Your vote requested for 2.0 feature: Clientside puffs enabled by default

General discussion of the port and Doom-related chat.

Should predicted puffs be on or off by default?

 
Total votes: 0

Uwe L. Bendoverson
 
Posts: 62
Joined: Wed Nov 12, 2014 12:32 am
Location: Testing Zan 2.0

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#21

Post by Uwe L. Bendoverson » Thu Nov 13, 2014 5:54 pm

Ijon Tichy wrote: Either way, the puff and the shrapnel are now in wildly different places, causing things to be worse off than if you simply had the puff spawn by the server's command.
I could see the shrapnel spawning late (if it's set up in such a way that only the server can control spawning of those), but I don't see how they'd be in wildly different places unless the puff moves around on it's own.
Watermelon wrote: As yet another addon that I did not mention (I'm sorry... argh), only bullet puffs with +NONETID are predicted. Anything else will not be, no matter what the setting (prediction on or off).
So that basically solved Ijon's concern.
Last edited by Uwe L. Bendoverson on Thu Nov 13, 2014 5:56 pm, edited 1 time in total.

Cruduxy
Zandrone
Posts: 1059
Joined: Fri Jun 08, 2012 4:24 pm

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#22

Post by Cruduxy » Sat Nov 15, 2014 11:15 pm

While at it you guys should enable autorun and mouselook by default!
[][][][][][][][][][][][][][][]
Nothing to see here
[][][][][][][][][][][][][][][]

Kara Kurt
Frequent Poster Miles card holder
Posts: 887
Joined: Sat Oct 12, 2013 6:58 pm
Location: Strasbourg, France
Contact:

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#23

Post by Kara Kurt » Sat Nov 15, 2014 11:25 pm

Cruduxy wrote: While at it you guys should enable autorun
Hell yeah, most newbies have no clue why everyone else moves faster with their "hey how you move this fast xD", at least with that there's a better chance they score a flag that they missed by an inch because too fucking god damn slow!

Uwe L. Bendoverson
 
Posts: 62
Joined: Wed Nov 12, 2014 12:32 am
Location: Testing Zan 2.0

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#24

Post by Uwe L. Bendoverson » Sat Nov 15, 2014 11:31 pm

Or have it so if someone doesn't run for nearly an entire map, it just says in bold letters on the screen: "THERE'S A SHIFT/CAPSLOCK KEY FOR A REASON!".

Just kidding... Or am I?

Ijon Tichy
Frequent Poster Miles card holder
Posts: 901
Joined: Mon Jun 04, 2012 5:07 am

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#25

Post by Ijon Tichy » Sun Nov 16, 2014 2:31 am

Uwe L. Bendoverson wrote:
Ijon Tichy wrote: Either way, the puff and the shrapnel are now in wildly different places, causing things to be worse off than if you simply had the puff spawn by the server's command.
I could see the shrapnel spawning late (if it's set up in such a way that only the server can control spawning of those), but I don't see how they'd be in wildly different places unless the puff moves around on it's own.
Hitscan check hits player on client and spawns on the player for the client, but misses the player on the server and so keeps going, presumably to hit a wall and spawn there for everybody else, including the server.

And yeah the NONETID thing should help, assuming BulletPuff subclasses don't have that flag (edit: by default). If they do, then things will still desync for the reasons discussed above.

edit2:
Uwe L. Bendoverson wrote:(if it's set up in such a way that only the server can control spawning of those)
The client never spawns server-side things of its own accord. If the client tried to spawn things, it'd cause horrible desyncs constantly, not just from simple latency issues, but from the fact that the client doesn't keep track of the server's RNG, is not guaranteed to know TIDs, is guaranteed to not know health values for anything not a player (and even then only sometimes), ACS variables, and most probably some other shit that escaped my mind.
Last edited by Ijon Tichy on Sun Nov 16, 2014 2:38 am, edited 1 time in total.

Uwe L. Bendoverson
 
Posts: 62
Joined: Wed Nov 12, 2014 12:32 am
Location: Testing Zan 2.0

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#26

Post by Uwe L. Bendoverson » Thu Nov 20, 2014 7:20 pm

Ijon Tichy wrote:
The client never spawns server-side things of its own accord. If the client tried to spawn things, it'd cause horrible desyncs constantly, not just from simple latency issues, but from the fact that the client doesn't keep track of the server's RNG, is not guaranteed to know TIDs, is guaranteed to not know health values for anything not a player (and even then only sometimes), ACS variables, and most probably some other shit that escaped my mind.
I'm aware. I just meant that if puffs are still spawned serverside, but predicted clientside, the clientside predicted puffs would not spawn the shrapnel, the shrapnel will spawn later when the server tells the client to spawn the shrapnel by the server's instance of the puffs.

I hope that makes sense.
Last edited by Uwe L. Bendoverson on Thu Nov 20, 2014 7:21 pm, edited 1 time in total.

Watermelon
Zandrone
Posts: 1244
Joined: Thu Jun 28, 2012 9:07 pm
Location: Rwanda

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#27

Post by Watermelon » Mon Nov 24, 2014 11:20 pm

Note that I would like to sync the RNG's between the client and server if using compat_oldrandom. That way predicted shots should still look okay minus touchy desyncs.
Last edited by Watermelon on Mon Nov 24, 2014 11:25 pm, edited 1 time in total.

User avatar
Torr Samaho
Lead Developer
Posts: 1543
Joined: Fri May 25, 2012 6:03 pm
Location: Germany

RE: Your vote requested for 2.0 feature: Clientside puffs enabled by default

#28

Post by Torr Samaho » Wed Nov 26, 2014 7:34 am

Watermelon wrote: Note that I would like to sync the RNG's between the client and server
This would be the only correct solution, but unfortunately is quite difficult to achieve with our C/S setup.
Watermelon wrote: if using compat_oldrandom
I don't think that this restriction would make the undertaking that much easier, but considerably less useful.

Post Reply