[cfe-commits] r147730 - in /cfe/trunk: include/clang/Sema/Sema.h lib/CodeGen/CodeGenFunction.cpp lib/Parse/ParseDecl.cpp lib/Sema/SemaDecl.cpp test/CodeGen/vla.c

[Enea Zaffanella]

Il 09/01/2012 18:42, John McCall ha scritto:
> 
> On Jan 7, 2012, at 2:52 AM, Abramo Bagnara wrote:
> 
>> Author: abramo
>> Date: Sat Jan  7 04:52:36 2012
>> New Revision: 147730
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=147730&view=rev
>> Log:
>> Fixed TypeofExpr AST and code generation.
>>
>> Modified:
>>    cfe/trunk/include/clang/Sema/Sema.h
>>    cfe/trunk/lib/CodeGen/CodeGenFunction.cpp
>>    cfe/trunk/lib/Parse/ParseDecl.cpp
>>    cfe/trunk/lib/Sema/SemaDecl.cpp
>>    cfe/trunk/test/CodeGen/vla.c
>>
>> Modified: cfe/trunk/include/clang/Sema/Sema.h
>> URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Sema/Sema.h?rev=147730&r1=147729&r2=147730&view=diff
>> ==============================================================================
>> --- cfe/trunk/include/clang/Sema/Sema.h (original)
>> +++ cfe/trunk/include/clang/Sema/Sema.h Sat Jan  7 04:52:36 2012
>> @@ -1146,8 +1146,6 @@
>>                                    DeclSpec &DS,
>>                                    MultiTemplateParamsArg TemplateParams);
>>
>> -  StmtResult ActOnVlaStmt(const DeclSpec &DS);
>> -
>>   Decl *BuildAnonymousStructOrUnion(Scope *S, DeclSpec &DS,
>>                                     AccessSpecifier AS,
>>                                     RecordDecl *Record);
>>
>> Modified: cfe/trunk/lib/CodeGen/CodeGenFunction.cpp
>> URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CodeGenFunction.cpp?rev=147730&r1=147729&r2=147730&view=diff
>> ==============================================================================
>> --- cfe/trunk/lib/CodeGen/CodeGenFunction.cpp (original)
>> +++ cfe/trunk/lib/CodeGen/CodeGenFunction.cpp Sat Jan  7 04:52:36 2012
>> @@ -929,19 +929,33 @@
>>
>>   // We're going to walk down into the type and look for VLA
>>   // expressions.
>> -  type = type.getCanonicalType();
>>   do {
>>     assert(type->isVariablyModifiedType());
>>
>>     const Type *ty = type.getTypePtr();
>>     switch (ty->getTypeClass()) {
>> +
>> +    default:
>> +      // Only sugared types (different from typeof_expr) can reach this point.
>> +      assert(!type.isCanonical() && "unhandled canonical type!");
>> +      type = type.getSingleStepDesugaredType(getContext());
>> +      break;
> 
> This would look through typedefs and therefore be very wrong if it could
> actually execute.  Fortunately, it can't, except with decltype.  Still, please
> don't leave this code in there.

You are right, here is a patch for review.

For the non-canonical types, we now stop (doing nothing) on Typedef,
Decltype and Auto; we keep going (after single-level desugaring) on
Paren, TypeOf, UnaryTransform, Attributed and SubstTemplateTypeParm.
(The default case has gone.)

Enea.


> Also, rather than re-introducing a default case, please just nullify and
> enumerate the metaprogramming case you care about
> (NON_CANONICAL_UNLESS_DEPENDENT_TYPE).  Alternatively,
> change the metaprogramming file so that it provides opt-out filters for
> specific cases the same way that StmtNodes.inc does.
> 
> John.
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: vla-2.patch
Type: text/x-patch
Size: 1926 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20120109/28a2ddfd/attachment.bin>