When resetting tick records each frame, a fairly large shared ptr destruction chain occured due to the final references to the various bits of context data (e.g. state machine scopes, mirror scopes) being removed.
This change switches individual shared ptrs to context data to unique ptrs, while allowing shared ownership (e.g. via the FAnimNotifyEventReference) via holding a shared ptr to the entire array.
#jira UE-130378 - Optimization pass on new UE5 anim node features
#preflight 61b35056c01c89f906e1a29b
#rb Timothy.Daoust
#ROBOMERGE-AUTHOR: thomas.sarkanen
#ROBOMERGE-SOURCE: CL 18430259 in //UE5/Release-5.0/... via CL 18435292
#ROBOMERGE-BOT: STARSHIP (Release-Engine-Staging -> Release-Engine-Test) (v897-18405271)
[CL 18435518 by thomas sarkanen in ue5-release-engine-test branch]
This represents UE4/Main @18073326, Release-5.0 @18081140 and Dev-PerfTest @18045971
[CL 18081471 by aurel cordonnier in ue5-release-engine-test branch]
Adds 'template' anim BP concept. These anim BPs have no TargetSkeleton and as such cannot have direct references to animations placed inside of their anim graphs
Adds function call support from anim nodes. All anim graph nodes can now call functions when initialized, updated, evaluated or when they first become relevant.
Relevancy is established using a new FAnimSubsystemInstance_NodeRelevancy which tracks nodes in a local map, per UAnimInstance, if nodes require it.
Functions are displayed on the node if bound, and in the details panel.
Added a new override point to FAnimSubsystemInstance to allow for WT init.
Moved FMemberReference customization into a public header so it can be used on anim node functions.
Wrapped functions up into FAnimNodeFunctionRef structure so they can be re-used more effectively. Converted CallFunction node to use them.
Added a couple of simple BP function libraries to demonstrate the use of the new FAnimNodeContext (this is intended to be a generic way of exposing FAnimNode_Base types to script without the node types themselves having to be known to script directly).
Added the ability to set exposed properties as 'always dynamic' so they appear in mutable data rather than being merged into constants. This allows for the anim node data API to be changed to be less strict (and more script friendly). Now values can be set in mutable data (via GET_MUTABLE_ANIM_NODE_DATA_PTR) and attempted sets to constant data can be ignored and reported to client code.
A few minor crash fixes with edge cases (inc when trying to open an asset editor fails because of a missing skeleton/cancellation).
Also fixes an issue where literal linked anim graph/control rig/custom poroperty node inputs didnt get copied
#rb Jurre.deBaare,Aaron.Cox
#ROBOMERGE-SOURCE: CL 16703644 in //UE5/Main/...
#ROBOMERGE-BOT: STARSHIP (Main -> Release-Engine-Test) (v835-16672529)
[CL 16703653 by thomas sarkanen in ue5-release-engine-test branch]
- Added animation notify mirroring functionality
- Extended IGraphMessage to allow users to create context data that is attached to notifies
- Notify, NotifyBegin, NotifyTick, and NotifyEnd methods have been extedned to take an FAnimNotifyEventReference that can provide generic data
- StateMachines can optionally attach active state information to notifies
- Added function libraries to query if a notify was triggered by a particular state or state machine
- Added support for detecting if a notify was generated by a mirrored animation
#github https://github.com/EpicGames/UnrealEngine/pull/7204 Fredrik Lindh
#rb thomas.sarkanen, aaron.cox
#jira UE-111797
[CL 16099692 by timothy daoust in ue5-main branch]
Per-node constant data is now held on a generated struct as part of sparse class data.
Per-node mutable data (i.e. pin links/property access mappings) is now held on a generated 'mutable data' struct that is compiled as part of the generated class.
The anim BP compiler is now extended more conventionally using UAnimBlueprintExtension, derived from UBlueprintExtension. This directly replaces the older 'compiler handler' pattern that was added in an emergency fashion for 4.26. Anim graph nodes now request their required extensions and these are held on the UAnimBlueprint in the UBlueprint::Extensions array. The Extensions array is potentially refreshed with any node addition or removal. The Extensions array is force-refreshed each time an anim BP is compiled for the first time to deal with newly added or removed requirements.
Const-corrected a bunch of UAnimInstance/FAnimInstanceProxy APIs that rely on (now truly) const data.
Added a split state/constant version of FInputScaleBiasClamp to allow some of its data to be split into constants.
Tweaked alignment/ordering of FPoseLinkBase to save a few bytes per pose link.
Deprecated FAnimNode_Base::OverrideAsset in favor of a more UAnimGraphNode_Base-based approach. Individual nodes can still have runtime overrides via specific accessors. The new approach will also give us the oppurtunity to override multiple assets per node if required in the future.
Moved property access into Engine module & removed event support from it - this was never used.
Reworked property access compilation API a little - construction/lifetime was a bit confusing previously.
Optimized path used to create UK2Node_StructMemberSet nodes in per-node custom events. When using mutable data, the structure used is large and very sparsely connected (i.e. only a few properties are written) so we only create pins that are actually going to be used, rather than creating all of them and conly connecting a few.
Patched the following nodes to use the new data approach:
- Asset players (sequences, blendspaces, aim offsets)
- Blend lists
- Ref poses
- Roots
#rb Jurre.deBaare, Martin.Wilson, Keith.Yerex
[CL 16090510 by Thomas Sarkanen in ue5-main branch]
Per-node constant data is now held on a generated struct as part of sparse class data.
