I have gotten bitten by lingering bytecode in migrations on several
occasions. Steps to reproduce are a little bit complex, but here is an
rough example:
Local Box:
— have *.pyc rule in .gitignore
— create some migrations
— commit them
Server Box:
— pull repository
— execute migrations
Local Box:
— remove migrations
— commit
— create new initial migration
Server Box:
— pull repository
— remove migrations history from db
— run fake initial migration — get an error here
At this point, the server box has a migrations directory that looks
something like this (bytecode is still around because it was in the
gitignore):
— migrations
— __init__.py
— 0001_initial.py
— 0001_initial.pyc
— 0002_second.pyc
— 0003_third.pyc
When running "python manage migrate" the bytecode appears to be referenced,
which often causes errors when running the migrations. When using git, for
example "git clean -df" clears out the problem. What I am wondering is if
it might make sense to make a deliberate attempt to ignore bytecode without
accompanying *.py files. This appears to be an issue because of the way
that migrations dynamically import python code, but I am not sure. Perhaps
the problem is more subtle as I am not deeply familiar with the way that
migrations work.
aryeh
--
You received this message because you are subscribed to the Google Groups
"Django developers (Contributions to Django itself)" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/django-developers.
To view this discussion on the web visit
https://groups.google.com/d/msgid/django-developers/1dbf1434-b24a-47c9-b45b-ae04530d7867%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
