Firefox 52 Released With WebAssembly Support, Security Fixes, CSS Grid

[franglais125]

http://html5doctor.com/html5-forms-input-types/

Thanks, that was quite useful, as I got the picture now.

Here is the meta bug: https://bugzilla.mozilla.org/show_bug.cgi?id=888320

[lunarcloud]

(And bulletxt )

What's the deal with these input types? I keep seeing posts like this on every new Firefox announcement, but I remain ignorant as to what they would bring to the user/developer.

Any pointers ?

If a website has a date field, it will simply be a text field on Firefox, but an actual calendar widget on Chrome.
Chrome ensures a valid date. Firefox makes the developer of each and every website write their own validation and add text to tell the user the right way to write the date.
Or, more practically, it forces people to make their own calendar widgets and abandon hope of using the date field until Firefox supports it.

Ever been on a website on your phone where you need to enter a phone number and are given a full keyboard instead of just a number pad? Same sort of thing (although nowadays it's usually because the website author used input's normal/default text type instead of number type).

[franglais125]

If a website has a date field, it will simply be a text field on Firefox, but an actual calendar widget on Chrome.
Chrome ensures a valid date. Firefox makes the developer of each and every website write their own validation and add text to tell the user the right way to write the date.
Or, more practically, it forces people to make their own calendar widgets and abandon hope of using the date field until Firefox supports it.

Ever been on a website on your phone where you need to enter a phone number and are given a full keyboard instead of just a number pad? Same sort of thing (although nowadays it's usually because the website author used input's normal/default text type instead of number type).

Thanks for the explanation! Yes, especially on the phone that's really annoying.
Now I'm going to be one of the people complaining about the lack of date input support!

[jrch2k8]

I don't mean to sound rude but in general large projects don't much care for users personal technical objections based on personal reasons unless they are writing the code. The reason I suspect they are not depending on gstreamer for this is to reduce supported code (supporting pulseaudio is actually very easy comparatively) considering most of the Linux desktop distributions ship pulseaudio nowadays it's not that surprising to be honest.

I Guess OP is another "lennart is cthulhu" but anyway for all the other 99.99999% of firefox users this is a nice idea and probably will set multi tab stream into parity with other OS/Browsers and so far with ArchLinux have years without 1 glitch with PulseAudio, sometimes i even i forget is installed, is seamless this days and the latest memfds patches should have killed any latency issue remaining and jack is very easy to integrate with it this days too.

So no issue here

[stqn]

Using openal or portaudio or whatever instead of alsa/pulseaudio directly seems like the right way to go about it indeed. But Mozilla hasn’t done anything right since Firefox 4, so no surprise there.

[L_A_G]

WebAssembly

I hope I'm not the only one who thinks that "web" and "assembly" go together about as well as raspberries and mustard?

You generally want to keep arbitrary code from the web about as far away from low level access on your machine and what this does is basically undo an absolutely enormous amount of protections against low level code execution. With the rise of low level web standards it's almost as we're back in the "good old days" of ActiveX and how badly that got exploited all the time before people eventually understood why using it was a very bad idea and dropped it.

[Mystro256]

WebAssembly sounds nice, but does that mean that sketchy ads will now execute as blobs in a browser?

It still runs in the JS sandbox, so if your browser properly sandboxes JS, wasm will be nicely contained as well. The only difference is that those "sketchy ads" won't be slowing down your browser as much, but on the security side, nothing will change.

As well, It's less like a blob and more like a better version of asm.js, made easier to parse and smaller by being minified into a hex representation instead of text/characters. I guess you could call it the closest thing to js bytecode.

[Mystro256]

I hope I'm not the only one who thinks that "web" and "assembly" go together about as well as raspberries and mustard?

You generally want to keep arbitrary code from the web about as far away from low level access on your machine and what this does is basically undo an absolutely enormous amount of protections against low level code execution. With the rise of low level web standards it's almost as we're back in the "good old days" of ActiveX and how badly that got exploited all the time before people eventually understood why using it was a very bad idea and dropped it.

It's still run in the JS sandbox, so WASM can't run "arbitrary code" unless there's a bug in the JS sandbox in your browser. The security implications are the same as using JS, thus not nearly as horrible as ActiveX was. ActiveX deserved to die in the fire that it came from.

[Mystro256]

Using openal or portaudio or whatever instead of alsa/pulseaudio directly seems like the right way to go about it indeed. But Mozilla hasn’t done anything right since Firefox 4, so no surprise there.

That would make a lot of sense, because they could unify their sound backend on multiple platforms.

[cl333r]

wondering how much longer i will stay a [Firefox] user.

One reason for me: Chrom[ium] sucks at the top 2 extensions I use:

1) Chrom[ium] youtube download forbidden by Google
2) its adblock extensions suck compared to Firefox.

But I also use Chromium because it does lots of minor things better than Firefox.