Per-node mutable data (i.e. pin links/property access mappings) is now held on a generated 'mutable data' struct that is compiled as part of the generated class.
The anim BP compiler is now extended more conventionally using UAnimBlueprintExtension, derived from UBlueprintExtension. This directly replaces the older 'compiler handler' pattern that was added in an emergency fashion for 4.26. Anim graph nodes now request their required extensions and these are held on the UAnimBlueprint in the UBlueprint::Extensions array. The Extensions array is potentially refreshed with any node addition or removal. The Extensions array is force-refreshed each time an anim BP is compiled for the first time to deal with newly added or removed requirements.
Const-corrected a bunch of UAnimInstance/FAnimInstanceProxy APIs that rely on (now truly) const data.
Added a split state/constant version of FInputScaleBiasClamp to allow some of its data to be split into constants.
Tweaked alignment/ordering of FPoseLinkBase to save a few bytes per pose link.
Deprecated FAnimNode_Base::OverrideAsset in favor of a more UAnimGraphNode_Base-based approach. Individual nodes can still have runtime overrides via specific accessors. The new approach will also give us the oppurtunity to override multiple assets per node if required in the future.
Moved property access into Engine module & removed event support from it - this was never used.
Includes a thread-safety fix for 4.26 that hasnt made it over to 5.0 yet.
Reworked property access compilation API a little - construction/lifetime was a bit confusing previously.
Optimized path used to create UK2Node_StructMemberSet nodes in per-node custom events. When using mutable data, the structure used is large and very sparsely connected (i.e. only a few properties are written) so we only create pins that are actually going to be used, rather than creating all of them and conly connecting a few.
Patched the following nodes to use the new data approach:
- Asset players (sequences, blendspaces, aim offsets)
- Blend lists
- Ref poses
- Roots
#rb Jurre.deBaare, Martin.Wilson, Keith.Yerex
[CL 16071104 by Thomas Sarkanen in ue5-main branch]
Refactored tick record sync into utility structure - FAnimSync. This improves the odd API around tick records, better encapsulating functionality and leaving less for the caller to get wrong. This will also eventually allow this to be refactored out into a scriptable pipeline stage.
Added sync node and scoped sync message for new 'graph based sync'. Nodes that subscribe to graph-based-sync determine their sync group based on the scope that they are in.
Removed 4.26-style sync scopes - pushed all syncing up to the main anim instance. Linked anim instances no longer sync their own tick records.
To make graph based sync more useful, surfaced graph attributes and their visualizations as labels on pins and parallel wires to visualize flow.
This involves statically determining the attribute flow of the graph at compile time. Added a new compiler handler to deal with this new debug data.
Updated a lot of nodes to specify their attributes so graph flow can be correctly visualized.
Added tracing of attributes and sync records and visualization of traced records when debugging the anim graph.
#rb Jurre.deBaare, Martin.Wilson
[CL 14998555 by Thomas Sarkanen in ue5-main branch]
This takes the idea behind Braeden's node tracker/ancestor system and makes it a bit more generic, removing the bits of the API that were bound to FAnimNode_Base and allowing easier cross-plugin message dependencies.
The idea is that at any scope in the graph update you can add a TScopedGraphMessage with a specified payload (which can just be a node, but can also have state of its own), then further towards the leaves, that message can be detected and acted upon.
This changelist includes a re-working of the inter-node communication around inertial blending and cached poses as examples.
#rb Braeden.Shosa,Jurre.deBaare
#jira none
[CL 14640646 by Thomas Sarkanen in ue5-main branch]
Fixed issue with sync group collisions when syncing with 'component' scope
This change moves sync group management to use a name->group map instead of indices. Staying with indices cannot work across multiple instances as they are simply allocated at the point the single anim BP is compiled, so can easily collide in anything other than simple configurations.
Deprecated all index based access to sync groups in the runtime. Sync groups can still be enumerated on the generated class.
#jira UE-101030 - Sync groups with non-local scope can collide, causing incorrect behaviour
#rb Martin.Wilson,Jurre.deBaare
[CL 14521985 by Thomas Sarkanen in ue5-main branch]
+ Added SetSequenceLength
+ Moved GetPlayLength virtaul from AnimSequenceBase to AnimationAsset and deprecated GetMaxCurrentTime (only used in a single place)
#rb Martin.Wilson
[CL 14298631 by Jurre deBaare in ue5-main branch]
Initial basic implementation, more features to follow.
Accessed via the context menu of a graph track.
Shows anim graph nodes in a linearized tree similar to ShowDebug ANIMATION.
Select a node to 'pin' all the properties that it traced (multi select supported) to the bottom of the window.
Scrub the timeline to see those properties change over time.
Asset references use clickable hyperlinks to go to the specified asset editor.
#rb Jurre.deBaare
[CL 11041140 by Thomas Sarkanen in Dev-Core branch]
#rnx
#rb none
#ROBOMERGE-OWNER: ryan.durand
#ROBOMERGE-AUTHOR: ryan.durand
#ROBOMERGE-SOURCE: CL 10869210 via CL 10869511 via CL 10869900
#ROBOMERGE-BOT: (v613-10869866)
[CL 10870549 by ryan durand in Main branch]
Merging using Fortnite_Main<->Fortnite_ControlRig
#rb none
#ROBOMERGE-OWNER: helge.mathee
#ROBOMERGE-AUTHOR: helge.mathee
#ROBOMERGE-SOURCE: CL 7321580 via CL 7323175
#ROBOMERGE-BOT: (v371-7306989)
[CL 7323351 by helge mathee in Main branch]
