternatives like Tree Style Tab but they weren't as clean.

Finally, one or two pages didn't render / operate properly (guessing due to non-cross-compatible JavaScript statement buried somewhere in them).

Unfortunately I don't have the time to create good step-by-step reproducibles and log bugs. I know some of this can be tweaked out via preference settings. I even spent some time tailoring my userChrome.css per suggestions here. But just couldn't get everything to work quite perfect. FF will have to spend some more time as my secondary browser, may give it a shot again in another year or so or next time Google pedals something stupid. Will try out the new devtools when I'm working on a page, maybe that'll be the hook for another attempt.

>it may be worth looking into whether or not your reasons for using Chromium for the devtools are simply habit

After deciding that life's too short and that I really should give it a try - I've found out that source maps doesn't work on Firefox's dev tools. Back to chrome.

Each has their pros and cons. My favorite features in FF devtools which are missing from Chrome are CSS diff and the advanced filtering in the network panel. Chrome filtering only works on hosts, while FF can filter in query strings and other sections of a request.

I often use network filtering in Chrome devtools to find requests by query string. If you start typing into the search box it matches it against the whole request including url parameters.

You're right, that works. Either that landed recently, or I'm confusing with another filtering feature that Chrome is missing.

Having recently switched to Edge Dev after having used Firefox for roughly a year (and Chrome before that), I'd say Firefox devtools' strongest point are the style editing stuff.

Somehow Firefox devtools feel a bit clunkier when working on JS projects but they offer a lot of neat gimmicks if you dig into the depths of CSS. I also find the guidelines when highlighting an element more helpful than in Chromium for aligning elements on the page.

Personally I don't get to use grid layouts as my code has to work in IE11 (I gave autoprefixer a try but in the end it was no significant improvement over flexbox because of all the caveats and limitations). If you do, I'd imagine Firefox devtools would be a gamechanger for you.

I find myself occasionally switching back to Firefox for CSS work, then going back to Edge Dev for everything else.

What percentage of your users are on IE11? Did you measure if there's at least one user on IE11?

Nearly 100%. Welcome to German enterprise intranets.

The only good thing is that when the new Edge enters general availability, we can likely get rid of 100% of our IE11 users for good.

Having to fight with IE11.. You are one of the heroes of modern times ^^

I guess it's a question of habit.

I have the same problem in the opposite directions: the interface of Chromium developer console is unfriendly to me, I never manage to get comfortable with it...

>I have the same problem in the opposite directions

Me too, as one example: Chrome had no way to edit/re-send AJAX requests for the longest time

Other than the web socket support in Chrome, the FF dev tools are a billion times better than Chrome now. I cannot work without the css layout overlays from FF dev tools. Makes working with grid layouts insanely easy.

The Chrome profiler is a lot better (IMO) than the one Firefox ships. But then there is https://profiler.firefox.com/

The debugger/sources view in Firefox also seems buggier still. E.g. a lot of times while it correctly displays source-mapped locations in the console, clicking on those source links will not actually go to the right place in the source view (it will either end up in the packed source or it will not go to the correct line in the source-mapped file even tho the console link gave the correct line).

One thing I miss that Firefox does not have at all is the Layers view. Which would be especially important since such things are browser specific and debugging performance/memory issues and making sure the engine uses a sane set of compositing layers can be important sometimes.

But the feature I miss most in my tiny corner of the world by a large margin is the Websocket support indeed. This is the thing that always makes me go back to Chrome devtools.

WebSocket inspector will be in DevEdition next week!

I think they're almost done with the first iteration of WebSocket support in the Firefox DevTools. Might even already be in Developer Edition: https://bugzilla.mozilla.org/show_bug.cgi?id=885508

Great news. This is the only thing I open chrome for! Thanks for the info.

DevTools guy here, WebSockets will be in DevEdition in the next build, b2, next week!

Thanks for your work!

While I mostly agree, I do switch to Chrome every once in a while when the javascript stacktrace in FFdev is a bit unclear. Chrome usually seems to be a bit better. But I always switch right back to FF.

Debugging fancy JS frameworks in firefox (for me, unexperienced js user) is more painful than in chrome, but debugging js in webbrowser is pain anyway

In Angular (which I like) I have found that there is a bug in error logging that only occurs on Firefox.

So every time I stumble upon that I'll open Chromium, find the stack trace and go back.

how many production websites you work on use Grids? just curious.

Only for internal apps and personal stuff. I don’t really work on client facing stuff.

I tried FF dev edition with Angular, it was getting randomly slow. Weird issues popping while debuging typescript code. In the end I have to use Chrome for my work, even though I would like to drop Chrome personaly. That said, I sometimes have to debug Angular applications on IE11 which is worst nightmare...

I had similar experiences with a Vuejs application where I rendered thousands of "reactive" elements. The Chromium JS engine had substantially better performance then the Firefox JS engine.

