[PATCH] D78424: [flang] Added Semantic Checks for 2 Data constraints and fixed the semantic errors in 3 test cases

[Kiran Chandramohan via Phabricator via llvm-commits]

kiranchandramohan added a comment.

Thanks for addressing the comments.

Also check the following.

  type t1
   integer, allocatable :: arr(:)
  end type
  type(t1) d1
  data d1/t1((/0,1/))/

Besides the constraints for the data statement, there are several other points mentioned in the data statement section. Like point 8 in 8.6.7 (test below) which says that the number of variables and data-statement-constants shall be the same. Are these outside the scope of your work? If so, please create a ticket somewhere so that we can come back to it later.

  program mn
    character(len=3) :: n1, n2
    data n1, n2 / 'f18' /
  end



================
Comment at: flang/lib/Semantics/check-data.cpp:44
+          "Data Object must not be accessed by host association"_err_en_US);
+    } else if (IsUseAssociated(*symbol, scope)) {
+      context_.Say(name.source,
----------------
kiranchandramohan wrote:
> Should this check fire for the data statement in the following program where the type is defined in a module but the variable itself is not?
> ```
> module dt
>   type d
>     real r
>   end type
> end module
> program test
>   use dt
>   type(d) :: d1
>   data d1%r/2.5/
> end program
> ```
I wasn't sure whether using the type from the module makes the variable d1 a use associated variable. What is your conclusion?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D78424/new/

https://reviews.llvm.org/D78424