2,524 Commits over 639 Days - 0.16cph!
Update: turn GamePhysics.DefaultMaxResultsPerQuery into a server var
- breaking API changes for GamePhysics.CheckSpheres, CheckCapsules, EnvironmentManager.Get
- Ran codegen
Tests: none, trivial change
▅▇▍▍█▋▊ ▌▅▍▊ ▇▍█▉█▉▇▄▇█▄▉▄▋▇▅▋▆▅▍▆▌▊▌▊▄▅▊▉▄ ▇▊▆▅▆▅ ▇▌▉ ▇▌▇▅▅▄ ▋▌▅▄▉▍▉▍ ▍▇ █▇▉▍▇▇ ▊█▇▄▇ ▆▍▉▆ ▌▌▋▌▅▍▋ ▉▋▌▅▄▊▄▄▌▋▆▌▋▌▋▌▅▍▆▌▋▄ ▊█▍▊▌▊▄▌▄ ▌▇▋▌▌▅ ▅▇▍█ ▆ ▆▊▄▄▋▇ ▅▄▊▍▍▅▍█▅ ▇▆▇▌█ ▍▍ ▊██▄▄ ▍▋ ▍▉▆ █▉▋▍ █▌ ▍██ ▄▅▅▍ ▆ ▌▇▊▋ ▍▍▅
▉▄▆▋▍▌▆▇ ▍▉▌▉▆▍▌▋▅ ▊▄▍▇▋▇▋ ▊▋▇▅▍▇▊▄ █▉█▉▅▉ ▉▄▆ ▅▄▅▊▇▌▆▋ ▇▅█▊▍▍▋▊█▄▆█▅ ▇▊▍▌▅▊█▆ ▉▊▅▄▋ ▆▋▆▄▆▍▄▇▊▅▆█▉▊▋▍▋█▆▉▅▋█▅▋ ▆▅▌█▋▉▆▉▆▊▄██▋▇█▋▋█▋█▉ ▍▋▇▌▅ ▌▉▉█▆▄▆▄▋▅ ▍▌▊▍▌▇▍▊█▆▊▍▄▇ ▄▍█▇▊▅ ▅▄▍▇ ▇ ▄▊▆▋▇▅ ▋▉▆▌▌▌▋▉▆ ▍▄▌▉█ ▍█ ▆█▆▊▍ ▇▉ ▄▌▍ ▆▊▋▋ ▋▌ ▅▅▍ █▋▉▅ ▅ ▊▋█▅ ▅▆▇
Merge: from hascloseconnections_fix
- Bugfix for BaseNetworkable.HasCloseConnections and GetCloseConnections mot seeing outside of small layer
Tests: ran unit tests
Bugfix: ensure players are detected on the border of the network cell
- expanded unit tests to cover these cases
Tests: ran unit tests
Bugfix: fix HasConnectionsClose & GetCloseConnections missing players in medium and large ranges
- also add a unit test to validate Connections overload
Tests: ran unit tests
Bugfix(tests): adjust player spawn distance for TestNetworkRange.OutsideRange cases
They can land on a boundary, making it more difficult to reason when working across small/medium/large layers
Tests: ran the tests, same results
Bugfix(tests): expand TestServer's network grid from 1k to 2k
Previously it was too easy to accidentally setup an out-of-bounds scenario, leading to confusing test failures
Tests: ran related tests, discovered couple tests are failing (unrelated to this change)
New(tests): tests for BaseNetworkable.HasCloseConnections and GetCloseConnections
- named cell sizes for various layers in NetworkVisibilityGrid
Tests: ran tests, failing where expected
Merge: from serverprofiler_recordscope_pause
- Update: ServerProfiler recorder scopes can be paused/resumed (but needs more work on export side, looks bad)
- Bugfixes for recorder scopes corrupting memory and breaking perfsnapshot
Tests: unit tests + recorded a multiframe coro with recorder scope + perfsnapshots
Update: add Pause/Resume to ScopeRecorder and SlowScopeRecorder
Tests: tried them in throw-away code, works but exported result looks wonky and needs more work
Update: ServerProfiler - add ability to pause & resume active recorders
- added couple unit tests
- release binaries built from 61cba2fc
Tests: ran unit tests
Bugfix: Simplify RecordContext memory management in ServerProfiler.Core
- release binaries built from 39f0cfa4
Worst bug was that 2 threads sharing one recording buffer
Tests: recorded a bunch of test coroutines (2 frames worth each) via RecordScope, then ran a perfsnapshot for 10 frames - no issues.
