Still pursuing the Topic tree fix on vbz, though I probably should not. I did identify a bunch of related code that's no longer being used, and have commented it out.
I guess I find it difficult to do a major organization task (figuring out what should go on the home page) while there are still substantially untidy bits elsewhere (the topic tree).
For some reason, it feels vitally important that the topic tree should show the number of titles available in twigs below it, rather than just the current twig (otherwise how do you know where to look for stuff?).
This despite the fact that untangling all the related code to make this happen has taken hours so far and looks like it will require more hours.
Yesterday: got the bare framework of a vbz homepage working. I stupidly put all the wiki/node logic into a dropin, though, so had to spend a lot of time separating it out so it can be used outside of admin, and it looks like two of the class-groups also need to be split into logic and UI.
Maybe I should just hard-code a home page for now.
PsyCrit on standby because the accelerating crisis of our money situation re-orders my priorities: vbz, FinFer, migration (incl tootcat)
PsyCrit: form layout fixed, now properly retrieves list of target articles; titles need to be cleaned for display, tho, and form-processing has not been checked at all.
Need to revamp NukePlus a bit to not depend on RecentChanges, so I can flush out old spam. Currently thinking it should look for edits by accounts that haven't been marked as "trusted".
To that end, I'm setting vbz up on the live server, finally, so I can continue development without worrying about making things worse.
Then maybe I'll kluge up a front page, and then maybe I'll see if I can push a test order all the way through and pretend-fill it. (Or maybe I'll order myself a t-shirt and *actually* fill it.)
And maybe here, on Mastodon, there will be some useful critiques/ideas from people who can really picture the fire I'm trying to start.
I feel like vbz.net has spent more than enough time in an interrupt state. It needs its own time-slice. It shouldn't have to keep waiting.
So I'll work on PsyCrit during quiet times when I can focus (it currently involves some rather tricky queries, reverse-engineered by examining the data because I can't see any other way to make this work), and during more active times I'll work on getting vbz WORKING. Not perfect, not unit-tested -- just usable.
Reverse-engineering SMW's v3 data schema. Can't find any documentation on it. There may be class wrappers that would make this unnecessary, but can't find anything likely-looking in the code docs.
Fortunately I had apparently spent some time on this earlier and made a little progress.
What mystifies me is that they seem to have dropped the idea of having an index for all the different field-types; you have to look through all of them now...
