News for hackers. Stuff that matters.
Re:PlayStation 4 System Compiler Support Landing in LLVM by forgottenpass on Thursday January 29, 2015 @12:58PM (#8967749)
Is there anything interesting about this? It's not like end users will get any utility from this, do development studios? Or is this just a way to ease Sony's maintenance burden?

Yeah, it's probably a good idea in the grand scheme of things to get their patches back upstream. But is it worth reporting on? Worth posting/upvoting on social media? It doesn't actually mean anything does it?

Re:PlayStation 4 System Compiler Support Landing in LLVM by kentonv on Thursday January 29, 2015 @05:14PM (#8969226)
I find it interesting in light of the licensing debates between GCC and Clang. GNU goes to great lengths to ensure that anyone modifying or extending GCC is legally obliged to release their changes -- including technical design decisions that actually make GCC's code difficult to reuse even in GPL-compatible projects. Clang is under a permissive license, so Sony has no obligation to release their changes, but apparently they are anyway.

I realize that GCC's policies were actually important for getting Next to release their Objective C code in the 90's, but it seems like times have changed. Companies have learned that keeping their code a secret actually isn't a huge competitive advantage whereas releasing it has a lot of real benefits (such as reduced maintenance cost). It's always better to have people doing the right thing because it's in their interests rather than because you forced them to.

Re:PlayStation 4 System Compiler Support Landing in LLVM by Nursie on Thursday January 29, 2015 @05:29PM (#8969312)
The devkit for the PS3 was gcc-based IIRC, though I have no idea how Sony behaved with respect to releasing stuff back to the community.

OTOH the main chip was an IBM product so it might have been big blue that we should look to.

Re:PlayStation 4 System Compiler Support Landing in LLVM by cperciva on Thursday January 29, 2015 @03:37PM (#8968681)
It's useful to Sony to reduce their maintenance burden. It's useful to everybody else because there are probably pieces of their code which can be reused or adapted for non-PS4 platforms.

Heck, even if the code is just used as a reference by people saying "gee, I wonder how Sony solved this problem" it's still very useful.

Re:PlayStation 4 System Compiler Support Landing in LLVM by krschultz on Thursday January 29, 2015 @02:29PM (#8968294)
People upvote what they find interesting. I don't own a PS4 and I'm never going to write code for it. I find this interesting because it is a good example of the value of upstreaming patches. The next time I give a talk about open source, maybe this is one of the examples I cite because the PS4 is something everyone knows.
Re:PlayStation 4 System Compiler Support Landing in LLVM by Sanddancer on Thursday January 29, 2015 @01:32PM (#8967973)
It could potentially be interesting if they worked on the Fusion platform code generation much. The little AMD SoCs are pretty awesome, and if Sony has a few tricks up their sleeve for making things like OpenCL nicer, we could see some use elsewhere. Dev studios will get the advantage that because the Sony code's not in llvm mainline, it's a lot easier for Sony to give them more recent, faster compilers.
Re:PlayStation 4 System Compiler Support Landing in LLVM by nocut12 on Thursday January 29, 2015 @01:05PM (#8967810)
It could possibly lead to PS4 support from some more (especially open source) game engines. That would be a pretty big deal.
Re:PlayStation 4 System Compiler Support Landing in LLVM by Mikeb85 on Thursday January 29, 2015 @01:49PM (#8968062)
Not really. The only hurdle is the developer agreements. Even game engines that are able to run on consoles (Godot, for instance), generally won't give you the export templates unless you're an official developer.
Re:PlayStation 4 System Compiler Support Landing in LLVM by Narishma on Thursday January 29, 2015 @01:12PM (#8967845)
How would it lead to that?