However, to be fair, this is a strong sign that as a developer, we are doing something wrong. If a JS application feels slow on an average browser in an average computer, it will feel slow on all slower computers. The usability will be crap. In this particular example, the frameworks are not suitable for such a large number of "reactive" elements.

If its having performance issues on Firefox.... Leaving Firefox and making it work on Chrome is part of the problem with the browser wars... I also would consider that a bug in either that framework or your code or both. I rather websites worked the same in all browsers otherwise why bother with JS frameworks that are to be compiled for all browsers.

I use Firefox developer edition as my main browser, and there does seem to be some issue with Angular sites. admin.google.com is particularly bad – but so bad it makes me wonder if they only test in Chrome.

Error logging used to be broken in Firefox at least until last year. I'm not working in frontend at moment and I never got around to submitting a bug (it always happened while at work) but my opinion was that there was a bug somewhere in there where they used a Chromium only API.

And yes, I too have wondered if they only test in Chrome.

In Chromium in the Network tab how do you see the request? In Firefox you can edit and resend a request and see it or modify it and send it again.

Does anyone know why Chromium hasn't fixed this supposedly old regression bug where opening large minified js file in the source tab hangs the whole tab/browser, sometimes for more than 10 seconds even on a fast machine. Can't this be done asynchronously or at least give the user a way to cancel loading it? If you open a large file and then close developer tools, reload the page and open developer tools again, it will try to reopen the last opened file again, blocking chrome for another 10 seconds. Very annoying.

meanwhile I'm the exact opposite. I've been using the firefox dev tools too much that using chrome on someone else's machine is something I feel like I have to put up with.

That's not me saying Firefox's tools are actually better though, there's still a number of weird quirks or features that could be lacking, while also having features I can't find in chrome that I also take for granted

I generally start with Firefox but recently needed to switch to Chrome for JS debugging. There were errors and they simply did not show in the FF console, and whole script tags of inline/embedded javascript was not to be found anywhere I spent 20-30 minutes mucking around looking here and there, and finally switched to Chrome and was done in a few minutes.

Firefox (hardened config) for browsing.

Chromium for web dev.

I miss ctrl-g as go to line. Is it possible to remap Firefox keyboard commands?

To my knowledge, no. They said they were working on it ages ago when they introduced the commands API for webextensions, but I don't think it is high on their priority queue.

I use Firefox devtools only when I need special features: Mostly: Grid editor CSS Shape editor

FYI, this post is because Firefox 70 was released for Firefox Quantum: Developer Edition.

Should the title be changed?

Why Firefox is not allowing us to edit JS code as allowed by Chrome.

What’s the Mozilla community like for external contributions? Before the big google dollars it seemed cliquey and now seems mostly paid people. Is that accurate ?

I wonder what happens to Firefox if the search dollars run out and companies are uniting behind chrome/WebKit

I regularly watch commits from webkit, chromium, gecko-dev and servo.

Webkit is developed almost exclusively by Apple (or more maintained because development is very slow, btw I suspect that Apple will drop webkit as it become too far behind) There are also a few igalia devs and one Sony dev + one Gmail dev.

Chromium is developed by so many people, Google, Microsoft, opera, QT, hardware makers (ARM, Intel, nvidia), some enterprises, and many contributors (I wonder if chromium has more opensource non paid contributor than gecko)

Servo is mostly developed by paid mozilla devs, but had major improvements made by contributors. Proportionally far more non paid contributor than the gecko of today, I expect mostly because they work on github which is more friendly than bugzilla and mailing lists...

Gecko dev today non paid developers are a few minority.

If Firefox does not die by himself, expect Google to slowly stop paying them Thus mozilla will die which would be a great loss.

Mozilla could save it's future, mozilla could become sustainable: I made a proposal here: https://github.com/servo/servo/issues/24026

What a terrible argument. I really don’t understand how people don’t see the negative effects of browser monoculture. I feel like I’d just be rehashing the same old points :(

Which argument? Firefox is dying, it's a fact, watch the marketshare over time.

Were you referring to my proposal on github?

> Were you referring to my proposal on github?

I was, yes.

Sorry but what's your alternative? I understand that it does not please you (I actually think that mozilla merging with chromium would be a good thing and we disagree on that)

But you can't disagree with my plan? It would mean you have an alternative (you probably don't), or that you wish that the trend continue and that mozilla die (you don't actually wish that)

Google can't even get its own house in order. Working with web brokenness is all to do with Chrome monoculture these days. You have to make crazy workarounds even for electron-based apps.

A monoculture inherently creates fragility and inflexibility without the need to compete or collaborate.

Chrome removing ad blocker functionality is one reason why Mozilla is sorely needed.

Unfortunately, people don't realize that they are the product with Google, and ever more invasive spying will always the norm with them.

Dont drink the Kool-aid.

Did you ever post your proposal to the mozilla dev mailing as requested by jdm?

