The June has happened, and with it groundwork for threadview took place. The progress achieved makes it now possible to read content posted on forums. The next steps are going to be improvement of Django template for posts list, "goto" views that handle redirections to specific posts (eg. "first unread", "first unapproved", "last in thread"), once those land, I'll be moving on to js UI for thread view, posts API and then, finally, posting content.
In meantime I've decided to stop for a while and look back to think trough the issues in codebase and ways in which those negatively affect the development.
The good news is most of those can be delayed to after first alpha and refactored one after another in small steps, saving Misago from another costly and harmful rewrite.
I like to think of this as sign of me progressing as software developer, but sadly there's one thing that'll take some effort to fix:
The way Misago's JavaScript is being done now makes its extremly costful and fragile to test currently. In fact this is so bad, I'm currently not writing any integration and acceptance tests for React.js components. The time expense to do this is simply unacceptable to me currently.
Luckily this may not be a tragedy as much as it may appear at first, and those tests may be instead done in Django using LiveServerTestCase, ran against headless Chrome, FF and maybe even Safari or Edge, and additionally used to test not only JavaScript, but also it's workings with backend. Still, this too has been delayed to after first alpha, and holds lower priority than backend cleanups.
From other news, Misago is now using JSONField that was introduced by Django 1.9 and is supported in PostgreSQL since 9.4. This field allows Misago to support some edge cases that don't fit into RDBMS relational model nicely, like storing object's unstructured metadata. This makes features like custom admin-defined user fields or storing small amounts of extension-specific data on other models easier.