python-pgsql - stacktrace when using arrays

[Guillaume Pratte]

Binary package hint: python-pgsql

The following code causes a stacktrace in python-pgsql :

------
import PgSQL
conn = PgSQL.connect(user="user",password="pass",database="db",host="host")
cur.execute("SELECT ARRAY[now()]")
res = cur.fetchall()
------

Here is the stacktrace (test.py is the test file containing the above code):

------
Traceback (most recent call last):
  File "/usr/lib/python2.5/site-packages/pyPgSQL/test.py", line 5, in <module>
    res = cur.fetchall()
  File "/usr/lib/python2.5/site-packages/pyPgSQL/PgSQL.py", line 3247, in fetchall
    _list.append(self.__fetchOneRow())
  File "/usr/lib/python2.5/site-packages/pyPgSQL/PgSQL.py", line 2814, in __fetchOneRow
    _r.getvalue(self._idx_, _i)))
  File "/usr/lib/python2.5/site-packages/pyPgSQL/PgSQL.py", line 890, in typecast
    return self.handleArray(colinfo, _list)
  File "/usr/lib/python2.5/site-packages/pyPgSQL/PgSQL.py", line 999, in handleArray
    if _fvt != PG_INTERVAL:
NameError: global name '_fvt' is not defined
------

This bug is present in edgy, feisty and gutsy, but not in dapper. It was not tested in hardy.

We can conclude the following :
* the bug appeared in version 2.5.1 (since edgy)
* the bug is not caused from the upgrade of python 2.4 to python 2.5 (since the bug is in edgy, which uses python 2.4)

This bug is pretty strange, since '_fvt' seems to be used in the live above the error (line 998).

[Luca Falavigna (dktrkranz)]

Reported upstream: http://sourceforge.net/tracker/index.php?func=detail&aid=1854028&group_id=16528&atid=116528.

[Luca Falavigna (dktrkranz)]

python-pgsql (2.5.1-2ubuntu1) hardy; urgency=low

  * pyPgSQL/PgSQL.py: fix typo which caused NameError (LP: #177268)
  * debian/control: update Maintainer field as per spec

 -- Luca Falavigna <email address hidden> Wed, 19 Dec 2007 15:28:04 +0100