[workspace] Fix OpenUSD plugin loading

[jwnimmer-tri]

Towards #20898.

Something is still weird with operator== stuff on the stage, but I'm hoping we can co-evolve that with real parser regression tests. This at least moves us closer.

---

This change is 

[jwnimmer-tri]

OMG I really hate custom smart pointers, but I think I finally got to the root cause and this should be ready to go now.

+@rpoyner-tri for feature review, please.

\CC @hong-nvidia FYI

[jwnimmer-tri]

Reviewable status: 1 unresolved discussion, LGTM missing from assignee rpoyner-tri(platform), needs at least two assigned reviewers

---

tools/workspace/openusd_internal/patches/dlopen_forbidden.patch line 1 at r2 (raw file):

WIP

Working

On the next push, I'll write some explanatory text here.

[jwnimmer-tri]

With fresh eyes this morning I noticed a few tweaks, and pushed some changes just now.

Reviewed 4 of 8 files at r1, 2 of 3 files at r2, 12 of 12 files at r3, 2 of 2 files at r4, all commit messages.
Reviewable status: LGTM missing from assignee rpoyner-tri(platform), needs at least two assigned reviewers

[rpoyner-tri]

Reviewed 4 of 8 files at r1, 2 of 3 files at r2, 12 of 12 files at r3, 2 of 2 files at r4, all commit messages.
Reviewable status: 1 unresolved discussion, needs at least two assigned reviewers

---

tools/workspace/openusd_internal/patches/weakptrfacade_cxx20.patch line 11 at r3 (raw file):

`this` and `p` were different, so the swapped ordering was formally
ambiguous. We can fix this by using matching the type of `this` and
`p`. There's no need to downcast to the derived type for any reason.

nit the grammar of this sentence lost me. Not sure what to suggest exactly.

Code quote:

ambiguous. We can fix this by using matching the type of `this` and
`p`. There's no need to downcast to the derived type for any reason.

[jwnimmer-tri]

+@sherm1 for platform review per schedule, please.

Reviewed 1 of 1 files at r5, all commit messages.
Reviewable status: LGTM missing from assignee sherm1(platform)

---

tools/workspace/openusd_internal/patches/weakptrfacade_cxx20.patch line 11 at r3 (raw file):

Previously, rpoyner-tri (Rick Poyner (rico)) wrote…

nit the grammar of this sentence lost me. Not sure what to suggest exactly.

I gave it another shot.

[rpoyner-tri]

Reviewed 1 of 1 files at r5.
Reviewable status: LGTM missing from assignee sherm1(platform)

[rpoyner-tri]

Reviewed all commit messages.
Reviewable status: LGTM missing from assignee sherm1(platform)

[sherm1]

Platform

Reviewed 3 of 8 files at r1, 1 of 3 files at r2, 12 of 12 files at r3, 1 of 2 files at r4, 1 of 1 files at r5, all commit messages.
Reviewable status: 1 unresolved discussion

---

multibody/parsing/test/usd_parser_test/box_plane.usda line 18 at r5 (raw file):

    }
    def Cube "boxActor" (
        prepend apiSchemas = ["PhysicsCollisionAPI", "PhysicsRigidBodyAPI", "PhysicsMassAPI"]

BTW consider wrapping these to 80 chars for easier reading

[jwnimmer-tri]

Reviewable status: complete! all discussions resolved, LGTM from assignees rpoyner-tri(platform),sherm1(platform)

---

multibody/parsing/test/usd_parser_test/box_plane.usda line 18 at r5 (raw file):

Previously, sherm1 (Michael Sherman) wrote…

BTW consider wrapping these to 80 chars for easier reading

This is an good thought, but I'm going to put it on the back burner for now. I think many times these files are saved by a GUI, not written by hand. As we work with more and more USD files, we'll need to come up with a policy one way or another.