Now that everything is a <li> can set the margin-top value to make a gap, instead of making a visible "invisible" hr. And with the template thread.html fixed, it seems to be working, showing everything as a tree, instead of a flat list.
Oh for Pete's sake I'm an idiot. The descendants array isn't going to contain the root post at all, because it's totally separate. That's the missing post everything was trying to reply to!
Right, I forgot I was in the middle of changing the name post to toot. Thus explaining why op.post had an empty result, since op.toot was what I changed.
We have the post ID, so why would Pleroma not send us the post? Request any missing posts it hasn't sent us, to try and get a thread that doesn't lose all its replies.
1) initialize seen to empty
2) get all posts not in seen
3) add posts to seen because the generator was iterated over lazily
4) wonder why it looks like there weren't any posts in seen
yeah... leftovers has to be lazy too.
I couldn't tell who was replying to whom, so I thought I'd give a shot at implementing a tree of threads, instead of a flattened list, with only the root distinguishable from the rest.
Because when you add one field to a class, you clearly want to add two fields to a class, and there's no need to consider the field you never asked to add to be something other than a field.
No need to list the preference fields three times in three places in the code. A class decorator ought to be able to sleuth them out from the model itself. Should make it easier to add new preferences.
Recreate the mastodon object if there's an error complaining about a missing feature set. Only happens for status_post I think. Could be further generalized...
Pleroma will send records that cause brutaldon to make pleroma-specific responses, which the mastodon python module wigs out on claiming it doesn't support that "feature set" so allow for a feature set to be specified...
commit 0a80206abb8fae7785a59aab88043b2b1974756b
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Tue Nov 5 19:22:00 2019 -0500
Fix oxford comma in bundled notifications, remove unused dependency
commit e96bd22bdce996734aaaf1d5625e08add3c8fcf7
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Tue Nov 5 19:19:42 2019 -0500
Now template works with bundled or un-bundled notifications
commit 6f46bef7fdd0defe2f02e09e28558de882ce4456
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Tue Nov 5 19:02:51 2019 -0500
Bundled toots work; now fix unbundled case
commit 07d9de49f943d019d04a5a5203081e57dc0741d8
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Tue Nov 5 14:09:14 2019 -0500
Notifications are now sorted by groups, but not collapsed
commit f62666929f12cf0c7db4c68a1468f7e138318a5c
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Tue Nov 5 13:58:41 2019 -0500
Fix saving of bundle_notifications setting
commit 335d5f985c968bb84e4b459dabf77d1d7ecad646
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Mon Nov 4 18:57:54 2019 -0500
Forgot to include migration for bundle notifications preference
commit 0e8232591c4f1bb972e9694433c546c9f66b5419
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Mon Nov 4 18:57:35 2019 -0500
Bundle notifications setting front-end
commit 6e945f1ceb2ff19470e164a946a6a48de4142812
Author: Jason McBrayer <jmcbray@carcosa.net>
Date: Mon Nov 4 18:54:49 2019 -0500
Backend code to group notifications
autocommit
Cy
GCU Prosthetic Conscience
Diana Thayer