Committed, adjustments for trivial test for .gnu.warning.SYMBOL

[Hans-Peter Nilsson]

Testing with the mips-elf toolchain showed that it default-sets
the type of symbols to OBJECT, so the readelf line didn't match.
As the purpose of that part of the test is to check that the
symbol attributes weren't munged by the warning support, it
seems better to test typical use and particularly where
attributes aren't default.  Hence, this is what I committed as
trivial changes, instead of the previous submitted code.  Same
changelog entry.

--- /dev/null	Tue Oct 29 15:57:07 2002
+++ symbol1w.s	Mon Feb  7 02:38:24 2005
@@ -0,0 +1,10 @@
+	.section .gnu.warning.symbol1
+	.asciz "witty one-liner"
+	.text
+	.type symbol1,@function
+	.global symbol1
+symbol1:
+.L1:
+	.long 0
+.L0:
+	.size symbol1,.L0-.L1
--- /dev/null	Tue Oct 29 15:57:07 2002
+++ warn1.d	Mon Feb  7 02:42:19 2005
@@ -0,0 +1,13 @@
+#source: start.s
+#source: symbol1ref.s
+#source: symbol1w.s
+#ld: 
+#warning: ^[^\\n]*\): warning: witty one-liner$
+#readelf: -s
+
+# Check that warnings are generated for the .gnu.warning.SYMBOL
+# construct and that the symbol still appears as expected.
+
+#...
+[ 	]+[0-9]+:[ 	]+[0-9a-f]+[ 	]+[48][ 	]+FUNC[	 ]+GLOBAL DEFAULT[ 	]+[1-9] symbol1
+#pass

brgds, H-P