Firefox 18 | Hacker News Reader

Firefox 18

thenextweb.com

105 pointsPankajGhosh70 comments

I've used Firefox since the time it was called Phoenix. I still use it as my default browser, but there are two quite simple features in Chrome that I'm really missing in Firefox:

1. Exemplary implementation of user accounts. I have a separate user account for my private Google account (Gmail, Calendar etc), a second for Facebook, a third for work (separate Twitter and Google accounts) and a third for web development. I know Firefox has a Profile manager and that it is possible to use -no-remote to run several profiles at once. But: it does not work very well at all. Not even with the https://addons.mozilla.org/en-US/firefox/addon/profileswitch... addon.

2. Incognito window can run in parallel with the "normal" browsing session. This is lovely when I want to see how a website looks in a clean browser. In Firefox the private session replaces the normal session until i chose to go back to "unprivate" again.

Seconding 2). Incognito is not just for porn, but very handy to check for cookie or session related troubles too. Chrome's implementation should be the default for all private browsing sessions in browsers.

Per-window private browsing is implemented and available in Nightly according to [1] (so, most likely in Aurora after the release shift this week). So hopefully it'll be in Firefox 20.

Meta-bug for per-window private browsing in Bugzilla is [2].

[1] https://hacks.mozilla.org/2012/12/firefox-development-highli...

[2] https://bugzilla.mozilla.org/show_bug.cgi?id=463027

Great news!

> Chrome's implementation should be the default for all private browsing sessions in browsers

I actually like to have an option to have a private tab. If it's just for a quick test then windows is fine, but if I want to continuously work with two logins on one page or something similar then I found the tab to be more appropriate.

There are UX problems with the Chrome style, though. Many, many times I've debugged issues developers are seeing because they thought they had closed their last incognito window and then opened a new one, but there was another one minimised in the background somewhere that persisted their cookies.

Yes, the Firefox mode is far more awkward for quick testing, but it does have the advantage of making it explicitly clear when you start and end your private browsing session.

So true. Took me a while before I realized each Incognito Window wasn't isolated (but shared cookies etc).

A better behavior here would perhaps be that after having the first Incognito Window open, the next would be become "Incognito Window 2". The "detective icon" could have a number overlay and each window would isolated. If you want to share data between the open pages the natural thing to do would be opening a new tab.

+1 to numbering the incognito sessions. Sometimes I need to simulate two user accounts polling some web service, or perhaps simulate a chat-like environment. Point is: I need two separate cookie jars. I've accomplished this using profiles, or by using my incognito window for "user A" and a profile window for "user B." However, there's less overhead with incognito windows and for development scenarios (which are inherently ad-hoc and temporary), overhead sometimes makes all the difference.

I'm on the nightly (version 20) and they added the feature where you can run private windows with non-private ones concurrently. Took them long enough..

#2 is shipping as part of Firefox 20.

Built in PDF viewer is the only Chrome feature I wish Firefox had.

PDF.js is Mozilla's project to implement plugin-less support for PDF in Firefox. Please give it a try.

https://addons.mozilla.org/en-US/firefox/addon/pdfjs/

PDF.js comes builtin with the recent releases of Firefox.

http://www.ghacks.net/2012/09/08/enable-the-built-in-pdf-rea...

And it's enabled by default in the current Aurora (pre-beta) test builds:

https://www.mozilla.org/en-US/firefox/19.0a2/auroranotes/

Hooray!! Thank you so much for telling me about this.

Firefox also missing: Close tabs to the right/left of this tab instead of only "close other tabs".

This is a feature that I figure is available in many addons that add functionality to the tabs in Firefox. One of the pros with Firefox is that "low level" GUI behavior can be modified.

Tree Style Tab is my favorite example of this: https://addons.mozilla.org/sv-se/firefox/addon/tree-style-ta...

(Poorly informed guess ahead) That may be a consequence of the "each browser tab in its own process" approach I've read somewhen that Chrome uses and Firefox does not, or not yet.

How does Chrome handle user accounts?

