[cfe-commits] r55417 - in /cfe/trunk: include/clang/AST/ASTContext.h include/clang/AST/Type.h include/clang/Basic/DiagnosticKinds.def include/clang/Basic/LangOptions.h include/clang/Parse/DeclSpec.h lib/AST/ASTContext.cpp lib/AST/Type.cpp lib/AST

steve naroff snaroff at apple.com
Wed Aug 27 13:58:28 PDT 2008

On Aug 27, 2008, at 4:38 PM, Chris Lattner wrote:

> On Aug 27, 2008, at 1:02 PM, Eli Friedman wrote:
>> On Wed, Aug 27, 2008 at 10:02 AM, Ted Kremenek <kremenek at apple.com>
>> wrote:
>>> There is no public documentation yet (Apple-related documents/talks
>>> are currently under NDA), although we plan on putting some
>>> information
>>> in the Clang internals documents.
>> I don't think it's appropriate to be putting new features into clang
>> without some reasonable documentation publicly available.  Ideally,
>> there should be a specification, but even if that isn't possible,
>> there should at least be some documentation available for how it's
>> supposed to work, especially if it's an extension to C.  Apple is
>> definitely the primary contributor to clang, but it isn't an
>> Apple-internal project.  Consider it like this: what would the
>> reaction be if someone else checked in an unused extension to clang's
>> type system without review, and claimed that the documentation is all
>> under Company X's NDA.
> To be clear, getting the documentation in shape is part of the
> process.  This is one of Steve's short-term todo items.  If you
> prefer, he can send out an old and very out-of-date version of the
> document, but that might cause more confusion than it will clear up.

I think it's important to note that *many* language features (ObjC,  
many GCC extensions) don't have specifications (yet we've made forward  
progress, using GCC as our reference implementation).

Just to be clear, I'm not trying to condone the practice (of  
implementing without a formal spec), but it isn't a new phenomenon.  
There is no doubt that having formal specs and documentation is  
goodness. I will modernize the documentation for "blocks" before  
committing it to the repository. Creating specifications for all of  
ObjC is another story:-)


> -Chris
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

More information about the cfe-commits mailing list