Re: Regression: Re: [3/3] RFC: `function:label' linespecs

[Tom Tromey <tromey@redhat.com>]

Tom> Sorry about that.  I am working on a fix.

Here is what I am checking in.

This restores some old code that was exercised by this particular test
case.

Built and regtested on x86-64 (compile farm).
I also ran the failing case locally by hand.

Tom

2011-03-25  Tom Tromey  <tromey@redhat.com>

	* linespec.c (symbol_found): Restore line-based result for
	non-LOC_LABEL symbols.

Index: linespec.c
===================================================================
RCS file: /cvs/src/src/gdb/linespec.c,v
retrieving revision 1.116
diff -u -r1.116 linespec.c
--- linespec.c	24 Mar 2011 20:25:14 -0000	1.116
+++ linespec.c	25 Mar 2011 17:11:20 -0000
@@ -2174,9 +2174,7 @@
     }
   else
     {
-      if (funfirstline && SYMBOL_CLASS (sym) != LOC_LABEL)
-	error (_("\"%s\" is not a function"), copy);
-      else if (SYMBOL_VALUE_ADDRESS (sym) != 0)
+      if (SYMBOL_CLASS (sym) == LOC_LABEL && SYMBOL_VALUE_ADDRESS (sym) != 0)
 	{
 	  /* We know its line number.  */
 	  values.sals = (struct symtab_and_line *)
@@ -2201,6 +2199,20 @@
 
 	  return values;
 	}
+      else if (funfirstline)
+	error (_("\"%s\" is not a function"), copy);
+      else if (SYMBOL_LINE (sym) != 0)
+	{
+	  /* We know its line number.  */
+	  values.sals = (struct symtab_and_line *)
+	    xmalloc (sizeof (struct symtab_and_line));
+	  values.nelts = 1;
+	  memset (&values.sals[0], 0, sizeof (values.sals[0]));
+	  values.sals[0].symtab = SYMBOL_SYMTAB (sym);
+	  values.sals[0].line = SYMBOL_LINE (sym);
+	  values.sals[0].pspace = SYMTAB_PSPACE (SYMBOL_SYMTAB (sym));
+	  return values;
+	}
       else
 	/* This can happen if it is compiled with a compiler which doesn't
 	   put out line numbers for variables.  */