I'd like to read the discussion but can't find anything in the archives.

No, I didn't because.. <shame>I'm too lazy and the site was ugly on Android</shame>

>If Firefox does not die by himself, expect Google to slowly stop paying them

If the web is going towards a Chrome monoculture, then perhaps Microsoft would step in to keep some competition in the marketplace.

Stranger things have happened!

Quantum has been out for some time but it seems have updated some of the dev tools as well as added in a new shape editor and fonts panel, something that is not available in chrome. looking forward to trying these out.

Would love to see something like this for Android (Android firefox doesn't have developer tools to save space).

It's not to save space, they could download it opt-in.. It's because they don't have the resources and have many P1 features to catch up.

You can turn on usb debugging and attach to it from your computer. Then go to about:debugging to inspect, add breakpoints, etc.

Even though I stayed away from FF because of it's macos issues, I found myself using it(normal ff) for development. I'm not the only one at work either.

It's a weird angle for a browser to compete from, but I guess it's working.

Am I crazy or does Option-Command-S not open the Debugger as it says in the Debugger tutorial?

You made me discover a new shortcut... <ctrl>-<shift>-<s> opens the screenshot tool ... too bad they removed the cloud hosting feature though...

Not a front end dev but this looks really neat.

What happened to Servo?

Did they fix the memory bloat on Mac OS? It’s pretty much useless to me on MacOS.

I saw a tweet recently about macOS power consumption fixes: https://mobile.twitter.com/pcwalton/status/11661937700160798...

Haven't noticed any memory bloat myself though (= it consumes a lot of memory, like any browsers). It's somewhere between Safari and Chrome.

Random question: Do you use safari? Do you see significant performance differences with chrome? Do you see broken/partially broken websites on safari vs chrome ?

I only use Safari on my personal computer. And at work I use Firefox for development, and Safari for personal stuff. I used to use Chrome but nowadays it's very rare (testing something specific).

The thing is, Chrome actually feels snappier than both Firefox and Safari when I am using it, probably because it's profile is empty (almost no extensions, no history, etc.).

I rarely see broken websites, with any browser. And usually it's because of my ad-blocking extensions. Google websites are slower on Firefox, but rumours are saying that it's almost on purpose… (one more reason to not use Google products like Chrome).

Good to know thanks. The last time I used it it was taking up nine gigs of RAM with only one tab open. Maybe I’ll give it a go again

Just a few days ago one of the first fixes got into the code (v70 already has this): https://bugzilla.mozilla.org/show_bug.cgi?id=1429522#c32

It's night and day on some pages, like google docs word document, for example.

PS: this one is about power\cpu usage though. Never really had any memory problems on my macs.

Wondering about this as well.

On my 2017 MacBook Pro with touchbar I don't notice the performance hit, but on my 2015 MacBook Pro the whole system freezes with Firefox. What completely solves this however is enabling 'low res mode' (right click app icon -> get info -> check the low-resolution checkbox). But, as the name suggest, Firefox now runs in non-retina resolution. It looks crappier, but I still prefer it to other browsers.

Should the title have a (2017) at the end, since Quantum initially came out then[0,1]? (I understand it points to a current-version but it seems like, from the comments, it's being mistaken for having recently been released.)

[0] - https://blog.mozilla.org/blog/2017/09/26/firefox-quantum-bet...

[1] - https://en.wikipedia.org/wiki/Firefox_version_history#Firefo...

No, i'm using the developer edition and got the update only this morning. But I don't know if something have changed...

Today updated to 70.b1, according to the About Firefox Developer Edition window.

It has a "What's new" link next to the version number, but it links to a 404: https://www.mozilla.org/en-US/firefox/70.0beta/releasenotes

EDIT - I put in a bug report on the release notes link, response says that release notes will come out next week

Hi there. You can also refer to the 70 nightly notes at https://www.mozilla.org/firefox/70.0a1/releasenotes/ for now.

The Quantum version of Firefox (and Firefox Developer) did not come out this morning. You likely just got a routine build update.

It bumped up from 69 to 70 this morning, a "major" update so much as browsers have those anymore. But it's not clear what's actually new in this.

Here's what they listed for 70.0a1 back in July, but the release notes link for today's update is a dead end.


Dev edtition gets the last nightly build two weeks before the same release hits Beta channel.

One of the main changes for me as a Mac user is this: https://bugzilla.mozilla.org/show_bug.cgi?id=1429522#c32

Ooh, that's a very nice change. I've stubbornly stuck to Safari on my laptop because of its low power usage.

Now I'd be losing the tabs/bookmarks integration with iOS's unchangeable default browser, but that's not as big a deal to me.

Does this mean Firefox won't always be the "apps using significant energy" list?

Web browsers are basically the new higher-level operating system. They likely should be using more energy than anything else considering how much we run in them.

Maybe. This was but one improvement in a backlog.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact