[rfa/testsuite] Add test for PR gdb/38

[rfa/testsuite] Add test for PR gdb/38

[David Carlton]

This patch adds a KFAILed test for PR gdb/38.  That PR is about
watchpoints set at the same location as breakpoints, which can cause
the watchpoint to get triggered late.  Right now, the only testsuite
case that tests this is in gdb.c++/annota2.exp, which is bad because
it has nothing to do with C++, nothing to do with annotation, and
because level 2 annotation is on its way out.  So I'd like to get the
bug recorded in a more correct location.

Tested on i686-pc-linux-gnu, GCC 3.1, DWARF 2.  I haven't been able to
test the pass branch, because it reliably fails on this platform, but
I did try to be careful with the pass regexp.

OK to commit?

David Carlton
carlton@math.stanford.edu

2003-03-17  David Carlton  <carlton@math.stanford.edu>

	* gdb.base/watchpoint.exp (test_watchpoint_and_breakpoint): New
	test, for PR breakpoints/38.
	Call test_watchpoint_and_breakpoint.
	* gdb.base/watchpoint.c (func3): New function.
	(main): Call func3.

Index: watchpoint.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/watchpoint.exp,v
retrieving revision 1.7
diff -u -p -r1.7 watchpoint.exp
--- watchpoint.exp	4 Feb 2003 22:49:19 -0000	1.7
+++ watchpoint.exp	17 Mar 2003 18:29:11 -0000
@@ -648,6 +648,27 @@ proc test_complex_watchpoint {} {
     }
 }
 
+proc test_watchpoint_and_breakpoint {} {
+    global gdb_prompt
+
+    # This is a test for PR gdb/38, which involves setting a
+    # watchpoint right after you've reached a breakpoint.
+
+    if [runto func3] then {
+	gdb_breakpoint [gdb_get_line_number "second x assignment"]
+	gdb_continue_to_breakpoint "second x assignment"
+	gdb_test "watch x" ".*atchpoint \[0-9\]+: x"
+	gdb_test_multiple "next" "next after watch x" {
+	    -re ".*atchpoint \[0-9\]+: x\r\n\r\nOld value = 0\r\nNew value = 1\r\n.*$gdb_prompt $" {
+		pass "next after watch x"
+	    }
+	    -re "\[0-9\]+\[\t \]+y = 1;\r\n$gdb_prompt $" {
+		kfail "gdb/38" "next after watch x"
+	    }
+	}
+    }
+}
+    
 # Start with a fresh gdb.
 
 gdb_exit
@@ -799,6 +820,17 @@ if [initialize] then {
         timeout {fail "(timeout) access watches disallowed"}
       }
     }
+
+    # See above.
+    if [istarget "mips-idt-*"] then {
+	gdb_exit
+	gdb_start
+	gdb_reinitialize_dir $srcdir/$subdir
+	gdb_load $binfile
+	initialize
+    }
+
+    test_watchpoint_and_breakpoint
 }
 
 # Restore old timeout
Index: watchpoint.c
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/watchpoint.c,v
retrieving revision 1.1.1.2
diff -u -p -r1.1.1.2 watchpoint.c
--- watchpoint.c	28 Jun 1999 16:04:22 -0000	1.1.1.2
+++ watchpoint.c	17 Mar 2003 18:29:07 -0000
@@ -83,6 +83,18 @@ func2 ()
   static_b = local_a;
 }
 
+void
+func3 ()
+{
+  int x;
+  int y;
+
+  x = 0;
+  x = 1;				/* second x assignment */
+  y = 1;
+  y = 2;
+}
+
 int
 func1 ()
 {
@@ -170,5 +182,8 @@ int main ()
   recurser (2);
 
   marker6 ();
+
+  func3 ();
+
   return 0;
 }

Re: [rfa/testsuite] Add test for PR gdb/38

[Daniel Jacobowitz]

On Mon, Mar 17, 2003 at 10:35:51AM -0800, David Carlton wrote:
> This patch adds a KFAILed test for PR gdb/38.  That PR is about
> watchpoints set at the same location as breakpoints, which can cause
> the watchpoint to get triggered late.  Right now, the only testsuite
> case that tests this is in gdb.c++/annota2.exp, which is bad because
> it has nothing to do with C++, nothing to do with annotation, and
> because level 2 annotation is on its way out.  So I'd like to get the
> bug recorded in a more correct location.
> 
> Tested on i686-pc-linux-gnu, GCC 3.1, DWARF 2.  I haven't been able to
> test the pass branch, because it reliably fails on this platform, but
> I did try to be careful with the pass regexp.
> 
> OK to commit?

Yes, thank you.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: [rfa/testsuite] Add test for PR gdb/38

[David Carlton]

On Mon, 17 Mar 2003 14:42:30 -0500, Daniel Jacobowitz <drow@mvista.com> said:
> On Mon, Mar 17, 2003 at 10:35:51AM -0800, David Carlton wrote:

>> OK to commit?

> Yes, thank you.

Thanks, committed.

David Carlton
carlton@math.stanford.edu