Changes between Initial Version and Version 1 of Ticket #23599, comment 6


Ignore:
Timestamp:
Oct 17, 2014, 2:46:56 PM (10 years ago)
Author:
Andrew Godwin

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #23599, comment 6

    initial v1  
    11Well, the bug mentioned above is definitely not to do with the empty migrations folder (which, as was correctly posted, is so new apps are not ignored by makemigrations - we can remove the empty folder when we deprecate the syncdb-style path in 2.0, as at that point makemigrations can assume everything has migrations).
    22
    3 The root of the cause here seems to be the custom user model, which as the docs mention in several places, must be made in the first migration of an app. ``makemigrations`` is meant to have code that ignores this style of error while it's running (as you may be doing it to make the migration for said user model), but I've not seen that result in this error before - usually it manifested as a graph dependency error (as the migration depending on the custom user model would have a dependency on an app with no migrations).
     3The root of the cause here seems to be the custom user model, which as the docs mention in several places, must be made in the first migration of an app. `makemigrations` is meant to have code that ignores this style of error while it's running (as you may be doing it to make the migration for said user model), but I've not seen that result in this error before - usually it manifested as a graph dependency error (as the migration depending on the custom user model would have a dependency on an app with no migrations).
    44
    55I'm not sure what the Django bug is here - the original reporter or people more familiar with CMS will need to dig down and either give a simple project that can reproduce the issue (i.e. one without the intricacies of CMS) or work out why it's trying to import PageUser when, as far as the CMS bug says, it's someone's own personal custom user.
Back to Top