diff mbox

C++ PATCH for c++/60379 (bogus error with loop in template)

Message ID 531249B5.9010209@redhat.com
State New
Headers show

Commit Message

Jason Merrill March 1, 2014, 8:57 p.m. UTC 
We shouldn't complain if a loop condition isn't a valid constant 
expression; it isn't supposed to be.

Tested x86_64-pc-linux-gnu, applying to trunk.
diff mbox

Patch

commit 5266ff07b6077df5d7bf6d5748856a43e48ad6ee
Author: Jason Merrill <jason@redhat.com>
Date:   Sat Mar 1 14:22:02 2014 -0500

    	PR c++/60379
    	* semantics.c (begin_maybe_infinite_loop): Use
    	fold_non_dependent_expr_sfinae.

diff --git a/gcc/cp/semantics.c b/gcc/cp/semantics.c
index 9c1c29d..eaeeb24 100644
--- a/gcc/cp/semantics.c
+++ b/gcc/cp/semantics.c
@@ -503,7 +503,7 @@  begin_maybe_infinite_loop (tree cond)
   bool maybe_infinite = true;
   if (cond)
     {
-      cond = fold_non_dependent_expr (cond);
+      cond = fold_non_dependent_expr_sfinae (cond, tf_none);
       cond = maybe_constant_value (cond);
       maybe_infinite = integer_nonzerop (cond);
     }
diff --git a/gcc/testsuite/g++.dg/template/loop1.C b/gcc/testsuite/g++.dg/template/loop1.C
new file mode 100644
index 0000000..aa6d177
--- /dev/null
+++ b/gcc/testsuite/g++.dg/template/loop1.C
@@ -0,0 +1,9 @@ 
+// PR c++/60379
+
+template <int> struct A {
+  void m_fn1(int p1) {
+    int *a;
+    while (p1 && *static_cast<int *>(static_cast<void *>(a)))
+      ;
+  }
+};