LLVM Clang Compiler Now C++11 Feature Complete

An anonymous reader writes "With the latest development work on Clang ahead of the release of LLVM version 3.3, Clang is now C++11 feature complete. The last remaining features of the ISO C++11 feature specification have been implemented. C++11 support for GCC is also more or less complete."

Re:Thank you, Apple!

[Anonymous Coward]

It has long been accepted fact that GPL does hurt open source in the long run.

Re:Thank you, Apple!

[Microlith]

Because the AC in the initial post in this thread did.

Re:Thank you, Apple!

[PhamNguyen]

I imagine that the issue is not so much that they want to fork LLVM, but they want to integrate LLVM with XCode (I'm guessing they already do, but when I stopped using XCode it was still using GCC) for static analysis. The main difference between LLVM and GCC that allows this is that LLVM is not monolithic, so you can use its code analysis/parsing features without actually compiling, and you can rely on the stability of these components. Licences may or may not be an issue depeding on whether it is necessary to link to LLVM components at compile time or not. But at least with LLVM you have this option (while staying closed source).

Re:Executable performance

[spitzak]

You might want to try the Intel compiler. This was three years ago but it produced obviously faster results than either gcc or msvc.

Re:Nothing could be more wrong

[SuperKendall]

The answer to that is simple.

Yes it is, but it is not the answer you gave. If Apple cared about pushing out BSD and Linux, wouldn't they put more than a token effort into OS X server? Apple doesn't need to push Linux of the desktop, because the desktop has become irrelevant.

No, the simple answer is that even if you could just use all the GCC code you liked any way you wanted, the way GCC is built does not lend itself to being part of other tools. As in, it's not so easy to just parse some arbitrary code and get a sense of where the tokens are at.

In reality the "huge savings" came about in writing a whole new compiler chain from scratch, which was easier overall than trying to get GCC parsing integrated into anything else!

The side benefit going forward is a WAY better compiler design that can integrate new ideas very quickly, and as I said be used as part of many different tools.