Firefox 71 | Hacker News Reader

Firefox 71

hacks.mozilla.org

211 pointsblendergeek65 comments

The increased performance of the developer tools window is highly appreciated. Please prioritize #1219917 [1] to allow a single undocked developer tools window to be reused across all tabs! Would be such an improvement for multi-monitor developers.

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=1219917

Oh, man. I want this so bad. But the Bugzilla thread makes it sound like it's a really hard thing to do (grrr).

I bet. Safari has it, as a small drop down hidden in the corner of the window. Easy enough to switch tabs if you remember it’s there.

Firefox just keeps looking better and better. Lots of features added and QoL increases. I have switched over to using it for general browsing, but continue to use chrome for development. That might be changing real soon!

I frequently find myself bouncing between them for development, but lately...I'm not really sure why. FF has everything I need. Even a dark mode (Chrome does too, I think).

I might be a top 5% Firefox fan and I dislike Chrome. I'll still keep a Chromium instance around to test in when I do frontend work.

Luckily though if it works in Firefox then it usually[0] works in all other browsers as well.

[0]: While doing Android development I can remember once where it failed in Chrome but worked in Firefox.

You should try Firefox Developer Edition. Once I started, I ditched Chrome entirely.

Same

What's the difference other than being an alpha (2 major versions ahead) and having some nicer default settings?

No Google telemetry, for a start.

I'm the opposite - I use Chrome for general browsing and Firefox for development. Firefox's developer tools are better for the types of work I do.

I'm guessing more CSS than JavaScript debugging for you? That's where the difference in dev tool strength is for me at the moment, at least.

I use FF for both, only using Chrome to test and debug cross-browser. I haven't really found I needed anything more than FF offers for JS debugging, but I guess you can't miss what you don't know.

It's actually purely a performance issue, and I honestly don't know what is causing it. I suspect there might be certain quirks in certain frameworks that I use that make the FF debugger slow down to a crawl, while Chrome handles it just fine.

Ironically today's bump of Firefox dev edition to 72.0b1 introduced a really ugly UI regression on macOS. However the QOL improvements are nice.

I do like Firefox for political reasons, but the UI is pretty bad looking on macOS compared to Chrome or (especially) Safari.

The contextual menus keep getting weirder and weirder looking so now I guess it's hard to tell if they support Dark Mode or not.

FF is my browser of choice now. I work with large XMLs as a part of my dev process and Chrome is simply unresponsive with large XML files that must display on the page.

Websocket message inspector, yay!

That's a big "I would use Firefox if ..." vanishing, at least judging from the most common complaints here.

Well, playing with SignalR and securing real-time web apps just got more interesting.

Along with full-text network panel search--I didn't even know I wanted that!

I switched from Chrome to Firefox 6 months ago when Google started playing games with it's ad blockers.

All of my Chrome extensions and features were supported by Firefox and I was a happy camper.

Then two problems got to me:

1. Slow performance on web apps like Gmail (use it for work)

2. Firefox regularly crashes (more than 10 times per day)

I'm trying to switch to the Brave browser (Chromium) now, but I was almost a Firefox convert.

> 2. Firefox regularly crashes (more than 10 times per day)

Needless to say, crashing 10 times a day is not normal. :)

Can you share any Crash Report IDs from your Firefox's about:crashes page? The crash reports will say whether a bug has already been filed or fixed.

It might be your setup, I do not experience that. (I too use Gmail in FF.)

Firefox does not seem to crash on me. However, the FF performance has never been up to par. Especially with Google products(Gmail, Docs, etc.).

Doesn't that make you suspicious?

Too bad Electron isn’t based on FF.

Indeed. Scrolling Monaco editor here https://microsoft.github.io/monaco-editor/ feels much smoother on Firefox than on Chrome for me in my very unscientific macOS test.

I hope one day we get a new embedded FF through Servo or something.

I remember how excited I was back in the old days when I could change KDE's browser Konqueror's engine from KHTML to Gecko in just a setting (or was it a tab?)

There used to be a similar project, PRISM: https://en.m.wikipedia.org/wiki/Mozilla_Prism

Products like Zimbra Desktop were based on it.

