Commit Graph

792 Commits

Author SHA1 Message Date
Steve Naroff af2a022563 - Add Sema::CheckStringLiteralInit, Sema::IsStringLiteralInit.
- Use previous hooks to simplify Sema::CheckInitializerTypes()...

llvm-svn: 46235
2008-01-22 00:55:40 +00:00
Steve Naroff f9eb598de3 Sema::CheckInitializerTypes(). Start simpliying and cleaning up...
llvm-svn: 46234
2008-01-21 23:53:58 +00:00
Fariborz Jahanian 7718e36e25 Elimate bogus warning when va_start is correctly used in
a method.

llvm-svn: 46232
2008-01-21 22:59:53 +00:00
Steve Naroff acb6fa6bea Sema::FinalizeDeclaratorGroup(): Tighten up the tentative definition rule when dealing with arrays.
Bug submitted by Eli.

llvm-svn: 46179
2008-01-18 20:40:52 +00:00
Steve Naroff b716fbab44 Sema::FinalizeDeclaratorGroup()...make sure we emit an diagnostic for tentative definitions with incomplete types. Touch up all test cases that are effected.
llvm-svn: 46152
2008-01-18 00:39:39 +00:00
Fariborz Jahanian 9a5124a67e Don't ICE on missing interface declaration when declaring one of its protocols.
llvm-svn: 46141
2008-01-17 20:33:24 +00:00
Ted Kremenek 32a3358eab Fixed typo introduced by r46133 that caused -Wfloat-equal to break.
llvm-svn: 46134
2008-01-17 17:55:13 +00:00
Nate Begeman 1e36a85f64 Implement basic overload support via a new builtin, __builtin_overload.
__builtin_overload takes 2 or more arguments:
0) a non-zero constant-expr for the number of arguments the overloaded 
   functions will take
1) the arguments to pass to the matching overloaded function
2) a list of functions to match.

The return type of __builtin_overload is inferred from the function whose args
match the types of the arguments passed to the builtin.  For example:

float a;
float sinf(float);
int   sini(int);

float b = __builtin_overload(1, a, sini, sinf);

Says that we are overloading functions that take one argument, and trying to 
pass an argument of the same type as 'a'.  sini() does not match since it takes
and argument of type int.  sinf does match, so at codegen time this will turn
into float b = sinf(a);

llvm-svn: 46132
2008-01-17 17:46:27 +00:00
Fariborz Jahanian ad134b987e Renamed getSynthesizedSelectorSize to getSynthesizedMethodSize. Compute method name
size in accordance with how rewriter synthesizes method name.

llvm-svn: 46131
2008-01-17 17:37:26 +00:00
Ted Kremenek fff70962bb Added method Expr::IgnoreParens(), which returns the first non-ParenExpr Expr*.
Refactored the use of this method into both the Sema module and Analysis module,
which were using their own static functions that did the same thing.

llvm-svn: 46129
2008-01-17 16:57:34 +00:00
Fariborz Jahanian 386699515b Computed length of a __func__ identifier used in an objective-c method correctly,
when creating its type.

llvm-svn: 46109
2008-01-17 01:36:09 +00:00
Chris Lattner a65e1f3b31 Move promoteExprToType from being a static method in SemaExpr.cpp to being
a method named ImpCastExprToType in Sema.

Use this method to insert implicit casts for case statements from their 
operand type to the condition type of the switch.  This fixes a crash on
test/CodeGen/statements.c, reported by Eli Friedman.

llvm-svn: 46083
2008-01-16 19:17:22 +00:00
Steve Naroff 17832a48f8 Sema::MergeFunctionDecl()...make sure diagnostic is accurate (wrt function declarations/definitions). Patch by Carl Lewis!
llvm-svn: 46070
2008-01-16 15:01:34 +00:00
Steve Naroff abefc391dd Finish up handling all permutations of "complex int" (in Sema::UsualArithmeticConversions()).
A FIXME remains to verify the conversion rules are consistent with GCC.

Thanks to Eli for the new/improved test case...

llvm-svn: 46022
2008-01-15 22:21:49 +00:00
Steve Naroff 6fcfd0581d - Change Type::isComplexType() to exlude GCC's complex integer extension. In general, we will keep the lowest level Type predicates "pure" (i.e. true to the C99 spec).
- Modify Sema::UsualArithmeticConversions() to work with the new definition of Type::isComplexType().

