Anastasia marked an inline comment as done.
Anastasia added a comment.
missing test case
1 auto* accumulator()
2 {
3 __global int * glptr;
4 return glptr;
5 }
6 int i;
7 auto* ai = &i;
8 int* ii = &i;
In D65744#1646348 <https://reviews.llvm.org/D65744#1646348>, @rjmccall wrote:
> In D65744#1629055 <https://reviews.llvm.org/D65744#1629055>, @Anastasia wrote:
>
> > In D65744#1627589 <https://reviews.llvm.org/D65744#1627589>, @rjmccall
> > wrote:
> >
> > > I see. Is the deduction rule recursive or something, where a pointer to
> > > pointer is inferred to point to the same address space as the pointee?
> >
> >
> > It is recursive indeed and we currently deduce all pointees to generic AS.
>
>
> Is that likely to change? If all pointees are inferred to be generic (in the
> absence of an explicit qualifier) when building a pointer type, this seems
> very over-thought vs. just adding the generic AS whenever you build a pointer
> to an unqualified but non-dependent type.
I don't think this is likely to change. Are you suggesting to move the
deduction logic for pointee of pointers, references and block pointers into
ASTContext helper that creates a pointer/reference/block pointer type? I guess
it should work if the only way to construct the pointer is by using this
helper. I can certainly prototype this change and see.
================
Comment at: lib/Sema/SemaType.cpp:7441
+ // the initializing expression type during the type deduction.
+ (T->isAutoType() && IsPointee) || (IsAutoPointee) ||
// OpenCL spec v2.0 s6.9.b:
----------------
mantognini wrote:
> mantognini wrote:
> > Shouldn't the parentheses around `IsAutoPointee` be removed for style
> > consistency?
> With the `if` statement introduced above, `IsAutoPointee` can be true only in
> C++ mode. Could it be an issue to not guard `(T->isAutoType() && IsPointee)`
> for non-C++ mode? (I guess not, but I couldn't convince myself.)
I think `TreeTransforms` will only be used in C++ mode. But also `isAutoType`
should only be true in C++ mode. So I think we should be fine.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D65744/new/
https://reviews.llvm.org/D65744
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
