Notes |
|
(0002809)
|
Edward-san
|
2012-03-12 16:49
(edited on: 2012-03-12 16:51) |
|
It seems that the client thinks the elevator speed is lower than the right one. In fact, when I shoot with any hitscan weapon the puff's position is lower than yours. Also, when you wait a lot, at a certain point there's a high speed lowering down. The server's current elevator is near the stopping position, but the client thinks the elevator's current position is a lot higher.
This effect is visible with the second record demo. You have to wait a lot of time, I don't know how to measure the right time this happens.
|
|
|
|
Can somebody create a minimal example wad? |
|
|
|
Edward-san, can you reproduce the bug with my example WAD? |
|
|
|
Yes, just wait a bit and begin shooting with the pistol. At the beginning there's nothing notable, but wait a minute and refire again. The spawned projectile is positioned lower than the cross. |
|
|
(0003159)
|
TIHan
|
2012-04-07 00:24
(edited on: 2012-04-07 00:25) |
|
Me and Edward had a good long discussion about this, which I was able to reproduce the issue in two ways:
1. With unlagged off and with 500 ping - the ping determines the gap between where you fired and where it actually hits - with hitscan weapons.
2. With unlagged on, I start the floor moving down with 500 ping, everything feels good. If I bring the ping down to 0 while the floor is moving, you can clearly see the gap in hitscan and projectiles.
Notes: Apparently projectile weapons do have a gap, but it's very hard to notice without fluctuating the ping constantly. This is due to how latency works with floors; as we have no prediction of moving sectors and/or no unlagged on projectile weapons right when they fire (which we do not want that).
This needs to be thoroughly discussed and probably do something about it after 98e.
|
|
|
|
zandronum 1.0 is released, so I believe this should be fixed in 1.1. |
|