News for hackers. Stuff that matters.
Re:Sponsorship needed for new OS/2 web browser (2017) by mhd on Thursday July 05, 2018 @02:08AM (#17461166)
So what makes Rust impossible to port to OS/2?
Re:Sponsorship needed for new OS/2 web browser (2017) by ofrzeta on Thursday July 05, 2018 @02:27AM (#17461224)
They said "unlikely" not "impossible". I guess it is unlikely because no one is willing to put in the massive resources needed to port Rust to an archaic platform.

Here's a link about porting the Rust compiler to the Haiku OS: http://rust-on-haiku.com/wiki/PortingRust

Please note that Haiku is among the supported OSs of the LLVM compiler framework while OS/2 isn't. So unless you want to re-create a Rust compiler from scratch you would first need to port LLVM to OS/2.

Re:Sponsorship needed for new OS/2 web browser (2017) by mhd on Thursday July 05, 2018 @04:31AM (#17461703)
Yeah, the backend not being available is certainly a showstopper, thanks. GCC being ported to every silicon-based system under the sun does blind one to the fact that that isn't necessarily true for its rivals.
Re:Sponsorship needed for new OS/2 web browser (2017) by _verandaguy on Thursday July 05, 2018 @09:02AM (#17462896)
It probably helps that GCC was around when OS/2 and other now-archaic systems were relevant.
Re:Sponsorship needed for new OS/2 web browser (2017) by mkj on Thursday July 05, 2018 @03:33AM (#17461495)
mrustc might be an option to avoid LLVM.
Re:Sponsorship needed for new OS/2 web browser (2017) by Arnt on Thursday July 05, 2018 @02:30AM (#17461234)
Writing an OS/2 backend for a language that uses LLVM isn't impossible, but it's hard, hard work. Writing or porting a linker, implementing the low-level parts of the Rust runtime.

Fixing bugs likely requires being fluent in all of Rust, C++, LLVM IR and OS/2's variant of x86 assembly, and tracing causes from the Rust frontend through a dozen layers and subsystems until their effect show up in x86. You may think I'm exaggerating.

Re:Sponsorship needed for new OS/2 web browser (2017) by golergka on Thursday July 05, 2018 @04:58AM (#17461798)
I don't think that you're exaggerating, but writing a browser is also a lot of very hard work. And writing a browser is only writing a single application; implementing a LLVM backend, however, is supporting thousands of potential applications. Wouldn't it make much sense to raise funds towards this goal?

Also, browser is one of the main attack surfaces in modern web, and writing your own browser instead of sticking with a huge project that benifits from researchers all around the globe just doesn't sound right. Especially since OS/2 has a lot of legacy use in corporations that can be targetted for an attack.

Re:Sponsorship needed for new OS/2 web browser (2017) by exikyut on Thursday July 05, 2018 @06:38AM (#17462155)
> thousands of potential applications.

> ...OS/2

...probably not. Wouldn't mind if that weren't the case, but that's fundamentally a legal problem IMO (impossible to open source etc).