I use Chromium and have a few Google accounts and find switching profiles a chore (first world problems I know). If there's a feature I'm overlooking then that would be fantastic.

Lifesaver: Ctrl+Shift+M

An ugly hack for #2 is to run "firefox -no-remote -private -ProfileManager". This will pop up a new session of firefox that is in private mode. However, you will have to choose a profile other than the one you are currently using.

This isn't officially released yet.

If you're the type who would be willing to install this, you should consider using the Beta channel instead. Beta releases are generally very close to release quality, and you'd help gather telemetry data earlier.

I've been using the beta channel since FF6 (maybe) and, as you said, it's been quite reliable. It's fun because you feel like you're somehow an insider and privileged because you get features (that no one but you would care about!) earlier than everyone else. Yes, these are quite silly feelings, but what the heck.

Sounds like you might enjoy the Nightly channel, in my experience it's been pretty stable.

Aurora is a good compromise. I was on nightly for a few months, and I had one or two times where something would break. Usually it was just a GUI element getting misaligned from some change. The worst was when it broke web video for a day.

I wish there was a reliable Aurora channel on Arch. The pkgbuild system isn't well suited to keeping up to date with a moving filename target like Firefox.

I agree, nightly does break badly from time to time. Beta or aurora are better bets.

I've switched to Aurora for the past month and it's noticeably faster than FF18b on my older hardware.

I've used Aurora for some time, extensions often get borked or don't work, so I went back to Beta.

btw one tip for people who switch to beta/aurora - disable extensions compatibility check.

There's an Arch developer named heftig who maintains a custom repository with Aurora. I've been using it successfully the last few months. I think aurora is updated daily, but I haven't watched it closely.

Info: https://bbs.archlinux.org/viewtopic.php?id=117157 Repo 32-bit: http://pkgbuild.com/~heftig/repo/i686 Repo 64-bit: http://pkgbuild.com/~heftig/repo/x86_64

Another option is to download the latest build [1] and use the built-in updater.

[1] https://ftp.mozilla.org/pub/mozilla.org/firefox/tinderbox-bu...

They do this every time: Make a news event around the appearance of a new Firefox release on the FTP site before it is officially announced/released.

Although this stunt seems considerably stupider now that everyone running a recent version of Firefox will automatically get updates...

If anyone decides to change the release channel, you can do it now by modifying 'app.update.channel' in about:config to nightly, aurora, beta...

Update: http://support.mozilla.org/en-US/questions/940799

I'm on the stable release, and I just got the v18 release by going to Help --> About and installed the automatic update. Their release notes also indicate that version 18 is now released as well: https://www.mozilla.org/en-US/firefox/18.0/releasenotes/

Initial support for the CSS Flexbox Module has been landed. It is disabled by default but can be enabled by setting layout.css.flexbox.enabled to true.

This is actually a feature I'm waiting to see on at least both Chrome and Firefox (without vendor prefixes). I haven't been tracking the spec closely enough to see how close it is to being finalized since the "new" flexbox model was adopted.

Other changes that interest me for fairly obvious reasons:

   Preliminary support for WebRTC.
   New HTML scaling algorithm.
   Performance improvements around tab switching.
   Improvement in startup time through smart handling of signed extension certificates.
   Support for W3C touch events impemented, taking the place of MozTouch events.

The last one for better, less specific Javascript.

Firefox implements it without prefix (it is finalized). It will be enabled by default in Fx 20. It supports both horizontal and vertical flexboxes but not yet multiline flexbox.

Opera supports it fully and without prefix too.

Webkit tends to be (very) late in unprefixing (they are the only ones with prefix for Animation, Gradients, Transitions nowadays.)

Browsers that support TouchEvents: - Chrome (desktop and mobile) - Safari (mobile) - Firefox (mobile, and now desktop)

Browsers that support Microsoft's proposed PointerEvents: - Internet Explorer (desktop and mobile)

I get in theory why having a combined touch/pen/mouse API is a good idea, but MS's proposal makes handling multiple touches concurrently much harder than it should be. It'd be nice if they'd get over themselves and implement the same standard as everyone else (especially if they want to be relevant in a world where touch interactions are designed primarily for WebKit, while they have vanishingly low market share in touch-first devices.)

