[NearCai]

Fixes #4391

Motivation

lane.getShape and junction.getShape only return 2D positions; elevation is
dropped even though the simulation model and the net file contain it.
junction.getPosition(..., includeZ=True) already exists, but there was no way
to retrieve the 3D shape of a lane or junction.

Implementation

• Add the POLYGON_3D (0x05) wire type and VAR_SHAPE3D (0x3f) variable.
• Write and read three doubles per point for VAR_SHAPE3D; keep VAR_SHAPE 2D.
• Support the new variable in libsumo and libtraci, including subscriptions.
• Add lane.getShape3D, junction.getShape3D and the convenience alias
  junction.getPosition3D to the Python client.
• Document the new protocol type and value retrieval APIs.
• Add a complex test with a net that has elevation.

Because the change only adds a new variable and does not alter existing
payloads, TRACI_VERSION is not bumped. Old clients keep getting 2D shapes;
old servers will report VAR_SHAPE3D as unknown instead of silently returning
2D data.

Verification

The new complex test tests/complex/traci/misc/shape3D checks the 2D and 3D
lane/junction shapes, the getPosition3D alias, and a VAR_SHAPE3D
subscription. Existing 2D retrieval (including simulation.getNetBoundary)
still returns 2-tuples.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota.

[behrisch]

@NearCai Thanks for that one! Can you try to make your commits always under your "real" email address? If the commit shows something like @users.noreply.github.com the eclipse validation will fail

[NearCai]

Thanks! I rewrote the commit author/committer address to my real email (2094657017@qq.com) and force-pushed the branch; the PR commit now uses it. I will use this address for all future commits.