On Monday, 18 August 2014 19:07:06 UTC+1, Michael Bayer wrote:
>
>
> On Aug 18, 2014, at 11:43 AM, Stephan Hügel <[email protected] 
> <javascript:>> wrote:
>
> On Monday, 18 August 2014 16:07:30 UTC+1, Michael Bayer wrote:
>
>>
>> On Aug 18, 2014, at 11:02 AM, Michael Bayer <[email protected]> 
>> wrote:
>>
>>
>> On Aug 18, 2014, at 11:00 AM, Michael Bayer <[email protected]> 
>> wrote:
>>
>>
>> On Aug 18, 2014, at 10:21 AM, Stephan Hügel <[email protected]> wrote:
>>
>>
>> Yep, I'm using naming_convention, and can confirm that the drop operation 
>>> isn't being wrapped by f(). 
>>>
>>
>> (And also autogenerate, for the avoidance of ambiguity) 
>>
>>
>> the rendering of drop_constraint does not include the “type” parameter 
>> (it probably should but that’s a separate issue).  However I see this 
>> parameter rendered in your script at 
>> https://gist.github.com/urschrei/541fec05a3a82d71cbe9.
>>
>> also I have added a test for this case and am not able to reproduce - the 
>> op.f() is rendered in the drop case for the unique constraint.
>>
>> are you sure this is straight autogenerate for the “drop” ?
>>
>>
>>
>> still unsure about “type” but if i give the constraint a name, then we 
>> are seeing the op.f() omitted, so that must be your case.
>>
>>
>> ummm except that’s the correct behavior depending on the convention in use
>>
>> I need to see your naming convention fully as well as your Table 
>> definition please, database in use as well
>>
>>
>>
> I've reinstalled alembic just to be sure, and this time it's wrapped both 
> the create_constraint and drop_constraint name in f when naming conventions 
> and autogenerate are used, but downgrade() still fails with the same 
> IndexError.
>
> Gist below has naming convention setup (I'm *cough* manually adding it to 
> the MetaData instance because of the flask-sqlalchemy magic), mixins, table 
> model, alembic env.py, dump of MySQL variables, and autogenerated migration 
> (I've manually added the type_ kw to drop_constraint, though – I assumed 
> that was necessary)
>
> https://gist.github.com/urschrei/0ba40dff49b7845ebfc1 
>
>
> the stack trace is the same *with* op.f() sent to drop_constraint?   
> that’s strange because the whole naming convention thing should be skipped 
> entirely if that’s the case.   are you sure your SQLAlchemy is also on the 
> latest version?
>
>
Argh there was a global install of 0.9.4 hanging around. Sorry sorry sorry. 

-- 
You received this message because you are subscribed to the Google Groups 
"sqlalchemy-alembic" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to