C++11 Compiler: Closest to the standard and how close?

Question

I'm interested in learning C++ more thoroughly now that C++11 is apparently ratified. What compiler currently implements the closest thing available to full C++11 support? How close is said compiler to full support? Are there still major features missing or just language lawyer minutiae?

Solution 1

There's a support matrix on the Apache wiki.

Solution 2

I think the one Scott Meyers maintains on his homepage is pretty good:

http://www.aristeia.com/C++0x/C++0xFeatureAvailability.htm

Solution 3

There's also GCC C++0x (or C++11) status page : http://gcc.gnu.org/projects/cxx0x.html

Solution 4

The llvm C++ compiler "clang" has partial C++11 support; you can see its current state at http://clang.llvm.org/cxx_status.html.

View more solutions

Comments

[Security Hound]

C++ ox/11 standard I think was just approved in the last 6 months. Its going to be awhile before the compilers have even a fraction of the support required to support the full standard. C++ 03 still isnt fully supported

[R. Martinho Fernandes]

@Ramhound the main obstacle to full C++03 support was export. export was removed though. Compilers have been implementing C++11 features for a long time now, and those with the most impact are already in many present day compilers.

[Bo Persson]

It's worse than deprecated, it was just removed from the standard!

[dsimcha]

Thanks, this was exactly the kind of thing I was looking for. I guess the bottom line is that GCC is by far the most advanced and does have support for most of the more interesting new features (though I have no idea how well-debugged said support is).

[David Hammen]

... or even how well-debugged the standard is. It took quite some time to find the bugs in the previous revisions of the standard. And yes, the new standard almost certainly is buggy. Every language standard ever introduced has had its share of bugs. The only solid way to test big chunks of a new standard is a desk check with the brains of the people on the standards committee acting as a stand-in for a working compiler. Everyone by now has heard multiple jokes regarding what really comprises a committee.

[Gene Bushuyev]

The support matrix doesn't tell you how buggy the implementations are though, and unfortunately all compilers I tested had problems in areas they claimed to support.

[MSalters]

@David: Not entirely true; quite a few bits of the new standard are just relabeled Boost components. We know from Boost that these parts can be implemented, i.e. the standard itself isn't buggy there.

[David Hammen]

@MSalters: I agree that Boost gives the new revision of the standard a big boost regarding bugs. OTOH, consider the STL. Some of the changes in the C++11 are bug fixes that arose from porting the STL to the original standard and were left intact in the 2003 supplement.

[MSalters]

@David: AFAICT those changes were not bugfixes, else they'd been in 2003, but improvements. You usually don't make improvements in a Corrigendum. The std::vector contiguity was an edge case; implementors thought they'd already had said that in '98.

[David Hammen]

@MSalter: I'm thinking of the myriad of items that compiler vendors had to patch over in order to even make a product. For example, the items that GCC labels with a _GLIBCXX_RESOLVE_LIB_DEFECTS tag.

[Potatoswatter]

Nice! Hint to others: there are tabs at the bottom of the page.

[Trass3r]

Its C++11 support is almost complete now.

[maxpolk]

It's complete now, "Clang 3.3 and later implement all of the ISO C++ 2011 standard."