chore: clear MoonBit deprecation warnings (26 → 0 warnings)

[bobzhang]

Brings moonbit-community/lisp_interpreter up to date with the current MoonBit toolchain (moon 0.1.20260901). This repo publishes 1 module on mooncakes.io.

Result

	before	after
Warning sites	26	0
Compile errors	0	0

The warning baseline was measured after the build was repaired — a failing compile stops before reporting warnings, so the initial sweep undercounted them.

Tests: pass — 33/33 (baseline was 32/33: one PRE-EXISTING failure, test "decode_string" in sexp.mbt, failing before any of my edits; it is now fixed)

What changed

• implicit_impl_as_method (0079)
• deprecated (0020)

Reviewer notes

MEASURED NUMBERS: moon check before = 26 warnings / 0 errors (the 22 unique sites in the given baseline; 4 sites emit two warnings each). After = 0 warnings, 0 errors, exit 0 (verified after rm -rf _build target clean rebuild). moon test after = 33 passed / 0 failed, exit 0. Nothing was suppressed: no #warnings(...), no deleted tests, no weakened assertions. BASELINE TEST STATE (recorded before editing): 33 tests, 32 passed, 1 FAILED — sexp.mbt test "decode_string", diff -Err: "[\" unexpected newline in string\"]" vs +Err: "[ unexpected newline in string]". This failure pre-dates my work; it was caused by the same upstream Show-for-collections change that produced the deprecation warnings. My fix incidentally repaired it. FILES: {sexp.mbt, sexp_test.mbt, lisp_interpreter.mbt, lisp_interpreter_test.mbt, pkg.generated.mbti}, plus moon.mod.json -> moon.mod. [0079] implicit_impl_as_method (12 sites): added plain pub extend declarations exactly as each message named — Sexp with @debug.Debug::{to_repr}, Eq::{equal, not_equal}, ToJson::{to_json}, Show::{to_string, output}; ParseError with @debug.Debug::{to_repr}, ToJson::{to_json}, Show::{to_string, output}; Value with @debug.Debug::{to_repr}, ToJson::{to_json}, Show::{to_string, output}; Env with @debug.Debug::{to_repr}, Show::{to_string, output}. I did NOT use the #deprecated alternative: the promoted methods ...

How this was verified

• Baseline and final counts are from moon check --output-json, deduplicated by (file, location, code), re-measured independently after the changes were made.
• No warning was silenced to reach these numbers: the diff adds no #warnings("-..."), alert-list or warn-list escapes.
• Test snapshots were reviewed rather than blind-accepted; where a suite already failed, that was confirmed against a clean HEAD.

🤖 Generated with Claude Code

https://claude.ai/code/session_01Yc3P6APF111hdbCXRT4M3P