Fix total normal impulse

[erincatto]

Improved usefulness of total normal impulse on contact points
Distance joint now solves motor before limit
Code clean up and minor fixes
Logging
Update clang format
Dynamic tree unit test
b2FreeFcn now receives the allocation size
Added check for a missed call to b2Body_ApplyMassFromShapes
Calling b2CreateCapsuleShape with a very short capsule now returns b2_nullShapeId
Calling b2Shape_SetCapsule with a very short capsule now returns with no change.
Added b2ContactId to hit events.
b2Body_SetTargetTransform now has a wake option.

#1019
#1015
#1014
#1010
#1009
#1005

[Copilot]

Pull request overview

This pull request fixes the total normal impulse calculation in contact points, reorders distance joint constraint solving to solve motor before limit, adds logging functionality, updates clang-format configuration, and introduces comprehensive dynamic tree unit tests. The changes also include code cleanup, minor API updates, and fixes for memory allocation and mass computation.

Key Changes

• Contact Solver Fix: Corrected totalNormalImpulse to accumulate delta impulses instead of absolute impulses, making it more useful for tracking actual contact interactions
• Distance Joint: Motor constraint now solves before limit constraint for better stability
• Dynamic Tree Testing: Added comprehensive unit test suite with 462 lines covering tree operations, ray casting, queries, and movement

Reviewed changes

Copilot reviewed 37 out of 37 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
test/test_dynamic_tree.c	New comprehensive unit test suite for dynamic tree with ray casting, queries, and stress tests
test/main.c, test/CMakeLists.txt	Registration of new dynamic tree test
src/contact_solver.c	Fixed totalNormalImpulse to track delta impulse instead of newImpulse across warm start, solve, and restitution phases
src/distance_joint.c	Reordered motor constraint to execute before limit constraints; removed unused include
src/weld_joint.c	Experimental changes with commented spring code and hardcoded bias values
src/shape.c	Added b2_dirtyMass flag support and capsule validation; added b2Chain_GetSurfaceMaterialCount
src/body.c	Fixed missing shapeIndex increment; added b2_dirtyMass flag clearing; added wake parameter to SetTargetTransform
src/solver.c	Added assertion for b2_dirtyMass flag; added contactId to hit events
src/table.c	Fixed to use set.capacity instead of capacity parameter
src/core.c, src/core.h	Added logging infrastructure with b2Log function and customizable log handler
src/dynamic_tree.c	Code formatting fixes and minor comment corrections
src/constraint_graph.c	Simplified dynamic-dynamic body color assignment logic
samples/*.cpp	Updated all SetTargetTransform calls with wake parameter; added new CircleImpulse sample to demonstrate impulse tracking
include/box2d/*.h	Updated API documentation, added wake parameter, added contactId to hit events, improved bullet documentation
.clang-format	Updated configuration for VS 2022 with new spacing rules

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[Copilot]

Pull request overview

Copilot reviewed 35 out of 35 changed files in this pull request and generated 2 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The code adds new hardcoded test points that overwrite the previous values but doesn't remove or comment out the old initialization. This creates confusion about which values are actually used. Consider either removing the old initialization (lines 3240-3246) or adding a comment explaining why both sets of values are present.

[Copilot]

Adding the check "m_simplexCount > 0" before using m_simplexIndex is a good defensive programming practice that prevents potential issues when the slider is accessed with an empty simplex. However, this guards against a state that might indicate a logic error elsewhere. Consider whether this is masking an underlying issue or if an empty simplex is a valid state that should be handled gracefully.