I love everything related to css grids. Subgrids enabled by default is going to simplify my life greatly. Thanks, Mozilla!

Picture-in-picture is a game changer. This is awesome!

how do you activate it

Picture-in-picture video comes to Firefox for Windows: Select the blue icon from the right edge of a video to pop open a floating window so you can keep watching while working in other tabs. Learn how the feature works.

The FF dev tools is one of the reasons I switched. IMO, they are miles ahead of Chrome. Or at least, they were when I switched up until a few months ago when source maps stopped working. Without the possibility of debugging, it makes the dev experience much worse. I wish they would focus on fixing crucial bugs like that instead of speeding up the loading time. One gets you the dev tools a few milliseconds faster and the other one is a core functionality of the program. And yes, they work fine in Chrome and the way they are generated hasn't changed in 5 years.

Did the FireFox people ever explain why they abandoned a more typical version numbering system and instead started releasing a new major version every couple of weeks?

They have a write up about it here https://hacks.mozilla.org/2019/09/moving-firefox-to-a-faster...

Yes, see this[0] for example.

If memory serves me right, the TLDR is that the more typical release management system can cause certain features to not be released as soon as they're ready, because other features that should go in a major release are not ready yet.

By removing the semantic component of release numbers and releasing often, this does not happen. Feature ready -> shipped soon.

[0]: https://hacks.mozilla.org/2019/09/moving-firefox-to-a-faster...

I'd say that Firefox versions are mostly a unit of time now, since they aren't generally semantically meaningful in the way semver is. While working on Firefox, we often make plans that are delineated not in weeks, months, and quarters, but in Firefox releases and the nightly/beta/release cycle associated with them.

Well, actually it was introduced after Chrome switched to this same model. So it looked like Chrome is advancing faster because they released a new major version every few weeks and it got more press. I suppose that's why they really switched. If it would be just a matter of the management system you'd better adapt the system instead of version numbers. Anyway, after so many "major" releases the press has died down. So what we have now is version numbers that are quite meaningless because you can't pin major changes in the architecure to version numbers anymore.

The recent change to having a releaes every 4-5 weeks is a small modification of the existing practice of having a new version every 6-8 weeks.

The 6-8 week schedule is about a decade old now. It's original goals were to help ensure that unrelated features didn't block each other from releasing, and make releases more consistent, which helps localization, testing, and project management. In that regard I think it's been very successful.

It's probably not worth nitpicking a decade old decision though.

Google started it with Chrome, and Firefox had to adopt it too or look like an also-ran. More info here: https://news.softpedia.com/news/Firefox-to-Copy-Google-Chrom...

Chrome was three years old and on version 10.0 when this was published, and Firefox was on version 4.0 despite a decade of Mozilla legacy. It’s marketing, plain and simple, but I think it was the right decision.

I feel like consumers are smarter than this? I don't think anyone chooses iOS because it's on version 13 versus Android's version 10.

The difference also becomes less meaningful as the disparity grows. If Chrome is on version 78, and Firefox is on version 6 (or whatever), it's relatively intuitive that "versions" don't mean the same thing between the two. Same with the Playstation 3 versus Xbox 360, really.

Most people probably don't even know what version they're running, anyway.

Consumers are in fact this dumb.

That's why it was the Xbox 360 and not the Xbox 2, because XBox 2 < Playstation 3.

I'm aware that this was Microsoft's original thinking behind the Xbox 360's name, but then they called their next console "Xbox One".

That felt to me like Microsoft deciding that their original concerns had never been legitimate.

"One" instead of "1" suggests unity and oneness, rather than a version number.

I still think it was an awful branding decision, though.

Considering how the PlayStation has dominated the market without messing around with branding like that, it seems like MS dropped the ball by bike-shedding the name instead of nurturing quality content.

<ArmchairExecutive> Personally, I think Microsoft should have put way more effort into polishing their Kinect user experience. The X-bone could have been everything that the Echo/Siri/Google Home ecosystems evolved into if Microsoft had just worked harder on polish up front.

Instead, Microsoft pretty much alienated their core gaming base in order to half-assedly fail at a broader goal. They didn't even manage to push through more than a couple of decent Kinect v2 games </ArmchairExecutive>

PS didn't dominate that generation. It outsold the 360, but only by about 5%.

