06d47040

By: ValentinHirschi <valentin.hirschi@gmail.com>

keep composed sampling maps and densities in native precision

Generalize existing map components, affine/conditional compositions, prepared
cut records, E-surface callbacks, eager/dual evaluation, bridge results and
partitions to native Double/Quad/Arb scalars. Reuse the safeguarded native root
solver and preserve typed numerical errors through inverse/density evaluation.
Invalid original inputs remain structural failures; representational failures
must retain support and allow a later native retry.

Production bindings still use Double at this checkpoint. Native cache binding,
original-source stability retry and signed-distance root certification remain
explicit next steps; this commit does not claim production map rescue or a
GL638 variance improvement. Keep the shared amplitude/cross-section plan and
baseline comparability notes current.

Validation: 122 focused native/physical-geometry tests; 162 broader sampling
tests; both API saved-state and summed/explicit-channel regressions; core/API
cargo check --tests and all-target clippy; cargo fmt and git diff checks.

User prompts:
Continue as planned, but re-read the plan, and establish how to best orchestrate progress on both front of amplitude surface advanced sampling, but also the overall goal of cross-section advanced sampling, that must eventually apply to GL638 and demonstrate improvement there.

(continue as planned, but remember that for being able to quickly generate states for testing, specifying ot one orientation may help. But ultimately, things must be tested end-to-end on all orientations)

Timeline