Isn't there a patent by apple related to the touch events that makes it unsafe and risky to implement them?

http://blog.jquery.com/2012/04/10/getting-touchy-about-paten...

Does this apply in the EU?

It might or it might not. In theory, we don't have software patents, but some tend to still slip through (maybe this one did? I don't know).

It's totally beside the point anyways because the moment as a browser maker has an office in the US and is distributing the infringing software there, they can be sued there.

MS, as far as I remember, has an office in the US :p

From what I've heard Microsoft's PointerEvents offer a better complete model, without creating a web-standard which involves risks of getting sued by Apple should you ever dare to implement it.

For once, it seems Microsoft has a better option.

And no, just because it's in WebKit doesn't make it better, much less an official standard. WebKit is the new IE and it's destructive for the web. People design for "Webkit" as if it was the standard, and expect everyone else "to eventually come along". We did see how well that plays out not that many years ago, but then we did it for MSIE.

As a 5 minute user of Firefox mobile, I can attest to this happening right now. The mobile web is at this point not even usuable on a non-webkit browser.

This is bad for the web. You are doing it. Please stop doing it.

The only thing good for the web is following web-standards. If you're not complying to them, you are doing something wrong, and I don't care what your reason is: You are doing something wrong.

> For once, it seems Microsoft has a better option.

Not for the first time. * { box-sizing: border-box; } is a staple of mine.

> WebKit is the new IE and it's destructive for the web.

Microsoft has submitted a patch to the WebKit project to extend the open source rendering engine with a prototype implementation of the Pointer Events specification that the company is also working on together with Google, Mozilla, and Opera. ... The first specification, Touch Events, has been essentially abandoned.[1]

> The mobile web is at this point not even usuable on a non-webkit browser. This is bad for the web. You are doing it. Please stop doing it.

There's no need to hector web developers about this. (I assume that's what you meant by 'you'.) It's mostly up to the browser makers to hash this out. Microsoft's contribution to Webkit was an exemplary and, from what I can see, much-welcomed response to Apple's patent shenanigans over Touch. Let's just give it time.

It's worth bearing in mind that Android (which ships Webkit by default) is currently the only mainstream mobile OS that even allows unrestricted use of independent browser technology, which is partly why Firefox Mobile is only available for that platform.

[1] http://arstechnica.com/information-technology/2012/12/micros...

I forgot about that. Wonder when/if it will land in Chrome...

I'm not only partial to TouchEvents because they're better supported across browsers. IE makes you handle all the pointers independently, which is great for toy demos of drawing apps, but really cumbersome for actually implementing gestures with compared to simply iterating over TouchEvent.targetTouches.

Does anyone know the state of a Modern UI (what used to be called Metro) version of Firefox. I've recently switched to Windows 8 and love it apart from the lack of a native Firefox, obviously I can run the desktop version but I'd rather use the new UI where possible.

The last release/mention I found was a preview [1] in October last year and then nothing but tumbleweed ...

[1] https://blog.mozilla.org/futurereleases/2012/10/04/firefox-m...

(Mozilla dev here)

I can guarantee you that people are working on it. I don't know the specifics, but it's coming along. You can follow the commits here [1]. Of course because it is free software, if you are not afraid to test a early prototype, you can get a build for every push to the repo. It is a bit complicated (lots of clicks in our custom CI interface), tell me if you are interested.

[1]: http://hg.mozilla.org/projects/elm/, the elm branch being the repo where people work on Win8 support.

Thanks padenot, it would be great if you could point me in the direction of where to find the builds.

- Go to [1] (our CI)

- Find a green "B" on a Win opt line (meaning that a build has succeeded for a Windows Optimized build), click on it, a panel appears at the bottom

- On the bottom left of the page, you can see a link named "go to build directory". Click on it, that sends you to the location where the CI stores the build

- In the list, there is an installer executable, that you can use to install this particular Firefox build (for examples: "firefox-20.0a1.en-US.win32.installer.exe").

