Changeset fc9d167

Timestamp:

May 1, 2026, 5:06:35 PM (63 minutes ago)

Author:

Michael Brooks <mlbrooks@…>

Branches:

master

Parents:

16d9c3a

Message:

Slightly loosen a test that has failed on 32-bit build lately.

Since 16d9c3a, this test was failing on all builds. The present change in not 32-bit specific; it's expected to fix this particular test, on all builds.

Note there are further 32-bit issues remaining, even after this fix.

Test is: array-collections/c-dependent

Loosening is: Adjust test's scope to avoid exercising "truly incorrect" C code; which may receive compile-time array-bound warnings on sufficiently advanced C compilers. In test code and .expect, switch from passing incorrect bounds to passing correct bounds; the point of the test (newly clarified, always foremost) is that CFA handles C's dependently sized VLA parameter.

Location:

tests/array-collections

Files:

• .expect/c-dependent.txt (modified) (1 diff)
• c-dependent.cfa (modified) (3 diffs)

tests/array-collections/.expect/c-dependent.txt

@@ -1 @@
-bound_ptr_allow 999: 1.1 1.2 1.3
+bound_ptr_allow 42: 1.1 1.2 1.3
 bound_ar_allow 41:
 1.1 1.2 1.3

tests/array-collections/c-dependent.cfa

@@ -1 @@
-void iota1( int n, float * a, float base ) {
+// Purpose:  Demonstrate that aruably-dependent types, like `void (*)( size_t n, float[n] )`,
+// which are valid in C, work also in CFA.  It's a C-compatibility test.
+// So, here we exercise code that a C compiler that omnisciently warns about bounds would accept.
+
+// Note this test was once written with a broader intent, articulated after the fact as either of:
+// - demonstrate how permissive this C-compatibility feature is,
+//   even though gcc continually works to clamp down in these areas
+// - demonstrate that CFA lowering does not get in the way of C's warnings in these areas
+// Current decision is these qualities (while desirable) are too fickle to be pursued practically.
+
+void iota1( size_t n, float * a, float base ) {
     for (i; n) {
         a[i] = base + 0.1f * (float)(i + 1);
@@ -11 +21 @@
     float a[42];
     iota1( 42, a, 1.0 );
-    f__bound_ptr_allow( 999, a );
+    f__bound_ptr_allow( 42, a );  // pass actual size (42)  ==>  bounds ok
 }
 
+// note dimension `n + 1`, exercising nontrivial dim-expression
 void f__bound_ar_allow( size_t n, float a[][n + 1] ) {
     printf( "bound_ar_allow %zd:\n", n );
@@ -25 +36 @@
     iota1( 42, a[1], 2.0 );
     iota1( 42, a[2], 3.0 );
-    f__bound_ar_allow( 41, a );
+    f__bound_ar_allow( 41, a );  // n == 41  ==>  len(a) == 42  =>  bounds ok
 }