[cfe-dev] improved vector bool/pixel support

[Anton Yartsev]

On 05.08.2010 15:12, Douglas Gregor wrote:
> On Aug 5, 2010, at 1:52 AM, Anton Yartsev wrote:
>    
>> Added code for an implicit conversion between GCC and AltiVec vectors to Sema. Also attached tests with the relaxed vector conversions turned off. Please review.
>>      
> Index: lib/Sema/SemaExpr.cpp
> ===================================================================
> --- lib/Sema/SemaExpr.cpp	(revision 110284)
> +++ lib/Sema/SemaExpr.cpp	(working copy)
> @@ -4751,13 +4751,18 @@
>     }
>
>     if (lhsType->isVectorType() || rhsType->isVectorType()) {
> -    // If we are allowing lax vector conversions, and LHS and RHS are both
> -    // vectors, the total size only needs to be the same. This is a bitcast;
> -    // no bits are changed but the result type is different.
> -    if (getLangOptions().LaxVectorConversions&&
> -        lhsType->isVectorType()&&  rhsType->isVectorType()) {
> -      if (Context.getTypeSize(lhsType) == Context.getTypeSize(rhsType))
> +    if (lhsType->isVectorType()&&  rhsType->isVectorType()) {
> +      // If we are allowing lax vector conversions, and LHS and RHS are both
> +      // vectors, the total size only needs to be the same. This is a bitcast;
> +      // no bits are changed but the result type is different.
> +      if (getLangOptions().LaxVectorConversions&&
> +         (Context.getTypeSize(lhsType) == Context.getTypeSize(rhsType)))
>           return IncompatibleVectors;
> +
> +      // Allow assignments of an AltiVec vector type to an equivalent GCC
> +      // vector type and vice versa
> +      if (Context.areCompatibleVectorTypes(lhsType, rhsType))
> +        return IncompatibleVectors;
>       }
>
> I don't think that you want a warning when converting between AltiVec and GCC vector types, do you? If no, the "return IncompatibleVectors;" should be "return Compatible;".
>
> Why does the C++ test use __attribute__((overloadable))? Actually, why does the C test use it? It doesn't seem necessary in either case.
>
> It might make sense to test the ? : operator, which often gives us trouble with conversions.
>
> 	- Doug
Changed "return IncompatibleVectors" to "return Compatible".

Hmm.. The attribute is absolutely useless in both cases..
I used it in the C test to reach the IsVectorConversion() finction in 
SemaOverloaded.cpp, but there should be at least two overloaded 
candidates to reach the desired code.

Added "?:" operator to tests, no troubles.

-- 
Anton

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: Altivec_GCC_compat.patch
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20100805/38d72c34/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: test.c
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20100805/38d72c34/attachment.c>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: test.cpp
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20100805/38d72c34/attachment-0001.ksh>