| Message ID | 20141017122124.GV4197@redhat.com |
|---|---|
| State | New |
| Headers | show |
commit c388b3dc00c256c9cc5d8ae7b5bc37386bd14c58
Author: Jonathan Wakely <jwakely@redhat.com>
Date: Fri Oct 17 12:57:31 2014 +0100
* testsuite/27_io/basic_ostream/inserters_arithmetic/char/hexfloat.cc:
Add dg-require-string-conversions.
* testsuite/27_io/headers/cstdio/types_std.cc: Test for fpos_t.
diff --git a/libstdc++-v3/testsuite/27_io/basic_ostream/inserters_arithmetic/char/hexfloat.cc b/libstdc++-v3/testsuite/27_io/basic_ostream/inserters_arithmetic/char/hexfloat.cc
index 485a485..7fe6ff8 100644
--- a/libstdc++-v3/testsuite/27_io/basic_ostream/inserters_arithmetic/char/hexfloat.cc
+++ b/libstdc++-v3/testsuite/27_io/basic_ostream/inserters_arithmetic/char/hexfloat.cc
@@ -1,4 +1,5 @@
// { dg-options "-std=gnu++11" }
+// { dg-require-string-conversions "" }
// 2014-03-27 R??diger Sonderfeld
// test the hexadecimal floating point inserters (facet num_put)
diff --git a/libstdc++-v3/testsuite/27_io/headers/cstdio/types_std.cc b/libstdc++-v3/testsuite/27_io/headers/cstdio/types_std.cc
index a359b87..a34663f 100644
--- a/libstdc++-v3/testsuite/27_io/headers/cstdio/types_std.cc
+++ b/libstdc++-v3/testsuite/27_io/headers/cstdio/types_std.cc
@@ -23,5 +23,5 @@ namespace gnu
{
std::size_t s;
std::FILE f;
- std::FILE fpos_t;
+ std::fpos_t p;
}
testsuite/27_io/headers/cstdio/types_std.cc FAILs on dragonflybsd: /mnt/gcc-src/libstdc++-v3/testsuite/27_io/headers/cstdio/types_std.cc:25:13: error: aggregate 'FILE gnu::f' has incomplete type and cannot be defined /mnt/gcc-src/libstdc++-v3/testsuite/27_io/headers/cstdio/types_std.cc:26:13: error: aggregate 'FILE gnu::fpos_t' has incomplete type and cannot be defined These errors look correct to me, the C standard says that <stdio.h> declares FILE as an object type, but it doesn't say complete object type, so I think that's a bug in the test. I think there's another bug: #include <cstdio> namespace gnu { std::size_t s; std::FILE f; std::FILE fpos_t; } Surely that third declaration should be testing that fpos_t is a valid type, rather than declaring a variable of that name, so I'm committing the attached patch, which also fixes another fail on dragonflybsd. Tested x86_64-linux, committed to trunk.