This is a nice cleanup and also fixes a bug submitted by Eli (which I've added to the test suite).

llvm-svn: 46005
2008-01-15 19:36:10 +00:00
Steve Naroff 7f988889ef Rework commit r45976, which was incorrect.
- Add Type::isComplexIntegerType(), Type::getAsComplexIntegerType().
- Don't inlude complex types with Type::isIntegerType(), which is too general.
- Use the new predicates in Sema::UsualArithmeticConversions() to recognize/convert the types.

llvm-svn: 45992
2008-01-15 01:41:59 +00:00
Steve Naroff d8c7d484af Make sure Sema::GetTypeForDeclarator() deals with abstract declarators when diagnosing incorrect code.
Bug submitted by Eli.

llvm-svn: 45989
2008-01-14 23:33:18 +00:00
Chris Lattner 49f980c334 Fix build issue on cygwin, patch by Sam Bishop
llvm-svn: 45977
2008-01-14 21:47:29 +00:00
Steve Naroff 012484d6c8 Fix crasher when redefining functions. Not 100% pleased with this solution, but it is clearly an improvement. Will discuss with Chris later.
llvm-svn: 45975
2008-01-14 20:51:29 +00:00
Steve Naroff d32419de0d Record if a compound literal expression is @ file scope. This allows us to implement C99 6.5.2.5p6. This could have been done without modifying the AST (by checking the decl type and passing the info down to isContextExpr), however we concluded this is more desirable.
Bug/patch by Eli Friedman!

llvm-svn: 45966
2008-01-14 18:19:28 +00:00
Steve Naroff 826e91ae04 Change Sema::CheckAddressOfOperation() to respect C99-only addressof rules.
Remove diagnostics from Sema::CheckIndirectionOperand(). C89/C99 allow dereferencing an incomplete type. clang appears to be emulating some incorrect gcc behavior (see below).

void
foo (void)
{
 struct b;
 struct b* x = 0;
 struct b* y = &*x; // gcc produces an error ("dereferencing pointer to incomplete type")
}

With this patch, the above is now allowed.

Bug/Patch by Eli Friedman!

llvm-svn: 45933
2008-01-13 17:10:08 +00:00
Chris Lattner 0a8c2325fc allow __func__ in objc methods.
llvm-svn: 45915
2008-01-12 19:32:28 +00:00
Chris Lattner 317e6ba07f Tighten up handling of __func__ and friends: it should be an array
of const char, and it should error if it occurs outside a function.

Is it valid in an objc method?  If so we should handle that too.

llvm-svn: 45910
2008-01-12 18:39:25 +00:00
Chris Lattner a81a0279cc Fix the type of predefined identifiers like __func__. Patch by
Eli Friedman!

llvm-svn: 45906
2008-01-12 08:14:25 +00:00
Chris Lattner 38376f1595 Add first pieces of support for parsing and representing
extern "C" in C++ mode.  Patch by Mike Stump!

llvm-svn: 45904
2008-01-12 07:05:38 +00:00
Steve Naroff 98f7203680 - Teach Expr::isConstantExpr() about InitListExpr's (and offsetof, since I noticed it was missing).
- Rename CheckInitializer() to CheckInitializerTypes().
- Removed the isStatic argument to CheckInitializerTypes() and all of it's subroutines. Checking for constant expressions is now done separately.
- Added CheckForConstantInitializer().

llvm-svn: 45840
2008-01-10 22:15:12 +00:00
Fariborz Jahanian 939776756f Recover from user typo not having proper @interface decl and a bad foreach decl.
llvm-svn: 45839
2008-01-10 20:33:58 +00:00
Fariborz Jahanian 82ae0152a7 Allow messaging expression as foreach's collection expression.
llvm-svn: 45793
2008-01-10 00:24:29 +00:00
Steve Naroff e6b0ec8b5e Fix Sema::ActOnDeclarator() to call MergeFunctionDecl for function decls that aren't in scope. Since C functions are in a flat namespace, we need to give them special treatment (when compared with variables and typedefs).
llvm-svn: 45789
2008-01-09 23:34:55 +00:00
Steve Naroff c6edcbdb5d Fix ASTContext::typesAreCompatible to allow for int/enum compatibility (C99 6.7.2.2p4).
Fix Sema::MergeFunctionDecl to allow for function type compatibility (by using the predicate on ASTContext). Function types don't have to be identical to be compatible...

