Closed msheiny closed 7 years ago
Opened separate issues to track the follow-up changes to gulp and debug mode. I'm tempted to open yet another issue to resolve the fixtures madness, but don't have enough info yet—thoughts, @msheiny?
Re: fixtures madness, I have a theory that it might be due to: http://docs.wagtail.io/en/v1.7/releases/1.7.html#project-template-s-initial-migration-should-not-depend-on-wagtailcore-latest
@garrettr Believes that the fixtures madness is caused by the role running createsuperuser
before the fixtures are applied. That sounds logical! Let's block on merge here until that's resolved—all other changes look good.
After some review, I'm not sure about the cause of "fixtures madness" anymore. I tried various combinations of running manage.py createsuperuser
before manage.py loaddata
in an attempt to trigger the error locally, but was unable to do so. I checked the fixture docs and it looks like loaddata
always overwrites any conflicting entries in the database, so that probably isn't the cause of the issue.
Since I can't reproduce this locally, and it's only been observed on the production machine, my gut feeling is that the issue is with one of the biggest differences between the development and production environments: sqlite vs. postgres for the db. It's also possible that we're encountering a bug in Wagtail - I just upgraded to 1.7 last week, so it's possible that we're encountering an issue related to that.
I'll look over the Django deployment code and see if I can spot other potential causes.
Is this branch good to merge in @garrettr / @conorsch ?
@msheiny I'll open a separate issue to track "fixtures madness", then yes!
Below were the commits that @conorsch and I had to apply to get STN deployed to a staging server. We still have a number of issues that need to get debugged in a future PR:
debug
set toTrue