This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[commit] gdb.cp/temargs.exp: New GCC PR debug/49546 test
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: gdb-patches at sourceware dot org
- Date: Fri, 1 Jul 2011 20:38:38 +0200
- Subject: [commit] gdb.cp/temargs.exp: New GCC PR debug/49546 test
Hi,
checked in. Test:
class DW_AT_name does not match method's linkage name prefix
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49546
It currently PASSes with physname but going to XFAIL it by:
Re: [patch] Follow DW_AT_linkage_name for methods #2
http://sourceware.org/ml/gdb-patches/2011-06/msg00040.html
This is the only known regression of that DW_AT_linkage_name patch to me.
It is wrongly produced DWARF by GCC but sure GDB could workaround it.
I do not have a workaround at hand, I find the workaround either easy or nice
but never both.
So far I plan to just wait for a GCC fix, it should affect only pointers to
functions as template parameters which result in inaccessible template
types/values from GDB.
Thanks,
Jan
http://sourceware.org/ml/gdb-cvs/2011-07/msg00020.html
--- src/gdb/testsuite/ChangeLog 2011/07/01 18:27:25 1.2770
+++ src/gdb/testsuite/ChangeLog 2011/07/01 18:32:06 1.2771
@@ -1,3 +1,11 @@
+2011-07-01 Jan Kratochvil <jan.kratochvil@redhat.com>
+
+ Test GCC PR debug/49546.
+ * gdb.cp/temargs.exp (set sixth breakpoint for temargs)
+ (test type of F in k3_m, test value of F in k3_m): New.
+ * gdb.cp/temargs.cc (struct S3, struct K3): New.
+ (main): New variable k3. Call k3.k3_m.
+
2011-07-01 Jean-Charles Delay <delay@adacore.com>
* gdb.ada/packed_array.exp: Fix expected outout.
--- src/gdb/testsuite/gdb.cp/temargs.cc 2011/06/10 16:21:47 1.3
+++ src/gdb/testsuite/gdb.cp/temargs.cc 2011/07/01 18:32:06 1.4
@@ -65,6 +65,21 @@
}
};
+// GCC PR debug/49546
+struct S3
+{
+ static void m (int x) {}
+};
+template <void (*F) (int)>
+// or: template <void (F) (int)>
+struct K3
+{
+ void k3_m ()
+ {
+ F (0); // Breakpoint 6.
+ }
+};
+
int main ()
{
Base<double, 23, &a_global, &S::f> base;
@@ -72,12 +87,15 @@
// That would be worth testing, once g++ is fixed.
Base<long, 47, &a_global, &S::f>::Inner<float> inner;
K2<&S::somefunc> k2;
+ K3<&S3::m> k3;
+// or: K3<S3::m> k3;
base.base_m ();
inner.inner_m ();
func<unsigned char, 91, &a_global, &S::f> ();
base.templ_m<short> ();
k2.k2_m ();
+ k3.k3_m ();
return 0;
}
--- src/gdb/testsuite/gdb.cp/temargs.exp 2011/06/10 16:21:47 1.4
+++ src/gdb/testsuite/gdb.cp/temargs.exp 2011/07/01 18:32:06 1.5
@@ -75,6 +75,10 @@
gdb_test "break $srcfile:$line" "Breakpoint 6.*" \
"set fifth breakpoint for temargs"
+set line [gdb_get_line_number "Breakpoint 6" $srcfile]
+gdb_test "break $srcfile:$line" "Breakpoint 7.*" \
+ "set sixth breakpoint for temargs"
+
#
# Tests in Base::base_m.
#
@@ -166,3 +170,19 @@
setup_kfail gcc/49366 "*-*-*"
gdb_test "print F" "&S::somefunc" "test value of F in k2_m"
+
+#
+# Tests in K3::k3_m, GCC PR debug/49546.
+# The problem reproduces with DW_AT_MIPS_linkage_name. It does not happen with
+# GDB physname - GDB's own computation of the linkage name based on
+# (incorrectly output by GCC) DW_AT_name.
+#
+
+gdb_continue_to_breakpoint "continue to sixth breakpoint for temargs"
+
+if $have_older_template_gcc { setup_xfail "*-*-*" }
+gdb_test "ptype F" {type = void \(\*\)\(int\)} "test type of F in k3_m"
+
+if $have_older_template_gcc { setup_xfail "*-*-*" }
+gdb_test "print F" { = \(void \(\*\)\(int\)\) 0x[0-9a-f]+ <S3::m\(int\)>} \
+ "test value of F in k3_m"