gvalue: Add explicit casts in numeric transform functions

Compiling with `-Wfloat-conversion` warns about a few implicit
conversions from `double`/`float` to other numeric types in the `GValue`
transform functions.

These warnings are correct: value transformations can result in loss of
precision. That loss of precision is understood and expected, so add
some explicit casts to squash the warnings.

Signed-off-by: Philip Withnall <pwithnall@gnome.org>

Helps: #3405

gobject/gvaluetypes.c

@@ -190,7 +190,8 @@ value_collect_float (GValue      *value,
 		     GTypeCValue *collect_values,
 		     guint        collect_flags)
 {
-  value->data[0].v_float = collect_values[0].v_double;
+  /* This necessarily loses precision */
+  value->data[0].v_float = (gfloat) collect_values[0].v_double;
   
   return NULL;
 }