KInterbasDB

Logged In: YES
user_id=414645

Fixed in CVS rev 1.23 of _kinterbasdb.c

Logged In: YES
user_id=414645

I found the problem. It was the same silly strcat call
that gave rise to another bug:

http://sourceforge.net/tracker/index.php?
func=detail&aid=551207&group_id=9913&atid=109913


A fix is coming soon to a CVS repository near you.

Logged In: YES
user_id=414645

Thanks for the details.

I happen to have a Mandrake 8.1 installation, so I'll use
that to investigate.

Logged In: NO

This is kinterbasdb 3.0 final on Mandrake 8.1, kernel
2.4.8-34.1mdk talking to a LI-V6.2.796 Firebird 1.0 server
with a dialect 1 database on Debian Woody.

I left ZFirebird DA out of things, by using the Python 2.1
command line:

>>> import kinterbasdb
>>> kinterbasdb.__version__
(3, 0, 0, 'final', 0)
>>> conn = kinterbasdb.connect('server:/path/db.gdb',
'user', 'pass')
>>> curs = conn.cursor()
>>> curs.execute('update vardata set id=-999 where id=-999')
>>> curs.fetchone()
Segmentation fault

Logged In: YES
user_id=414645

I need more specifics:
- version of kinterbasdb
- operating system
- version of ZFirebirdDA

I tested a non-query Z SQL Method with kinterbasdb 3.0
(when I refer to 3.0 without further elaboration, I mean
3.0 final, not one of the prereleases), ZFirebirdDA 0.0.1,
Zope 2.5.0, Python 2.1.2, and WinNT 4.0 SP6a. It certainly
didn't segfault; it raised a kinterbasdb.ProgrammingError
(0, 'Attempt to fetch after non-query statement...').

I then tested with ZFirebirdDA 0.0.2; the improper fetching
has been fixed.

As you pointed out, the soundness of ZFirebirdDA is
immaterial in this context, because kinterbasdb should not
segfault no matter how it is abused by a client
programmer. However, *I observed no segfault with
kinterbasdb 3.0*.

Perhaps you are using a version of kinterbasdb previous to
3.0 final, and the segfault you observed is related to this
bug (which is fixed in 3.0 final):
http://sourceforge.net/tracker/index.php?
func=detail&aid=517842&group_id=9913&atid=109913


-------------
On a somewhat related note, I discovered a minor bug while
investigating your bug report:
http://sourceforge.net/tracker/index.php?
func=detail&aid=551207&group_id=9913&atid=109913