Update: ServerProfiler.Init now resets managed internal state
- also added handle cleanup to ScopeRecorder and SlowScopeRecorder dispose, in case double-dispose gets called
This reduces test boilerplate slightly (as scope recorders have a timeout)
Tests: ran unit tests - they pass. still chasing native storage corruption
Bugfix: when exporting a snapshot from recorder scopes skip fast-forward logic to 0 callstack depth
Tests: recorded coroutine, was able to see all 3 log calls (before yield, after 1st yield and before yield break). But something's corrupting memory for subsequent perfsnapshots
Merge: from serverprofiler_recordscope_pause
Need it for experiments with profiling coroutines
Bugfix: prevent taking perfsnapshots if a recorder scope is currently active and vice versa
- also fix recorder triggering NREs because it tries to run perfsnapshot code
Tests: ran profiler scope spanning multiple frames. Checked output, it's not making sense, investigating deeper
Bugfix: don't overwrite callstack depths for frame 0 with values for other frames when multiple frames are found
Tests: ran Export2FramesTorn - it passes and looks correct-ish (same wrong offset as in Export2Frames)
Bugfix(tests): fix invalid test logic in ExportExtraEnd2Frames
- renamed ExportExtraEnd2Frames to Export2FramesTorn
- added Export2Frames (shows invalid frame start - this is new)
Tests: ran unit tests, failures where expected
Merge: from networkgrid_iterationspeed_changes
- Optim: lazy allocate internals of NetworkGroup to save memory (adam.k's findings)
Tests: flew around on craggy, opened deep sea and entered-left, printed entity counts of deep sea, shot scientists in the distance
Update(tests): add ExportExtraEnd2Frames
Test passes but looks incorrect
Tests: ran unit tests
Bugfix: potentially support exporting profile of Recorder started inside a method and finished outside
Need to write more tests to validate it works across frames - not confident about it.
Tests: ran ExportExtraEnd - passes. ran all others - still valid
Update(tests): emit valid timestamps when running profile exporter tests
Tests: confirmed seconds-scale ranges in perfetto
Bugfix: handle cases where main thread doesn't have any profiling marks
Veeeeeeery unlikely event to happen, but should be supported
Tests: ran ExportEmptyMainThread test - it now passes
New(tests): ProfileExporter unit tests
Synthesizes basic profile streams and sees if it sucesfully exports them, without checking contents.
Tests: ran tests, 2 fail. inspected some outputs of passing ones
Bugfix: protect network group networkables access since it's lazy initted now
Tests: opened deepsea on craggy, ran `deepsea.printentitycount`, flew around craggy and open waters, entered and left deep sea, killed scientists from far away
Merge: from main
Will need to subtract some stuffs
Merge: from networkgrid_iterationspeed_changes
- Bugfix: handle more cases of lazy-initted net group subscribers instead of NREing
Tests: none, trivial changes
Bugfix: more null checks around lazy created subscribers
- utilities in BaseNetworkable
- BaseEntity.SignalBroadcast (in case npc shoots outside of players range)
- TreeManager.OnTreeSpawned (in case they spawn on an empty server)
Tests: none, trivial changes
Merge: from networkgrid_iterationspeed_changes
- Bugfix: avoid NREs from accessing uninitialized subscribers for net groups
Tests: booted into craggy with server occlusion and jobs 0, 1, 2, 3 modes
Bugfix: avoid NREs by skipping not yet initialized subscribers for the network group
Tests: booted into craggy with server occlusion and jobs 0, 1, 2, 3 modes
Merge: from serverprofiler_filters_april2026
- Update: remove a couple assemblies and utils from being recorded in snapshots to reduce overhead
Tests: exported snapshot in Jobs 3 mode and inspected
Update: april batch of ServerProfile filters
- Release binaries built from 8643e0d9
Excludes: websocket-sharp and UniTask assemblies, Facepunch.System's Deque and TimeSpanExt classes
Tests: inspected snapshot from editor in jobs 3 mode
Merge: from connections_pool_leak_fix
- Bugfix for leaking List<Connections>
Tests: ran around on craggy and checked print_memory showed low max value
Bugfix: return List<Connection> to pool in SendNetworkUpdate_Position
Submit on behalf of cipeaX, it's his find
Tests: ran around on craggy, checked pool.print_memory - max stayed low
Merge: from projectileweaponmod_allocs
- Optim: get rid of missed alloc in ProjectileWeaponMod aggregation calls
Tests: inspected snapshot taken while shooting
Optim: avoid IEnumerable Enumerator allocation
Not as generic, but saves an alloc
Tests: took a snapshot while shooting
Merge: from buildingprivlidge_invoke_allocs
- Optim: BuildingPrivlidge.AddDelayedUpdate caches it's invoke
Tests: none, trivial change
Optim: cache invoking callback in BuildingPrivlidge.AddDelayedUpdate
Tests: none, trivial change
Remerge of
147094 - Merge: from projectileweaponmod_allocs
Merge: from projectileweaponmod_allocs
- Optim: removed allocations related to ProjectileWeaponMod.Sum/Mult/Min/Max/Avg and their callers
Tests: used a modded AK to kill a scientist
Clean: ProjectileWeaponMod.Mult - use 1 as hardcoded default value
They've been manually specified as 1 everywhere, seems a bit redundant
Tests: compiles
Optim: cache all selectors that are used in mod value aggregations
Tests: got shot by a scientist. killed them with a fully modded ak
Optim: reimplement projectile mod aggregators to be alloc free
- added outstanding Avg, Min and Max helpers
- all helpers are now generic, to support the 2 overloads we have
- ripped out GetMods to avoid accidental reintroduction of allocs
Still need to fix remaining dynamic callback creation
Tests: got shot by a scientist
