[cfe-commits] r120890 - in /cfe/trunk: include/clang/AST/ include/clang/Sema/ lib/AST/ lib/Analysis/ lib/Checker/ lib/CodeGen/ lib/Parse/ lib/Sema/ test/CodeGen/ test/CodeGenObjCXX/ test/Rewriter/ test/Sema/ test/SemaCXX/ test/SemaTemplate/

[Abramo Bagnara]

Il 04/12/2010 07:42, John McCall ha scritto:
> 
> On Dec 3, 2010, at 10:37 PM, Abramo Bagnara wrote:
> 
>> Il 04/12/2010 04:47, John McCall ha scritto:
>>> --- cfe/trunk/lib/AST/Expr.cpp (original)
>>> +++ cfe/trunk/lib/AST/Expr.cpp Fri Dec  3 21:47:34 2010
>>> @@ -824,6 +824,8 @@
>>>     return "LValueBitCast";
>>>   case CK_LValueToRValue:
>>>     return "LValueToRValue";
>>> +  case CK_GetObjCProperty:
>>> +    return "GetObjCProperty";
>>>   case CK_NoOp:
>>>     return "NoOp";
>>>   case CK_BaseToDerived:
>>> @@ -1722,6 +1724,24 @@
>>>   }
>>> }
>>>
>>> +Expr *Expr::IgnoreParenLValueCasts() {
>>> +  Expr *E = this;
>>> +  while (E) {
>>> +    if (ParenExpr *P = dyn_cast<ParenExpr>(E)) {
>>> +      E = P->getSubExpr();
>>> +      continue;
>>> +    }
>>> +    if (CastExpr *P = dyn_cast<CastExpr>(E)) {
>>> +      if (P->getCastKind() == CK_LValueToRValue) {
>>> +        E = P->getSubExpr();
>>> +        continue;
>>> +      }
>>> +    }
>>> +    break;
>>> +  }
>>> +  return E;
>>> +}
>>
>> What about __extension__ skipping?
> 
> Good point.  I *think* we don't need that here because this method is meant to just be a hack to simplify workarounds in the static analyzer, but Ted might have thoughts.  For one, I don't know whether the analyzer generally pays attention to __extension__ nodes.

Whenever it calls IgnoreParen* helpers it *does* pay attention: they are
always skipped.