What steps will reproduce the problem?
1. Create a stored procedure with a date input value that is allowed to be
null.
2. Call the stored procedure with a null date
What is the expected output? What do you see instead?
I expect the procedure to not segfault.
What version of the product are you using? On what operating system?
ibm_db 2.0.4
ibm_db 2.0.5
xubuntu 12.10
Please provide any additional information below.
ibm_db 2.0.2 does not have this problem.
We came across a problem with running stored procedures that causes ibm_db to
segfault. This procedure allows nulls on a date. When the null is passed in
we suspect that it's hitting the new case SQL_TYPE_DATE and then it segfaults
when it accesses tmp_curr->date_value->year. It use to fall through to default
which correctly checks for nulls. The new types don't check for this.
Below is the code that should work for just the single case. I built and
quickly checked it and it worked.
diff --git a/IBM_DB/ibm_db/ibm_db.c b/IBM_DB/ibm_db/ibm_db.c
index 2aacf19..10b9074 100644
--- a/IBM_DB/ibm_db/ibm_db.c
+++ b/IBM_DB/ibm_db/ibm_db.c
@@ -10134,9 +10134,14 @@ static PyObject* ibm_db_callproc(PyObject *self,
PyObject *args){
paramCount++;
break;
case SQL_TYPE_DATE:
-
PyTuple_SetItem(outTuple, paramCount,
PyDate_FromDate(tmp_curr->date_value->year,
-
tmp_curr->date_value->month, tmp_curr->date_value->day));
- paramCount++;
+ if
(!NIL_P(tmp_curr->date_value)) {
+
PyTuple_SetItem(outTuple, paramCount, PyDate_FromDate(tmp_curr->date_value->ye
+
tmp_curr->date_value->month, tmp_curr->date
+ } else {
+
Py_INCREF(Py_None);
+
PyTuple_SetItem(outTuple, paramCount, Py_None);
+ }
+ paramCount++;
break;
case SQL_TYPE_TIME:
PyTuple_SetItem(outTuple, paramCount, PyTime_FromTime(tmp_curr->time_value->hour,
Original issue reported on code.google.com by eric.per...@gmail.com on 12 May 2014 at 9:46
Original issue reported on code.google.com by
eric.per...@gmail.comon 12 May 2014 at 9:46