Re: RFA: fix macro expansion bug

[Tom Tromey <tromey@redhat.com>]

>>>>> "Pedro" == Pedro Alves <pedro@codesourcery.com> writes:

>> +gdb_test "macro expand siginfo.si_addr" \
>> +  "expands to: siginfo. fields.fault.si_addr" \

Pedro> Just curious, as it's just a visual annoyance: do you know where
Pedro> this space comes from?  Do we store the definition with the space for
Pedro> some reason?  We don't get that extra space if the define came
Pedro> from the code, instead of from a 'macro define'.

Here's a patch to fix it.

This requires a minor tweak to the previous patch's test case, but
otherwise nothing.

As with the previous, I'm sending this through the regression tester,
but I don't expect problems.  Ok if it passes?

Tom

2008-12-11  Tom Tromey  <tromey@redhat.com>

	* macrocmd.c (macro_define_command): Skip whitespace after
	macro name.
	(print_one_macro): Print space after macro name.

diff --git a/gdb/macrocmd.c b/gdb/macrocmd.c
index 56e9a48..fa639d1 100644
--- a/gdb/macrocmd.c
+++ b/gdb/macrocmd.c
@@ -315,13 +315,17 @@ macro_define_command (char *exp, int from_tty)
 	}
       /* Skip the closing paren.  */
       ++exp;
+      skip_ws (&exp);
 
       macro_define_function (macro_main (macro_user_macros), -1, name,
 			     new_macro.argc, (const char **) new_macro.argv,
 			     exp);
     }
   else
-    macro_define_object (macro_main (macro_user_macros), -1, name, exp);
+    {
+      skip_ws (&exp);
+      macro_define_object (macro_main (macro_user_macros), -1, name, exp);
+    }
 
   do_cleanups (cleanup_chain);
 }
@@ -358,9 +362,7 @@ print_one_macro (const char *name, const struct macro_definition *macro,
 			  macro->argv[i]);
       fprintf_filtered (gdb_stdout, ")");
     }
-  /* Note that we don't need a leading space here -- "macro define"
-     provided it.  */
-  fprintf_filtered (gdb_stdout, "%s\n", macro->replacement);
+  fprintf_filtered (gdb_stdout, " %s\n", macro->replacement);
 }