TransactionMiddleware assumes that when process_request is called,
either process_exception or process_response is guaranteed to be
called. This isn't the case if another middleware raises an exception
first. If this happens, the call to
transaction.enter_transaction_management is not balanced
On Thu, Apr 16, 2009 at 11:14 AM, Horacio de Oro wrote:
>
> On Apr 15, 9:53 pm, Jacob Kaplan-Moss
> wrote:
>> On Wed, Apr 15, 2009 at 7:48 PM, Horacio de Oro wrote:
>>
>> > Is this a bug? Should I file a bug report?
>>
>> Yes,
On Apr 15, 9:53 pm, Jacob Kaplan-Moss
wrote:
> On Wed, Apr 15, 2009 at 7:48 PM, Horacio de Oro wrote:
>
> > Is this a bug? Should I file a bug report?
>
> Yes, please do. Thanks!
>
> Jacob
Testing with another applicatoin, I've found that
On Wed, Apr 15, 2009 at 7:48 PM, Horacio de Oro wrote:
> Is this a bug? Should I file a bug report?
Yes, please do. Thanks!
Jacob
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"Django
Hi!
I get an AttributeError while trying PostgreSql 8.4beta1 (compiled
from 'master' branch of the official git repository). See:
python2.4 src/dynamicware/manage.py test xx.FeedRetrieverServiceTest
Creating test database...
Traceback (most recent call last):
File "src/dynamicware/manage.py",
On Apr 15, 11:04 am, David Cramer wrote:
> I was never a fan of the profile model as it stands. It's not very
> practical in every situation I've ever been in. Being that 1.0 was
> supposed to be backwards compatible, and this is a public API I think
> it needs to stay.
>
>
On 15 Apr 2009, at 22:25, Karen Tracey wrote:
>> So am I missing something or should this be changed?
>
> TestCase testcases should not be running after TransactionTestCase
> testcases. I had this concern prior to checking in the test changes
> (see
>
(I'm posting this here rather than filing a ticket, because I'm not
sure whether I'm missing something. If I'm not than I will file a
ticket with a patch).
It strikes me that TransactionTestCase should be flushing the database
in the teardown stage of the test case rather than the setup.
I was never a fan of the profile model as it stands. It's not very
practical in every situation I've ever been in. Being that 1.0 was
supposed to be backwards compatible, and this is a public API I think
it needs to stay.
I'd love to see a way to swap out the User model in the future though,