[gpeairs]

This is the companion to #60 adding the Julia side of the wrapper

[SimonDanisch]

Let an LLM try this pr out: built the wrapper against patched Clipper2 2.0.1 per Yggdrasil #14270 — 112/112 tests pass, docs clean. Fuzzed ~1000 boolean ops and stress-tested GC in the result callbacks; no issues.

One defect: Clipper64() / ClipperOffset() don't check the NULL return documented at cclipper2.h:79 (engine.jl:71, offset.jl:23), so a failed create segfaults on first use instead of throwing.

One question: the *_z entry points have no Julia surface: only runtests.jl reaches them via raw ccall. Should those be wrapped?

[gpeairs]

Thanks for trying it out!

One defect: Clipper64() / ClipperOffset() don't check the NULL return documented at cclipper2.h:79 (engine.jl:71, offset.jl:23), so a failed create segfaults on first use instead of throwing.

Right, fixed.

One question: the *_z entry points have no Julia surface: only runtests.jl reaches them via raw ccall. Should those be wrapped?

Yeah, they should. Added Z versions of the types, dispatch on Z paths for the add_...! methods, and execute_polytree_z.

[gpeairs]

All the prereqs are merged/registered, and I pointed this at the new cclipper2_jll so it should be ready.

[SimonDanisch]

Great thank you!