RFC: fix PR 14635

[Tom Tromey <tromey at redhat dot com>]

This fixes PR python/14635.

Like PR 14634, this is a destructor bug.

In this case, the Python SAL code assumes that the 'symtab' field will
always be a valid object; but del_objfile_sal incorrectly sets this
field to NULL.

This fixes the problem in the obvious way.

Built and regtested on x86-64 Fedora 16.
New test case included.

Tom

	PR python/14635:
	* python/py-symtab.c (del_objfile_sal): Set 'symtab' field
	to Py_None.

	* gdb.python/py-symtab.exp: Test sal and symtab destructors.
---
 gdb/python/py-symtab.c                 |    5 ++++-
 gdb/testsuite/gdb.python/py-symtab.exp |    3 +++
 2 files changed, 7 insertions(+), 1 deletions(-)

diff --git a/gdb/python/py-symtab.c b/gdb/python/py-symtab.c
index b6f45e6..21a172c 100644
--- a/gdb/python/py-symtab.c
+++ b/gdb/python/py-symtab.c
@@ -470,7 +470,10 @@ del_objfile_sal (struct objfile *objfile, void *datum)
     {
       sal_object *next = obj->next;
 
-      obj->symtab = NULL;
+      Py_DECREF (obj->symtab);
+      obj->symtab = (symtab_object *) Py_None;
+      Py_INCREF (Py_None);
+
       obj->next = NULL;
       obj->prev = NULL;
       xfree (obj->sal);
diff --git a/gdb/testsuite/gdb.python/py-symtab.exp b/gdb/testsuite/gdb.python/py-symtab.exp
index da0c3e4..0d3a5f2 100644
--- a/gdb/testsuite/gdb.python/py-symtab.exp
+++ b/gdb/testsuite/gdb.python/py-symtab.exp
@@ -73,3 +73,6 @@ gdb_test "python print \"simple_struct\" in static_symbols" "True" "Test simple_
 gdb_unload
 gdb_test "python print sal.is_valid()" "False" "Test sal.is_valid"
 gdb_test "python print symtab.is_valid()" "False" "Test symtab.is_valid()"
+
+gdb_test_no_output "python sal = None" "Test sal destructor"
+gdb_test_no_output "python symtab = None" "Test symtab destructor"
-- 
1.7.7.6