obtained gcc-4.6.4.tar.bz2 from upstream website;upstream

verified gcc-4.6.4.tar.bz2.sig;
imported gcc-4.6.4 source tree from verified upstream tarball.

downloading a git-generated archive based on the 'upstream' tag
should provide you with a source tree that is binary identical
to the one extracted from the above tarball.

if you have obtained the source via the command 'git clone',
however, do note that line-endings of files in your working
directory might differ from line-endings of the respective
files in the upstream repository.

1 files changed, 131 insertions, 0 deletions

diff --git a/gcc/testsuite/c-c++-common/dfp/convert-bfp.c b/gcc/testsuite/c-c++-common/dfp/convert-bfp.c
new file mode 100644
index 000000000..710538d46
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/dfp/convert-bfp.c
@@ -0,0 +1,131 @@
+/* N1150 5.2 Conversions among decimal floating types and between
+   decimal floating types and generic floating types.
+   C99 6.3.1.5(4) Conversions, arithmetic operands, real floating types.  */
+
+/* Long double isn't supported yet at runtime, so disable those checks.  */
+
+#include "dfp-dbg.h"
+
+static int skip_long_double;
+
+volatile _Decimal32 d32;
+volatile _Decimal64 d64;
+volatile _Decimal128 d128;
+volatile float sf;
+volatile double df;
+volatile long double tf;
+
+int
+main ()
+{
+  /* Conversions from decimal float to binary float. */
+
+  if (sizeof (long double) == sizeof (double))
+    skip_long_double = 1;
+
+  /* Conversions from _Decimal32. */
+  d32 = 2.0df;
+  sf = d32;
+  if (sf != 2.0f)
+    FAILURE
+
+  df = d32;
+  if (df != 2.0)
+    FAILURE
+
+  if (skip_long_double == 0)
+    {
+      tf = d32;
+      if (tf != 2.0l)
+	FAILURE
+    }
+
+  /* Conversions from _Decimal64. */
+  d64 = -7.0dd;
+  sf = d64;
+  if (sf != -7.0f)
+    FAILURE
+  
+  df = d64;
+  if (df != -7.0)
+    FAILURE
+
+  if (skip_long_double == 0)
+    {
+      tf = d64;
+      if (tf != -7.0l)
+	FAILURE
+    }
+
+  /* Conversions from _Decimal128. */
+  d128 = 30.0dl;
+  sf = d128;
+  if (sf != 30.0f)
+    FAILURE
+
+  df = d128;
+  if (df != 30.0)
+    FAILURE
+
+  df = d128;
+  if (df != 30.0l)
+    FAILURE
+
+  /* Conversions from binary float to decimal float. */
+  sf = 30.0f;
+  d32 = sf;
+  if (d32 != 30.0df)
+    FAILURE
+
+  d64 = sf;
+  if (d64 != 30.0dd)
+    FAILURE
+
+  df = -2.0;
+  d32 = df;
+  if (d32 != -2.0df)
+    FAILURE
+
+  d64 = df;
+  if (d64 != -2.0dd)
+    FAILURE
+
+  d128 = df;
+  if (d128 != -2.0dl)
+    FAILURE
+  
+  sf = 30.0f;
+  d128 = sf;
+  if (d128 != 30.0dl)
+    FAILURE
+
+  if (skip_long_double == 0)
+    {
+      tf = -22.0l;
+      d32 = tf;
+      if (d32 != -22.0df)
+	FAILURE
+
+      d64 = tf;
+      if (d64 != -22.0dd)
+	FAILURE
+
+      d128 = tf;
+      if (d128 != -22.0dl)
+	FAILURE
+     }
+
+  /* 2**(-11) = 0.00048828125. */
+  d128 = 0.000488281251dl;
+  sf = d128;
+  if (sf != 0.00048828125f)
+    FAILURE
+  /* 2**(-25) = 0.298023223876953125E-7.  */
+  d128 = 2.98023223876953125E-8dl;
+  df = d128;
+  if (df < (2.9802322387695312e-08 - 0.00000000001)
+      || df > (2.9802322387695312e-08 + 0.00000000001))
+    FAILURE
+
+  FINISH
+}