* [requires.io] dependency update
* Server starts...
Various things are broken, but it runs!
* [requires.io] dependency update
* [requires.io] dependency update
* [requires.io] dependency update
* FIX: Broken migrations
* FIX: Update auth framework
* FIX: Correct static use in templates
* FIX: Fix supplier sort
* FIX: Remaining tests
* Revert "Disable password reset as temporary fix to vulnerability (#396)"
This reverts commit e0c6a56263.
# Conflicts:
# RIGS/urls.py
* FIX: Fix broken newlining in PDFs
Introduced by a change in Django 2.1 'HTML rendered by form widgets no longer includes a closing slash on void elements, e.g. <br>. This is incompatible within XHTML, although some widgets already used aspects of HTML5 such as boolean attributes.'
* FIX: Fix some Django4 deprecation warnings
Why not...
* Refactor dependency file
Should now only include dependencies we actually use, not dependencies of dependencies and unused things
* Add newlines to the paperwork print test event
This will catch the error encountered in 79ec9214f9
* Swap to pycodestyle rather than pep8 in Travis
And eliminate W605 errors
* Bit too heavy handed with the dep purge there...
* Whoops, helps if one installs pycodestyle...
* FIX: Re-add overridden login view
* Better fix for previous commit
* FIX: Bloody smartquotes
Co-authored-by: requires.io <support@requires.io>
* FEAT: Initial work on revision history for assets
The revision history for individual items mostly works, though it shows database ID where it should show asset ID. Recent changes feed isn't yet done.
* FEAT: Initial implementation of asset activity stream
* CHORE: Fix pep8
* FIX: Asset history table 'branding'
* FIX: Individual asset version history is now correctly filtered
* FEAT: Make revision history for suppliers accessible
* CHORE: *sings* And a pep8 in a broken tree...
* Refactored out duplicated code from `AssetVersionHistory
* CHORE: pep8
And another random bit of wierd whitespace I found
Co-authored-by: Matthew Smith <mattysmith22@googlemail.com>
Closes#358
Takes inspiration from, but does not use, django-reversion-compare. We do a lot of RIGS-specific stuff that requires a lot of hacking to get working nicely with django-reversion-compare. The main example of this is event-item “many-to-one” fields. The performance difference of my code compared to django-reversion-compare was found to be negligible.
