News for hackers. Stuff that matters.
Re:V8 Release v8.6 by benatkin on Monday September 21, 2020 @04:53PM (#24548352)
Another reminder to check out WebAssembly. Both updates in the release notes about WebAssembly sound very promising. One is getting towards complete parity with native code, and the other is reduced overhead when interfacing with JS.

I looked around the web and found this:

https://github.com/appcypher/awesome-wasm-langs

...which led me to this:

https://github.com/AssemblyScript/assemblyscript

Its description is awesome: "Definitely not a TypeScript to WebAssembly compiler :rocket:"

The thing that makes me most not want to miss the boat on WebAssembly so far is that Figma uses it (with C++):

https://madewithwebassembly.com/showcase/figma/

Re:V8 Release v8.6 by pjmlp on Monday September 21, 2020 @05:08PM (#24548505)
Re:V8 Release v8.6 by rvz on Monday September 21, 2020 @05:46PM (#24548826)
> Another reminder to check out WebAssembly.

Right. The actual reason you see all the big companies loving WebAssembly is that now you can fully have cross-platform black-box DRM binaries running on every browser. But of course what could possibly go wrong?

It is another Adobe® Flash® that is controlled by the big tech companies rather than just one.

Re:V8 Release v8.6 by gpm on Monday September 21, 2020 @08:34PM (#24549787)
Webasm doesn't help much with drm since you can reverse engineer it just like any code. It doesn't have any privileged access to allow tricks like using a key stored in hardware (and other drm does).

It does allow blackbox libraries. So does compiling to JavaScript though. Really it just means that the black box binaries are (even) faster than the white box scripts.

Re:V8 Release v8.6 by untog on Monday September 21, 2020 @05:58PM (#24548908)
We already have black box DRM binaries in browsers. That battle is lost.

WebAssembly is not another Flash. It might allow someone to create something that resembles Flash, but WebAssembly itself is not. You might as well criticise C because people can write DRM with it.

Re:V8 Release v8.6 by nonbirithm on Monday September 21, 2020 @06:17PM (#24549054)
One point that I don't seem to hear a lot: You can still block ads on YouTube in the browser, but you can't in the mobile app. The difference is that in the browser you can open the page and inspect everything that it consists of. In the app everything is bytecode. The solution of stripping ads from the YouTube app is to maintain a hacked version and follow a convoluted multi-step process to root your device and disable signature verification just to be able to install it. What will the solution be when webpages bundle both the content and advertisements into inseparable binary black boxes?

I think the population that consumes YouTube in general is much larger than the population that consumes paid content with DRM, and there would be a much larger pushback if webpages like YouTube started to be distributed in compiled format.

Re:V8 Release v8.6 by qppo on Monday September 21, 2020 @06:30PM (#24549144)
This has nothing to do with bytecode and everything to do with the security models used by Android and iOS.
Re:V8 Release v8.6 by untog on Monday September 21, 2020 @06:40PM (#24549216)
Ad blockers don’t work by inspecting the source code of YouTube’s JS and modifying it. Instead they inspect the network requests being sent, cookies being stored, elements being created and with which styles, and so on.

WebAssembly changes none of those things.

Re:V8 Release v8.6 by tgragnato on Monday September 21, 2020 @06:03PM (#24548952)
Isn’t EME limited to media sources? WebAssembly is not.

It’s not a tiny difference, and skepticism is not necessarily unfounded.

Re:V8 Release v8.6 by jopsen on Monday September 21, 2020 @06:15PM (#24549039)
Sufficiently obfuscated JS isn't easy to read either.
Re:V8 Release v8.6 by rad_gruchalski on Monday September 21, 2020 @06:53PM (#24549279)
I believe you mean “deny box”?
Re:V8 Release v8.6 by TheDong on Monday September 21, 2020 @06:03PM (#24548957)
Interestingly, webassembly is also helping make an open-source flash player possible.

Ruffle [0] is written in rust, compiles to webassembly, and already can play a reasonably sized subset of flash correctly at their intended speed.

[0]: https://github.com/ruffle-rs/ruffle

Re:V8 Release v8.6 by saagarjha on Monday September 21, 2020 @06:09PM (#24549003)
Another than can run Flash, among other things: https://leaningtech.com/pages/cheerpx.html
Re:V8 Release v8.6 by taf2 on Monday September 21, 2020 @09:26PM (#24550003)
This is awesome thank you for some reason last year my daughters school still needed a flash player for a typing program going to try this to see if works thanks!
Re:V8 Release v8.6 by x87678r on Monday September 21, 2020 @06:21PM (#24549083)
I always thought webassembly would be great that you can write rich guis in languages like Python or Java, but instead its mostly low level with Rust and C++ having momentum. Is it just not ready yet? or do I misunderstand the goals.
Re:V8 Release v8.6 by etaioinshrdlu on Monday September 21, 2020 @08:05PM (#24549673)
It could be that the bundle sizes for the user to download are smaller with low level languages. Scripting languages (and Java) have large runtimes.