Challenge submission: add trace metrics toolkit

[jianmosier]

Summary

This is an integrated TSPerf challenge submission branch for the Algora TSPerf bounty: https://algora.io/challenges/tsperf

It turns the extension from a trace viewer into a reproducible TypeScript performance measurement workflow for type complexity and load/check time:

• writes a per-run metrics.json artifact after Tracer: tsc trace
• supports stock TypeScript types.json files that do not include timestamped spans
• parses --extendedDiagnostics output into structured metrics such as types, instantiations, memory, check time, and total time
• adds Tracer: Compare trace metrics for before/after Markdown reports
• adds Tracer: Trace current file for focused mini traces from the active editor
• fixes trace tree binary search coverage and pos=0 source navigation
• documents the review path in docs/challenge-demo.md, docs/challenge-evidence.md, and docs/tsperf-submission.md

Installable Challenge Build

• VSIX release: https://github.com/jianmosier/tracer/releases/tag/tsperf-challenge-c3a4de8
• VSIX asset: https://github.com/jianmosier/tracer/releases/download/tsperf-challenge-c3a4de8/tracer-0.1.3.vsix
• SHA-256: 7a63eb5faacc197c6d29e62b8a98471c467510240cce1614510b4df72b5722be
• Submitted commit: c3a4de850e4afd84a0d44c3e6725fc4905e98d33
• Evidence doc included in the VSIX: docs/challenge-evidence.md

Review notes

I also opened the implementation as smaller review slices so maintainers can review/merge incrementally if preferred:

• test: cover trace tree binary search bounds #69 binSearch / source-position correctness
• feat: write trace run metrics artifact #70 trace run metrics artifact
• fix: tolerate untimestamped trace type data #71 stock types.json support
• feat: attach extended diagnostics to trace metrics #72 extended diagnostics in metrics
• feat: compare trace run metrics #73 trace metrics comparison
• feat: add active file mini trace command #74 active file mini trace

This PR is the complete integrated solution branch for challenge evaluation.

Validation

Run locally on this branch:

• pnpm lint
• pnpm typecheck
• pnpm build
• pnpm exec vitest run - 8 files / 21 tests passed

pnpm build completes with the repository's existing Rollup/Nuxt/Browserslist warnings only.

[jianmosier]

Added an installable VSIX release for easier challenge review:\n\n- Release: https://github.com/jianmosier/tracer/releases/tag/tsperf-challenge-a3d250d\n- Asset: tracer-0.1.3.vsix\n- Commit: a3d250d4eadbb5f62b10cb95fa47419aacd5865b\n- SHA-256: 99bf175d840475f2fa36170cdf84d63e4fdb39b0e69818410a83cd8ee63e5a03\n\nThe package includes the built extension plus docs/challenge-demo.md and docs/tsperf-submission.md.

[jianmosier]

Updated the challenge submission with a reviewer-facing evidence doc and a fresh installable VSIX build.\n\n- Commit: c3a4de8\n- Release: https://github.com/jianmosier/tracer/releases/tag/tsperf-challenge-c3a4de8\n- Asset: https://github.com/jianmosier/tracer/releases/download/tsperf-challenge-c3a4de8/tracer-0.1.3.vsix\n- SHA-256: 7a63eb5faacc197c6d29e62b8a98471c467510240cce1614510b4df72b5722be\n\nValidation re-run locally after the docs update: pnpm lint, pnpm typecheck, pnpm build, and pnpm exec vitest run all passed. The VSIX includes docs/challenge-evidence.md with representative metrics.json and comparison-report output.