llvm-svn: 45784
2008-01-09 22:43:08 +00:00
Steve Naroff 08ddb8c55c Teach Sema::ActOnCompoundLiteral about constraint C99 6.5.2.5p3.
llvm-svn: 45782
2008-01-09 20:58:06 +00:00
Steve Naroff 039ad3cf90 Fix Sema::CheckConditionalOperands(). The null pointer constant checks need to precede the check for two pointer operands.
llvm-svn: 45732
2008-01-08 01:11:38 +00:00
Ted Kremenek 1b0ea82459 Substituted all instances of the string "Objc" for "ObjC". This fixes
some naming inconsistencies in the names of classes pertaining to Objective-C
support in clang.

llvm-svn: 45715
2008-01-07 19:49:32 +00:00
Fariborz Jahanian d5450b7f07 Limit type of foreach's element and collection to be a pointer to
objc object type.

llvm-svn: 45709
2008-01-07 18:14:04 +00:00
Fariborz Jahanian fef287cb67 Issue diagnostics if more than one declaration in objectove-c's foreach-stmt header.
llvm-svn: 45708
2008-01-07 17:52:35 +00:00
Chris Lattner f5c973d061 minor cleanup.
llvm-svn: 45706
2008-01-07 17:51:46 +00:00
Chris Lattner 4d3b0f579c rearrange some code.
llvm-svn: 45666
2008-01-06 22:50:31 +00:00
Chris Lattner e2949f42d3 some simplifications/cleanups to ?: sema.
llvm-svn: 45665
2008-01-06 22:42:25 +00:00
Chris Lattner cfb6f430b6 Emit warnings like "dereferencing void pointer" instead of trying to
pretty print the type name for void.

llvm-svn: 45664
2008-01-06 22:21:46 +00:00
Fariborz Jahanian 7b1b947ca1 twiks based on Chris's comment. No need to iterate thru a list of declarations
in a foreach-statement which is illegal (but not yet caught) .

llvm-svn: 45615
2008-01-05 01:40:08 +00:00
Fariborz Jahanian bad3218954 Minor refactoring of foreach's semantics code per Chris's suggetion.
llvm-svn: 45604
2008-01-04 23:59:09 +00:00
Chris Lattner 881a212a70 unnest vector handling again.
llvm-svn: 45602
2008-01-04 23:32:24 +00:00
Chris Lattner f735409e11 we already test for exact type matches early, so we don't have to do
it explicitly for vectors.  This allows us to unnest some code.

llvm-svn: 45600
2008-01-04 23:20:56 +00:00
Chris Lattner a52c2f2e3e simplify some of this code by removing the else/if chain and by
stripping down to canonical types early.

llvm-svn: 45599
2008-01-04 23:18:45 +00:00
Chris Lattner a3fc41d616 move objc expr sema to its own file.
llvm-svn: 45597
2008-01-04 22:32:30 +00:00
Chris Lattner 940cfebf90 add comments for the various AssignConvertType's, and split int->pointer from pointer->int.
llvm-svn: 45591
2008-01-04 18:22:42 +00:00
Chris Lattner 9bad62c72a Merge all the 'assignment' diagnostic code into one routine, decloning
it from several places.  This merges the diagnostics, making them more
uniform and fewer in number. This also simplifies and cleans up the code.

Some highlights:
1. This removes a bunch of very-similar diagnostics.
2. This renames AssignmentCheckResult -> AssignConvertType
3. This merges PointerFromInt + IntFromPointer which were always treated the same.
4. This updates a bunch of test cases that have minor changes to the produced diagnostics.

llvm-svn: 45589
2008-01-04 18:04:52 +00:00
Fariborz Jahanian 775d5d02af Patch to add semantics check for ObjC2's foreacn statement.
llvm-svn: 45561
2008-01-04 00:27:46 +00:00
Chris Lattner 816dea2fc5 generalize some of the conversion warnings.
llvm-svn: 45560
2008-01-03 23:38:43 +00:00
Chris Lattner 874e024f10 regroup assignment-related diagnostics, fix a mistake in one.
llvm-svn: 45558
2008-01-03 23:13:36 +00:00