calderpg-tri

@drake-jenkins-bot mac-monterey-clang-bazel-experimental-release please @drake-jenkins-bot linux-focal-clang-bazel-experimental-address-sanitizer please @drake-jenkins-bot linux-focal-clang-bazel-experimental-thread-sanitizer please @drake-jenkins-bot linux-focal-clang-bazel-experimental-undefined-behavior-sanitizer please @drake-jenkins-bot linux-focal-clang-bazel-experimental-valgrind-memcheck please @drake-jenkins-bot linux-focal-gcc-bazel-experimental-coverage please

@drake-jenkins-bot linux-focal-clang-bazel-experimental-address-sanitizer please @drake-jenkins-bot linux-focal-clang-bazel-experimental-thread-sanitizer please @drake-jenkins-bot linux-focal-clang-bazel-experimental-undefined-behavior-sanitizer please @drake-jenkins-bot linux-focal-clang-bazel-experimental-valgrind-memcheck please @drake-jenkins-bot linux-focal-gcc-bazel-experimental-coverage please

Are there parameters that couldn't be set through the existing [`CollisionChecker::PerformOperationAgainstAllModelContexts`](https://github.com/RobotLocomotion/drake/blob/master/planning/collision_checker.h#L278) mechanism?

More importantly, IRIS using a `CollisionChecker` provides for IRIS using a collision checker that is not `SceneGraphCollisionChecker` (e.g. Anzu's `VoxelizedEnvironmentCollisionChecker`), where the robot and/or environment geometry is not necessarily part...

Only in `SceneGraphCollisionChecker` does the context contain all geometries. For any other implementation of `CollisionChecker`, the robot and/or environment geometries being considered could be arbitrarily different from those in the...

> Or is that the wrong idea, and we just need to implement it from scratch as a nonlinear program that only ever does signed distance queries, with no convex...

Relates to #19680 - we have again been bitten by a deadlock resulting from logging redirection from bound multithreaded C++ code. While a few of us at TRI are familiar...

> Minor concern about removing //add_model/@name is that it forces users to use names like carrot_1, carrot_2 (numbering) or left::carrot, right::carrot (scoping). It may preclude users from using model names...

I don't think we need to prototype dmd changes in Anzu to see that loosing renaming has negative consequences. 10245 is enough evidence already.

Closing as serializing `ScalarSystemFunction` is not itself that important - serialization for functions provided in pydrake bindings can be handled via proper GIL handling instead. The cv-rather-than-polling-wait elements of this...