On Tue, May 20, 2014 at 7:19 AM, Michael Manfre wrote:
>
>
> I didn't notice those problems back in January with my pull request
> (closed as won't merge) that made this exact change. [1] I haven't devoted
> much time toward testing schema migrations since, so perhaps the
>
On Tue, May 20, 2014 at 8:11 AM, Andrew Godwin wrote:
> Hi all,
>
> It's done as a raw DROP TABLE because it has to run in weird situations
> with transactions and models where the core delete_model doesn't work (if
> you swap it out, you'll see other failures). We can
Hi all,
It's done as a raw DROP TABLE because it has to run in weird situations
with transactions and models where the core delete_model doesn't work (if
you swap it out, you'll see other failures). We can definitely work towards
improving this, perhaps by moving this onto a method on the
Hi Maximiliano, Michael
Thanks for the clarification.
I'm not 100% up to speed with the new migrations code, but I agree that
this looks odd; it's certainly worth opening a ticket so that the issue
isn't forgotten. It's even arguable that this might be a 1.7 blocker; even
though it isn't a
Hi Michel,
Thank to explain it better than me.
Regards.
El martes, 20 de mayo de 2014 00:42:35 UTC-3, Michael Manfre escribió:
>
> Hi Russell,
>
> I understand what Maximiliano is asking. It's a flaw with
> SchemaTests.delete_tables [1] and it is essentially doing a raw drop table,
> instead
Hi Russell,
I understand what Maximiliano is asking. It's a flaw with
SchemaTests.delete_tables [1] and it is essentially doing a raw drop table,
instead of using the 3rd party backend friendly SchemaEditor.delete_model
[2]. There was the last time I checked (and still exists) a lot of raw SQL
in
On Tue, May 20, 2014 at 7:07 AM, Maximiliano Robaina
wrote:
> Hi,
>
> Running tests, (I'm using firebird) in schema app, I see that the tearDown
> method deletes all tables created by each test. The problem here is for
> tables with AutoInc field where sequences are created
Hi,
Running tests, (I'm using firebird) in schema app, I see that the tearDown
method deletes all tables created by each test. The problem here is for
tables with AutoInc field where sequences are created independently, when
the table is dropped we need delete the sequence object too.
The, to