-Create common methods create_segment_self and create_segment_target.
These reduce the amount of copy-paste boilerplate code since the IPC queueing stuff basically never changes
as far as I can tell. Now we simply specify the opcode and the data that goes with it, and off we go.
Create_segment_self retains the current behaviour of using the player's actor id as both the source and target,
and create_segment_target allows us to change the source and target in case commands are written that require
one or the other to be different (Sending targeted players to other coords/zones? Bringing a player to the GM/source user? Several possibilities.).
I originally did this thinking I would engineer a user management
interface afterwards, but this hasn't materialized yet and just makes
testing Kawari harder.
This should stop having to manually edit your Kawari World database to
gain access to GM and debug commands.
I don't know why I created a new message kind just for spawning NPCs
when ActorSpawn does this just fine. But what this actually means is
that the debug NPCs can now be spawned more than once.
It's a command that moves the user forward, backward, up, or down a specified amount of yalms.
It's quite helpful for exploration convenience, and doesn't replace setpos, rather, it's meant to complement it.
The set_position function was also extended to allow for lua code to specify the rotation/direction facing.
Instead of everyone starting out as a GM, they are a normal user by
default. Right now the only way to set GM rank is through editing the
database. The online status is set accordingly.
This doesn't work 100% reliably yet, but the world server should now try
to commit your player back to the database if it detects you disconnect.
I also fixed a mistake where the global server state never removed
clients when they errored out. There's now code to remove your actor
from the instance when disconnecting, but this doesn't work reliably yet
either.
This makes the Teleport action functional now, although it looks pretty
rough as it warps you instantly instead of waiting for the action to
actually finish.
This updates three major crates we use to the latest version with all of
their API changes, and the others too. Notably, this removes the Syn v1
dependency thanks to binrw 0.15 switching to Syn v2.
This begins figuring out how we are going to be propagating actor
control state: e.g. targets, poses, and other misc effects. I ended up
sending client triggers to the global server state, who then creates the
needed actor control packet for the other players.
Now players can see what other players are targeting!
There was two problems I was running into:
1. The move packet changed slightly, using a different way to encode the
rotation. In hindsight, it should seem obvious they would encode it the
same way for CommonSpawn and ActorMove.
2. write_quantized_rotation did it's order of operations wrong, and
it would spit out nonsensical rotations.
Both issues are fixed and you can see the other player's rotation
correctly now!
This is super simple, you get the player and the args (which you can
choose to split if you so choose.) I reimplemented the !setpos, and will
start doing the rest of the commands.
This means the three ARR inns are now "functional" (as in, you can enter
and exit them.) I also added some code to give you control of your
character in case you hit an unimplemented event.
It turns out that most warps are just a yes/no, and can be handled by
one single script. I scripted the lift attendants from Bulwark Hall <->
Airship Landing <-> Drowning Wench this way.
The inn bed is the most complex script we have written so far, and needs
a little bit more logic to when it finishes. I also made it play the bed
animation, although you *will* sleep forever as we don't know how to
disconnect you yet.
This kinda works, you can now talk to them without locking up your
character but you aren't warped yet. I also need to clean up the mess of
the Lua API.
