6,021 Commits over 974 Days - 0.26cph!
Initial attempt at a Train Ai that can patrol our existing rail network:
- Based off the backend systems for Boat AI
- Will reverse at dead ends
- Will only work if theres a driver in the train car
Add ability to send and load delta changes in the player metabolism system. Use a header mask bit to discern whats in the proto
Use HasGreatlyChanged() for wetness, calories and hydration
Use HasGreatlyChanged() for temp and comfort
Only set metabolism has changed provided there has been a slight variation rather than an exact one.
This prevents HasChanged() from firing if its like a 4th decimal place that changes
Set parent history projectile validation default to true
Fixed pooltable NRE'ing on spawn
Second set of cherrypicks
First set of cherry picks from game_room_dlc old because of plastics 'issues' up to
151245
50cal_animation_improvements -> main
Fix the camera animation going off in a random direction and causing a random pop - used jarryds method
Dont turn the yaw back to 0 when reloading 50cal - it's not needed
main -> 50cal_animation_improvements
deployable_snapping_settings_fix -> main
Fix diffferent snap modes not properly registering on the server
industrial_conveyor_network_optimize -> main
Unbusy has to call the full update because we are .5s later
- Prevent Unbusy from also calling a network update
- Prevent Invokes from allocating in client tweens and when unbusy is invoked
Stop industrial conveyors from sending 4 network updates rather than one every time the switch is set on or off (until flag changes are implemented)
network_analyser_rebuild -> main
Full remake of the network analyser UI.
Supports:
- looking at a delta (last x seconds)
- Pin/Compare snapshots side by side
- Anomaly Highlighting (needs more work), highlights rows that are particularly expensive considering our bandwidth
50cal_animation_improvements -> main
player_loot_mark_dirty_fix -> main
Remove repeated allocation in PlayerLoot::MarkDirty
Fix 'elbow-ness' a little on the front turret
New rear turret IK curves
Map hand ik to a curve instead of using raw values (just like the parachute)
- Good first pass curve on front
- Good first pass curve on rear
Map distance to turret during reload to a curve. Reduces snapping on the rear turret at the end of animations
- Good first pass curve on rear
Add some hand blend in at the end of reload animations. Not perfect but substantially better than what it was
main -> 50cal_animation_improvements
mountable_rollback -> main
Refactor tick parent history into TickHistory
Set parenthistory to false as default
Fix 50cal idle animations being set with ammo or no ammo incorrectly when sitting idle
Fix rear turret having incorrect animator time setting
main -> mountable_rollback
Use Jarryds hack to ennsure pt boat turret reload animation doesnt drift from the players reload animation (stupid unity issue)
steam_icons_upload_refactor_2 -> main
Update
145674 for the current codebase
Cleanup
Make sure its explicitly named/works with only parents
Parenting doesnt allow to mountables all the time so going down a pure parenting route catches the cases we need
main -> mountable_rollback
demo_breaking_fix -> main
Fix demos removing certain entities during initial playback
Was most apparent with global broadcast entities and would often result in Cargo Ship not being there during playback
- Add a serverside convar to be able to toggle it on/off
- Parameter-less constructor to Deque so it can be pooled
- Codegen
Initial attempt at mountable rollback to fix 'projectile invalid' when attacking a player thats on a mini or zipline etc.
We cheat by only storing player relative positions to avoid rolling back mountables physics positions
Each client will allocate and store the last 16 ticks. 192 bytes for each mounted player, which is reused via the pool.
Retain old behaviour if not mounted hit or if the last 16 ticks aren't available
demo_scrubbing_optimizations -> main
Dont allocate a new hashset everytime we jump to a new point on the timeline
Stop every demo scrub from allocating a new BaseNetworkable list
boomerang_model_dupe_fix -> main
