From: Ian Bicking <ianb@co...> - 2006-01-26 05:52:19
There was some question about the threadsafety of creating SQLObject
instances, and Kevin Dangoor tracked it down to .sqlmeta not being
created properly when you create an object (as opposed to fetching an
object). I *think* the commit (1539) I made fixed this in the trunk
(and should be backported to 0.7-bugfix). But I'll be damned if I can
get the tests to run. SQLite is giving me encoding errors out the
wazoo. Postgres has weird test interaction bugs -- I really should get
rid of all the crap in dbtest and just clear everything more
aggressively between tests. Sigh. Anyway, I'd appreciate it if other
people can run the tests, and either tell me they are totally broken for
everyone, or that the fix worked and my computer is messed up.
Ian Bicking | ianb@... | http://blog.ianbicking.org
On Wed, Jan 25, 2006 at 11:52:42PM -0600, Ian Bicking wrote:
> But I'll be damned if I can
> get the tests to run. SQLite is giving me encoding errors out the
> wazoo. Postgres has weird test interaction bugs
The test suite passed for me just fine. Python 2.3 and 2.4, SQLite
3.2.8, PySQLite 1.1.6 and 2.0.5, Postgres 7.4, psycopg 1.1.21 and 2.0b5.
The only minor problem I have is in test_ForeignKey.py test2() - it
cannot setup its two classes again it cannot drop TestComposerKey because
there are ForeignKeys in the TestWorkKey table.
Oleg Broytmann http://phd.pp.ru/ phd@...
Programmers don't die, they just GOSUB without RETURN.