[PATCH] D29197: Avoid implementation defined behavior in a test.

Dan Albert via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Thu Jan 26 13:38:23 PST 2017


danalbert created this revision.

num_put::put uses %p for pointer types, but the exact format of %p is
implementation defined behavior for the C library. Compare output to
snprintf for portability.


Repository:
  rL LLVM

https://reviews.llvm.org/D29197

Files:
  test/std/localization/locale.categories/category.numeric/locale.nm.put/facet.num.put.members/put_pointer.pass.cpp


Index: test/std/localization/locale.categories/category.numeric/locale.nm.put/facet.num.put.members/put_pointer.pass.cpp
===================================================================
--- test/std/localization/locale.categories/category.numeric/locale.nm.put/facet.num.put.members/put_pointer.pass.cpp
+++ test/std/localization/locale.categories/category.numeric/locale.nm.put/facet.num.put.members/put_pointer.pass.cpp
@@ -38,6 +38,12 @@
         char str[50];
         output_iterator<char*> iter = f.put(output_iterator<char*>(str), ios, '*', v);
         std::string ex(str, iter.base());
-        assert(ex == "0x0" || ex == "(nil)");
+        char expected_str[32] = {};
+        // num_put::put uses %p for pointer types, but the exact format of %p is
+        // implementation defined behavior for the C library. Compare output to
+        // snprintf for portability.
+        int rc = snprintf(expected_str, sizeof(expected_str), "%p", v);
+        assert(rc > 0);
+        assert(ex == expected_str);
     }
 }


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D29197.85961.patch
Type: text/x-patch
Size: 1028 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20170126/ccabe771/attachment.bin>


More information about the cfe-commits mailing list