This feature involved a lot of refactoring in order to get working
correctly. I wanted to continue using the timeline view for chats thus I
used alternative styling & structure for DM event kinds. This worked
create since the elements map does not care.
There is some queing that has to be done to decrypt message content thus
I allow viewing messages even if they haven't been decrypted yet. I
think this is good for transparency since you understand what is and is
not decrypted. I do think that the UX could improve, because even tho it
is fast, it's flashes on new messages.
I did not implement saving of latest messages. I will do this later, but
this feature is big enough to merge as is: an alpha state that works.
I further abstracted profile & name updating to work in a more global
manner. Additionally I rewrote code that had attribute scripts to use
addEventListener instead. This is needed to happen anyways for security
and made the codebase easier to manage.
Improved recognization of replying to thread and being able to open it.
Rewrote profile storing on the model. Additionally fixed issues where
the profile was not getting loaded for referenced pubkeys on an event.
This changes to always on explore (though this may change yet again).
This is so we always get new events and we simply filter out who we want
to see. This also replaces the recent history.
I also changed contacts to check if we have contacts from storage and if
not to fetch them. This solves login loading issues.
There are still issues (IMO) with profile loading and thread loading
that need investigation. Profile loading just seems slow and threads
seems to not have posts. I think this may be because the relay is not
sending them to me because it's busy with the explore subscription. The
other relays seem fine, but Damus relay seems slow or unresponsive as I
will send a REQ and it will go completely unanswered, yet return explore
events.
I did this because they are 2 separate concerns and I have diverged
so far from the original repo. I didn't start a new repo because it
contains valuable history of all the people who contributed to it.
Additionally if there is an issue with this we can revert, but I
rather not. Forks are forks.
