Development discussion. Logged to https://ddnet.tw/irclogs/ Connected with DDNet's IRC channel, Matrix room and GitHub repositories — IRC: #ddnet on Quakenet | Matrix: #ddnet-developer:matrix.org GitHub: https://github.com/ddnet
Between 2018-10-24 00:00:00Z and 2018-10-25 00:00:00Z
ofc all scripting languages are way slower than "real" languages like c/c++ but the question is whether it actually affects the execution of the program
well the mod doesnt matter, it's more of a "how would the client know what to calculate without making it script too"
probably heinrich thinks more like the gamecore changes only velocity etc.
and these values are then interpolated/handled like now
so basically only a small physic change
that doesnt create alot of possibilities anyway.
Still that would effect execution, bcs lua is easily 30-40 times slower than c(edited)
oh well in that case yeah i dont think replacing the core game elements with lua would help .. for any reason.. and it would also likely make it slower yes
a vm would deffs fit better, but even that is still 2x slower probs.
It's actually a matter of, if it's worth it, but e.g. as i said a switch from the server, telling the client to use a scrippted core or smth. would give more controll to a possible mod, so not that bad.
Still teeworlds deffs has other problems than supporting gamechanging mods xD
theres already alot of stuff possible with default client in the way of modding, so ye i think for now people can download custom clients if they want huge changes to a mod like the minetee one
well true, if ddnet makes another tab for mods that support this protocol, defenitelly possible i'd say, but this goes too deep in the direction of ddnet admins, they can explain you why probably(edited)
ddnet already has its own master server tho lol, they even have a whole separate section for only ddnet servers so it wouldnt seem so strange to me if they added "servers with http master support" to that tab or something
well theoretically you won't be much better between 100hz and 200hz in a game like teeworlds, since its only 50 ticks anyway.
tho i think our brain is much better, when it has realistic movements, independend of the reaction it could make
the way ive heard it explained is that higher fps makes it more likely that the game has a new frame drawn by the time the monitor refreshes because frames arent generated at a consistent rate
07:43
so even if it only refreshes/ticks 50-60x/sec having higher fps can still make a difference
yes, in the end we could basically discuss, can you feel the difference of 1ms.. and then i would say, we might be able to see it indeed, but that probably isn't game breaking for anyone. actually an interesting topic, bcs ppl always say "we can't do this and that", and 10 years later they are like. dude who says 1ms is no difference xD
Pretty late but I'm sure you can get away with having a huge portion of the game loop in lua. Love2D works pretty well and if you ever encounter a performance issue (interpolations) you can offload it to native code with a bridge like React Native has. A VM running optimized bytecode I doubt you could tell apart from benchmarking it.
11:33
Native/scripted hybrid solutions are getting more and more common nowadays and a lot of the performance issues with them are getting solved.
11:34
Regarding improving anti ping, can anyone even read the anti ping code?
@heinrich5991 a real time application where performance matters.. Like in games, there is no argument for performance of script languages vs compiled stuff doesn't matter. Sure it seems like nobody needs his 6k fps in teeworlds, but in the end partially implementing scripts will easily kill 1k of that frames. Script languages are good when they have a single task, because ez to use and debug
It makes more sense to specifically activate telegun on tiles than to specifically deactivate it.
As a player it's easier to find flashy spots than to find plain ones
The common usage of the feature seem to be parts that work by teleporting to specific spots. To cheat-proof them, everything else needs to be blocked off. This requires way more tiles than activating those specific spots.
Same goes for parts that work by teleporting to other tees or the air (with laser)
@heinrich5991 I've already played it, I don't know why they think it's a simulation game
20:15
i would say event based scripting is ok, like you finish a quest and some script does some things.
but i wouldn't add it to the physics in an action game
20:15
i doubt factorio needs as good fps as tw or csgo(edited)
20:16
if you like to add it to any game do as you want. i just wouldn't like to see it in teeworlds
20:16
factorio = ❤
20:16
+ mods
20:16
= <3<3<3
20:16
@Deleted User u nub u play solo
20:16
hahaha u have no friends
20:16
xD
20:17
I HAve fRienDs
20:17
explain me the word FRIENDS
20:17
!!
20:17
caughed
20:17
20:17
this is a friend
20:17
xD
20:18
he said me he isnt ur friend
20:18
let me ASK HIM, cuz of u I'll have to annoy him while he's coding(edited)
Currently you will always be ported to the middle (?) of the tee, no matter where you hit it. Imo this behaviour doesn't feel right, especially since it shoots you out of the tee. Instead it should be the same as with tiles, i.e. teleport to where you hit the tee.
I don't know if Pharice 2 uses the current behaviour, might break the map.
This would make the feature more dynamic and probably enable cool new parts, especially for solo maps. Not sure what the best implementation would be tho.
Currently it seems like it doesn't teleport if it hit a telegun-blocker during its lifetime. Imo it should always teleport as long as the last tile it hit was a non-blocked one/telegun-activator.