Add SelectorInfo (similar in spirit to IdentifierInfo). The key difference is SelectorInfo is not string-oriented, it is a unique aggregate of IdentifierInfo's (using a folding set).  SelectorInfo also has a richer API that simplifies the parser/action interface. 3 noteworthy benefits:

#1: It is cleaner. I never "liked" storing keyword selectors (i.e. foo:bar:baz) in the IdentifierTable.

#2: It is more space efficient. Since Cocoa keyword selectors can be quite long, this technique is space saving. For Cocoa.h, pulling the keyword selectors out saves ~180k. The cost of the SelectorInfo data is ~100k. Saves ~80k, or 43%.

#3: It results in many API simplifications. Here are some highlights:

- Removed 3 actions (ActOnKeywordMessage, ActOnUnaryMessage, & one flavor of ObjcBuildMethodDeclaration that was specific to unary messages).
- Removed 3 funky structs from DeclSpec.h (ObjcKeywordMessage, ObjcKeywordDecl, and ObjcKeywordInfo).
- Removed 2 ivars and 2 constructors from ObjCMessageExpr (fyi, this space savings has not been measured).

I am happy with the way it turned out (though it took a bit more hacking than I expected). Given the central role of selectors in ObjC, making sure this is "right" will pay dividends later.

Thanks to Chris for talking this through with me and suggesting this approach. 

llvm-svn: 42395

llvm-svn: 42014

Switch -parse-ast over to it.

llvm-svn: 41991

llvm-svn: 41611

out of the llvm namespace.  This makes the clang namespace be a sibling of
llvm instead of being a child.

The good thing about this is that it makes many things unambiguous.  The
bad things is that many things in the llvm namespace (notably data structures
like smallvector) now require an llvm:: qualifier.  IMO, libsystem and libsupport
should be split out of llvm into their own namespace in the future, which will fix
this issue.

llvm-svn: 39659

llvm-svn: 39478

Go back to having the clang driver create ASTContext explicitly, passing
it to Sema/ASTStreamer (separating the lifetime of ASTContext from
the lifetime of Sema). One day it might be useful to consider creating
a context object implicitly if one isn't provided (using default arguments in
Sema's constructor). At this point, adding this convenience isn't necessary.

llvm-svn: 39346

the prerocessor will be available via ASTContext).
- Removed the public data member "PP" in ASTContext.
- Changed ASTContext's contructor to take TargetInfo/IdentifierTable explicitly.
- Implicitly create an ASTContext in Sema's constructor. This simplifies
the clients job (and makes ASTContext more private).
--As a side effect, added a "PrintStats" hook to Sema.

To support this level of encapsulation, ASTContext is always dynamically
allocated (by Sema). Previously, ASTContext was being allocated on the
stack. I don't believe this should be a performance issue (since ASTContext
is fairly course grain and tied to the lifetime of Sema currently).

llvm-svn: 39345

*** AST Context Stats:
  30594 types total.
    19 builtin types
    3929 pointer types
    308 array types
    18883 function types with proto
    8 function types with no proto
    2988 typename (typedef) types
    4459 tagged types
      1476 struct types
      80 union types
      0 class types
      2903 enum types
  83298 slow type lookups

Next up, making type canonicalization not trivially silly.

llvm-svn: 39305

llvm-svn: 39161

interface exported by libast.

llvm-svn: 39154

llvm-svn: 39153

and should be implemented with an ASTBuilder subclass anyway.

llvm-svn: 39107

llvm-svn: 39095

declarations through the asm streamer.  For a testcase like:

int G;
int H, I, *J;
int func() {}

'clang -parse-print-ast' prints:

Read top-level decl: G
Read top-level decl: H
Read top-level decl: I
Read top-level decl: J
Read top-level decl: func

llvm-svn: 38992

llvm-svn: 38960

Add callbacks for same.
Add "full locinfo" mode.

llvm-svn: 38939

llvm-svn: 38938

llvm-svn: 38936

llvm-svn: 38935