Ironically the PS 4 did outsell the X1 almost 2:1.

That doesn't hold up because they went Xbox One vs Playstation 4.

At this level, the market demands non-linear value. http://achewood.com/index.php?date=01132006

Same reason it's such a cliché for the fourth movie in a franchise to drop the number and have a subtitle like 'Revolution' or 'Revelation' or 'Revenge'. I assume marketers imagine the public's reaction to a '4' like "Look at you with that new _sequel_. How could you BE so cocky?!"

I don't think anyone chooses iOS because it's on version 13 versus Android's version 10.

Users generally choose the phone, not the OS. And the most popular phones are pretty close on the numbering schemes.

Another example is SSL and TLS. SSL 3.0 was the last version, then TLS 1.0 took over, but a lot of people still thought SSL 3.0 was the latest and best.

Most non IT people I know, use Chrome because "one day it was just there". Installed through some freeware program where they forgot to uncheck a box.

Yeah...people are not that smart.

I was there when this decision was discussed and made. It was nothing to do with marketing.

The really important decision was to switch to a rapid, time-based release cycle from a traditional long, feature-based release cycle. We wanted to reduce the amount of work that was "fixed in the next release" but not yet shipped to users; to increase our responsiveness, so that if a new feature suddenly became important, we could quickly get it into the hands of users; and to avoid developer issues like schedule chicken and rushing to get not-really-done features into a release because the next release is far away. This change was indeed inspired by Chrome, and it worked well; those goals were achieved.

Then a much less important question was how to do version numbering in a world where you're doing these frequent releases. There is no technical reason why some releases should bump a major version number and others should not; the amount of new features or breaking changes will fluctuate from release to release, but the same kinds of changes are present in every release. It is therefore simpler to just have a uniform numbering scheme so no-one has to decide whether a release should bump the major version number, and users aren't trying to understand the difference between a major version bump and a minor version bump (when there isn't really a meaningful difference).

I don't care either way, but they'll have to switch to something else at some point, cause the numbers will look ridiculous. Firefox 157... Firefox 349... Firefox 1071

Intel's processor names already look ridiculous to me. Core i9-10980XE? Can't say that with a straight face, "normal" people seemed confused way back when Sandy Bridge got released

They could do like NVIDIA and switch back and forth between two, three, and four digit model numbers every decade or so as the numbering scheme gets stale. I remember having a GeForce2 MX400, then a Ti4600, then a 8600, then a GTS 250, then a GTX 460, then a GTX 970, then a Quadro P4000, and now they seem to go for a hybrid scheme where they have the two-digit "16 series" or "20 series" containing four-digit models like 2080 :p

Why is that a ridiculous name? It's from the Core X family of processors, is an i9 (which generally means more cores and cache than an i7), it's a 10th gen processor, which since 9th gen, also means that it has hyper-threading, as opposed to 9th gen and later i7s), and most importantly, as a product name is unique enough to make it easy to look up the technical specifications.

Considering `less` is at version 551, they still have quite some time left.

> There is no technical reason why some releases should bump a major version number and others should not; the amount of new features or breaking changes will fluctuate from release to release

Then why have three parts to the version? You could just have the cycle version based on the schedule and a patch/fix subversion.

ESR releases get minor version bumps. Eg 68.3 got released at the same time as 71.

I'm not sure what you mean by "more typical." If you mean the "major.minor.bugfix" thing that was the norm for a while, I think we've all realized it's basically useless. So many projects had no criteria for incrementing the major number, so it would just sit at 0. or 1. forever, informing no one of anything. So don't think of it as "a new major version every couple of weeks," think of it as dropping the implied (and useless) "1." major version with a new minor release every couple of weeks :-)

But now actual major changes are unnecessarily hard to communicate to non-technical users. If e.g. the "quantum" activation could have been celebrated with a once-in-a-couple-of-years major increment, far more people would have taken notice, with much less marketing effort. A rarely incremented major number is a marketing tool, a continually incremented number is merely a vector clock.

Just my 2 cents. I think that the traditional versioning system is most useful for libraries and other shared components. It is much less useful for "end programs" for end users (aka applications).

Reply on news.ycombinator.com