ajensen (@.) on 6/24/10 wrote:
---------------------------
>Hack, spaghetti, whatever. It is harder than the alternative to maintain. Much harder. That is the point.

It is also much, much bigger than *any* other compiler and has a much broader scope. How can you compare the maintainability of it with smaller projects?

>I'm not even going to try to understand those politics, way beyond my capabilities.
>But FreeBSD is making the switch quite fast these days.

Well it still has quite a number of problems building with LLVM as you can check on their project pages. And it's x86 only, good luck having a production build on that, let alone on any other architecture. LLVM is a nice project and it does a lot of things well but replacing GCC is not one of its tasks nor its goal.

>Don't see why Linux can't follow.

Because even ignoring the userland, LLVM cannot compile the Linux kernel and I doubt anybody would be crazy enough to even try it.

> Not that it is important to create mono cultures. The main idea is to try to move on from GCC.

Why? For what it does it works very well for me and a few other thousand of developers plus it's free. What's wrong with it? Sure, it's not perfect but it does just fine for most of its users and has no real alternative.