refactor: simplify Wright production Rust

[e54-bot]

Summary

This branch consolidates repeated control flow and data handling across the analyzer, CLI, driver, language, LPP, and LSP crates. It also removes redundant wrappers and unreachable source-edit analysis while preserving the existing supported behavior and external contracts.

Accepted result

The original target was a reduction of at least 30% in production Rust LOC. The measured result is 14,877 baseline lines to 12,854 lines: 2,023 fewer lines (13.60%), leaving a 2,441-line gap to the original target. On 2026-09-27, the user accepted the current simplification as completion of this effort despite the unmet 30% metric. This is an explicit acceptance adjustment; it does not claim that the 30% target was reached.

PR #419 was merged at 93a5914dc77899a9c674a692b27a38df4cd941b9.

Verification

Local checks passed on commit bce3c7feee3dd7ff465a56a909185d7754300b5b:

• cargo +1.85.0 fmt --all -- --check
• cargo +1.85.0 clippy --locked --all-targets --all-features -p wright-analyzer -p wright-transform -p wright-lpp -- -D warnings
• cargo +1.85.0 test --locked --all-targets --all-features -p wright-analyzer -p wright-transform -p wright-lpp </dev/null
• cargo fmt --all -- --check
• cargo clippy --workspace --all-targets --all-features -- -D warnings
• cargo test --workspace --all-targets --all-features </dev/null
• git diff --check

Hosted CI run 36313151674 completed successfully at the same commit: 22 jobs passed, including Rust 1.85.0, stable quality, integration and Workshop dogfood jobs, cross-platform builds, and distribution validation.

Tests and fixtures are unchanged from origin/main.

[indent]

This PR is still not bounded by an approved Issue, and its scope repeats the exact combination that got #370 CHANGES_REQUESTED on 2026-09-19 (#370 (review)): a ≥30% Rust-LOC-reduction target as the acceptance criterion, applied across analyzer, CLI, driver, language, LPP, and LSP.

The closest recorded contract is #355, and it excludes exactly this:

• "Non-goals: Removing code solely to meet a line-count or crate-count target."
• "Non-goals: Refactoring unrelated analyzer, transform, driver, or CLI code."
• Scope is limited to removing the legacy pre-canonical analyzer/transform path; it says nothing about wright-cli, wright-driver, wright-language, wright-lpp, or wright-lsp.

PR #419 touches crates/wright-cli/{cli.rs,main.rs,completion.rs,present.rs,update.rs,bin/wright-serve.rs}, crates/wright-driver/*, crates/wright-language/*, crates/wright-lpp/*, and crates/wright-lsp/main.rs — all outside #355's scope — driven by the same unapproved LOC mandate the PR body restates ("the agreed 30% production Rust LOC reduction target ... is not met").

No new Issue authorizing this cross-crate, LOC-driven mandate exists (checked open/closed issues #350-419). Per org policy, an agent must not self-authorize a scope/architecture decision reserved for the owner, and this exact mismatch was already raised once without a follow-up Issue being opened to settle it.

Required: either open/approve an Issue that explicitly authorizes this cross-crate, LOC-driven simplification (superseding #355's non-goals), or re-scope this PR to #355's actual boundary (legacy pre-canonical analyzer/transform removal only) and move the wright-cli/wright-driver/wright-lpp/wright-lsp changes into separately-scoped, Issue-bound PRs.

Separately, on execution quality: I traced the diff hunk-by-hunk against main for the highest-risk changes (the "remove unreachable source edit analysis" deletion in wright-driver/src/edit.rs, the wright-language hover/definition/rename stubs, the wright-driver/session.rs branch merges, the CLI completion rewrite, the LPP Capability-derived method dispatch, and the canonical-analyzer file split) and found no behavioral regression — the "unreachable" claims check out against actual call sites on main, not just local diff context. That part of the work is sound; the open issue is scope authorization, not correctness.

[indent]

Confirmed with the requester: this refactor was directed locally out-of-band (no Issue opened), which resolves the scope-authorization objection from my previous review — the owner has authorized superseding #355's non-goals for this effort. No commits changed since that review (still bce3c7f), and the earlier hunk-by-hunk correctness pass found no behavioral regressions. Approving.