[flang] Unnecessary call to CHECK() for call to NULL() as an actual argument

[PeteSteinfeld]

Summary

A program that passes NULL() to the intrinsic ASSOCIATED() should return an
error since section 16.9.16 requires that the pointer association status "shall
not be undefined".

In processing NULL() as an actual argument, the compiler ends up with an
'std::nullopt', which caused a call to 'CHECK()' to fail. But the value of
'std::nullopt' seems correct to me.

I fixed this by just removing the call to 'CHECK()' and adding a test.

[klausler]

NULL() isn't "undefined". It's a pointer (or in some other contexts, an unallocated allocatable), and the description of ASSOCIATED() in 16.9.16 just requires that its first (POINTER=) argument be a pointer.

(Even if NULL() were undefined, it would be okay -- see paragraph (2) of 16.1, which describes general rules for inquiry functions, of which ASSOCIATED is an example.) "Unless the description of an intrinsic inquiry function states otherwise, these arguments are permitted to be unallocated allocatable variables or pointers that are undefined or disassociated."

[PeteSteinfeld]

In D87859#2280407, @klausler wrote:

NULL() isn't "undefined". It's a pointer (or in some other contexts, an unallocated allocatable), and the description of ASSOCIATED() in 16.9.16 just requires that its first (POINTER=) argument be a pointer.

(Even if NULL() were undefined, it would be okay -- see paragraph (2) of 16.1, which describes general rules for inquiry functions, of which ASSOCIATED is an example.) "Unless the description of an intrinsic inquiry function states otherwise, these arguments are permitted to be unallocated allocatable variables or pointers that are undefined or disassociated."

This all sounds reasonable to me. If what you say is all correct, I would expect the call to ALLOCATED() to return .FALSE.. But every compiler I've tried (IBM, Intel, PGI, and GNU) all produce compile-time errors.

@klausler, is it your guidance that we should compiler the expression ALLOCATED(NULL)) without error?

[klausler]

In D87859#2280893, @PeteSteinfeld wrote:

In D87859#2280407, @klausler wrote:

NULL() isn't "undefined". It's a pointer (or in some other contexts, an unallocated allocatable), and the description of ASSOCIATED() in 16.9.16 just requires that its first (POINTER=) argument be a pointer.

(Even if NULL() were undefined, it would be okay -- see paragraph (2) of 16.1, which describes general rules for inquiry functions, of which ASSOCIATED is an example.) "Unless the description of an intrinsic inquiry function states otherwise, these arguments are permitted to be unallocated allocatable variables or pointers that are undefined or disassociated."

This all sounds reasonable to me. If what you say is all correct, I would expect the call to ALLOCATED() to return .FALSE.. But every compiler I've tried (IBM, Intel, PGI, and GNU) all produce compile-time errors.

@klausler, is it your guidance that we should compiler the expression ALLOCATED(NULL)) without error?

nagfor gets it right; we should do so too.

[PeteSteinfeld]

I've lost the repository where I was working on this, so I'm abandoning it.

flang/lib/Evaluate/intrinsics.cpp

std::optional<SpecificCall> IntrinsicInterface::Match(
   characteristics::DummyArguments dummyArgs;
   std::optional<int> sameDummyArg;
   for (std::size_t j{0}; j < dummies; ++j) {
     const IntrinsicDummyArgument &d{dummy[std::min(j, dummyArgPatterns - 1)]};
     if (const auto &arg{rearranged[j]}) {
       if (const Expr<SomeType> *expr{arg->UnwrapExpr()}) {
         auto dc{characteristics::DummyArgument::FromActual(
             std::string{d.keyword}, *expr, context)};
-        CHECK(dc);
         dummyArgs.emplace_back(std::move(*dc));
         if (d.typePattern.kindCode == KindCode::same && !sameDummyArg) {
           sameDummyArg = j;
         }
       } else {
         CHECK(arg->GetAssumedTypeDummy());
         dummyArgs.emplace_back(std::string{d.keyword},
             characteristics::DummyDataObject{DynamicType::AssumedType()});

flang/test/Semantics/call07.f90

contains
   end subroutine
 
   subroutine test
     !ERROR: CONTIGUOUS POINTER must be an array
     real, pointer, contiguous :: a01 ! C830
     real, pointer :: a02(:)
     real, target :: a03(10)
     real :: a04(10) ! not TARGET
+    !ERROR: Actual argument 'NULL()' associated with dummy argument is not a variable or typed expression
+    print *, associated(null())
     call s01(a03) ! ok
     !ERROR: Actual argument associated with CONTIGUOUS POINTER dummy argument 'p=' must be simply contiguous
     call s01(a02)
     !ERROR: Actual argument associated with CONTIGUOUS POINTER dummy argument 'p=' must be simply contiguous
     call s01(a03(::2))
     call s02(a02) ! ok
     call s03(a03) ! ok
     !ERROR: Actual argument associated with POINTER dummy argument 'p=' must also be POINTER unless INTENT(IN)