[flang-commits] [flang] [flang] Extension intrinsics INT8 and INT2 (PR #109433)

[via flang-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-flang-semantics

Author: Peter Klausler (klausler)

<details>
<summary>Changes</summary>

These are legacy conversion intrinsic functions supported by nearly all Fortran compilers (esp. INT8).
They are equivalent to INT(..., KIND=8 or 2), respectively.

---
Full diff: https://github.com/llvm/llvm-project/pull/109433.diff


2 Files Affected:

- (modified) flang/lib/Evaluate/intrinsics.cpp (+6) 
- (added) flang/test/Evaluate/int8.f90 (+5) 


``````````diff
diff --git a/flang/lib/Evaluate/intrinsics.cpp b/flang/lib/Evaluate/intrinsics.cpp
index 166dae93178c51..26661d9d534b3d 100644
--- a/flang/lib/Evaluate/intrinsics.cpp
+++ b/flang/lib/Evaluate/intrinsics.cpp
@@ -1169,6 +1169,12 @@ static const SpecificIntrinsicInterface specificIntrinsicFunction[]{
     // procedure pointer target.
     {{"index", {{"string", DefaultChar}, {"substring", DefaultChar}},
         DefaultInt}},
+    {{"int2", {{"a", AnyNumeric, Rank::elementalOrBOZ}},
+         TypePattern{IntType, KindCode::exactKind, 2}},
+        "int"},
+    {{"int8", {{"a", AnyNumeric, Rank::elementalOrBOZ}},
+         TypePattern{IntType, KindCode::exactKind, 8}},
+        "int"},
     {{"isign", {{"a", DefaultInt}, {"b", DefaultInt}}, DefaultInt}, "sign"},
     {{"jiabs", {{"a", TypePattern{IntType, KindCode::exactKind, 4}}},
          TypePattern{IntType, KindCode::exactKind, 4}},
diff --git a/flang/test/Evaluate/int8.f90 b/flang/test/Evaluate/int8.f90
new file mode 100644
index 00000000000000..ef8321f4c710b3
--- /dev/null
+++ b/flang/test/Evaluate/int8.f90
@@ -0,0 +1,5 @@
+! RUN: %flang_fc1 -fdebug-unparse %s 2>&1 | FileCheck %s
+!CHECK: warning: REAL(4) to INTEGER(2) conversion overflowed
+!CHECK: PRINT *, 32767_2, 4000000000_8
+print *, int2(4.e9), int8(4.e9)
+end

``````````

</details>


https://github.com/llvm/llvm-project/pull/109433