chore: reduce walls of comments

[gmelodie]

Summary

The FFI layer had accumulated large comment walls — multi-paragraph module headers, multi-line # rationale blocks, and long per-proc/field docstrings — that mostly restated what the code already says. This trims them to the project's comment budget: no multi-line non-doc comments (a single # line, or self-explanatory code), and docstrings capped at 3 lines (usually 1). Load-bearing WHY notes (malloc/free-vs-ORC-TLS thread hazards, single-fire callback, ABI wire-format seams, deadlock/reentry rationale) are kept — compressed to one line each rather than deleted.

Net effect across 51 files: 575 insertions, 1816 deletions. The insertions are almost entirely multi-line blocks collapsed into single terse lines, so the actual comment-text reduction is well past the line ratio.

Affected Areas

ffi/ runtime (thread, event, request-queue, context, cbor, alloc), ffi/internal/ macros (ffi_macro, c_macro_helpers, ffi_library, ffi_scalar), ffi/codegen/ generators (rust, cpp, c, meta, cddl, types_ir, string_helpers), the unit/bench tests, and the echo/timer examples. Changes are limited to comments and docstrings — no code tokens, identifiers, or emitted-code string literals were touched.

Impact on Library Users

No impact — comment-only change. No API, behavior, or performance change.

Risk Assessment

Minimal. Every touched file passes nph --check unchanged. Diff inspection confirms each changed code line differs only by a removed/shortened trailing comment; the code is byte-identical. No functional lines were altered.

References

N/A

[Ivansete-status]

We need to confirm there are no weird acronyms being used.
Thanks!