Please keep in mind this is very much a work in progress, and it is likely to break horribly, maybe destroying your computer in the process (you never know :D). No warranties, blah blah blah. If you find bugs, it would be much appreciated to have them filed on bugzila.mozilla.org.

[1]: https://tbpl.mozilla.org/?tree=Elm

Hi, I'm an engineer on the Metro Firefox team. Here's the best place to get a nightly build with Metro enabled. Needless to say, these are unsupported development snapshots and are not ready yet for everyday browsing:

http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/lates...

We hope to enable Metro in the main (mozilla-central) nightly builds within the next few weeks.

Please don't make the same mistake than Chrome (ie. make a Metro app which almost prevent using the desktop version).

http://code.google.com/p/chromium/issues/detail?id=140347

Can anyone confirm whether final flexbox support was added? http://caniuse.com/flexbox suggests that FF 18 only has support for the old version of the standard.

As referenced from Mozilla[1], Firefox and Opera support the new spec without prefix, Chrome supports it with the -webkit prefix, and Internet Explorer supports the old-flexbox.

[1]: https://developer.mozilla.org/en-US/docs/CSS/Using_CSS_flexi...

That page says it's behind a developer flag, and only limited single-line support is provided. Is the page up to date, or is the feature fully implemented, and accessible to anyone, regardless of flags?

Firefox 20 will contain css3 flexbox unpreffed and unprefixed. Assuming nothing is found that would make us flip the pref to false during the Aurora or Beta cycles.

The page is up-to-date.

According to the channel on freenode, full, up-to-date, and prefix-free flexbox support is on track to be enabled in the stable version of Firefox that is released exactly 12 weeks from tomorrow.

> FIXED: Disable insecure content loading on HTTPS pages (62178).

Very nice. Check out the test page here: https://people.mozilla.com/~bsterne/tests/62178/test.html

Also, if you can, switch your site to default to HTTPS (redirect from HTTP to HTTPS right away), especially if you run some kind of an API. I am looking at you Google, with your Charts API. Last I checked the only way to get those via HTTPS was through one employee's epic quest to make their API's HTTPS-capable.

Huh? :-( http://i.imgur.com/taamA.png

The relnote is subtly wrong. The functionality has been landed behind a pref, but it's not on by default yet because there are a few UI issues to sort out.

I used firefox up until I ran into a nasty Javascript bug which caused Firebug to conk out, never hitting the breakpoint. I switched to Chrome and the breakpoint was hit every time. I got the overall impression that Firefox's debugging tools weren't as good as Chrome's. Has this changed at all?

Firefox now has a built-in debugger (which appeared in Firefox 15, released in August 2012). This uses a new debugging back-end built by Mozilla's JS and devtools teams, totally separate from the older Firebug debugger code. It should be more robust, and has new features like remote debugging of code running in Firefox for Android.

More info at https://hacks.mozilla.org/2012/07/debugger-responsive-design...

Please speed up SVG animations in Firefox 19. d3.js animations can be painfully slow in Firefox when they aren't in Webkit.

You should fill a bug with a demo in http://bugzilla.mozilla.org (with timing in Fx and in a Webkit browser to show the problem)

They won't read you here.

That is not how Mozilla development works. Code changes go int the trunk/nightly builds currently FX 21. Only rarely does code get uplifted to Aurora or Beta. Usually this is because something regressed.

Please file a bug with example code. https://bugzilla.mozilla.org/enter_bug.cgi?product=Core&...

The retina support is really welcome.

Up until now, using firefox on a retina mac was kind of disappointing.

I'm still using 3.5.7. What more could you want?

Security updates? The huge number of performance boosts and resource use reductions later versions have seen? The very large number of css properties and selectors added?

Firefox version numbers could hardly be any more meaningless.

Let's not have this discussion again.

A > B is the only thing that matters. Anything else is bullshit that you can't trust and still have to read the patch notes for anyway.

Correct, which is why Mozilla themselves removed most of the version numbers from their website, and you have to go digging in sub-menus to find out which version you're running.

Reply on news.ycombinator.com