Better Todo Tree 1.1.20

Better Todo Tree 1.1.20

• release tag: v1.1.20
• target commit: 925479fa0202a40f08ca0aaa64547417674661ff
• previous release: v1.1.19

Included commits

• 559eeda fix: centralize regex sources and harden scan flows

  Regex construction routes runtime, tests, workflow checks, release
  checks, perf harnesses, and evidence tooling through a typed registry
  while preserving scan parity.

  Regex registry:

  • add shared fragments, named patterns, tag placeholders, default regex
    constants, builder helpers, and typed registry errors
  • replace duplicated literals in detection, utils, attributes,
    highlights, icons, tree, stream scanning, config, webpack, perf,
    release, workflow, parity, and test code
  • add an equivalence harness that covers 81/81 baseline sources,
    73/73 behavior cases, and zero non-registry regex literals

  Scan and command flow:

  • extend candidate workspace scans to JavaScript-compatible custom
    $TAGS regexes with balanced structure and zero-width tails
  • keep PCRE2-only backreference regexes on raw ripgrep normalization
  • wrap registered commands in a shared async boundary and return
    persistence, reveal, export, and migration work
  • roll back workspace-state writes before refresh work when
    multi-entry writes fail

  Ripgrep and evidence:

  • return RipgrepError for option, executable, spawn, JSON, cleanup,
    and cancellation failures
  • prevent stale search cleanup from clearing newer active searches
  • add registry, scan parity, ripgrep, workflow, release, and perf
    coverage plus accuracy and candidate-scan artifacts
  • ignore generated accuracy JSON and local settings exports

  Fixes #51

• fcd27fd fix: repair regex evidence and branch staging CI failures

  Branch staging applies staged and unstaged source patches separately, retains
  source-base snapshots, and merges fallback snapshot content against updated
  target files.

  Regex evidence resolves a CI-safe baseline, registers CodeQL workflow assertion
  patterns, and keeps busy-command icon assertions aligned with static product
  icons.

• b624123 Merge pull request #52 from FanaticPythoner/fix/issue-51-centralize-regex-sources-and-harden-scan-ripgrep-and-command-error-flows

  fix: centralize regex sources and harden scan flows

• 711118f fix: align reload tag offsets and incremental file order

  Ripgrep byte offsets are converted to JavaScript string offsets before
  custom-regex normalization so reload scans match editor scans for COBOL and SQL
  prefix patterns.

  Detection:

  • resolve ripgrep submatch byte positions against physical line text
  • use UTF-8 byte-to-string offset conversion for absolute and local offsets
  • cover issue #53 COBOL, SQL, bracket, and Unicode prefix cases

  Tree:

  • limit tag-order sorting to grouped root tag nodes
  • keep sibling file nodes sorted by path after incremental document replacement

  Packaging:

  • add npm publish ignores for source, tests, workflow tooling, and artifacts
  • preserve bundled dist runtime while excluding source maps

  Coverage:

  • add issue #53 regex matrix coverage for raw ripgrep payloads
  • add tree behavior coverage for sorted sibling files after updates
  • add issue #53 microbenchmark summary for scan-large-custom-regex

  Fixes #53

• dfcb666 fix: search regex audit history for usable baselines

  Walk first-parent ancestry from candidate refs so the equivalence harness selects
  the nearest commit that still exposes baseline regex entries after registry
  centralization.

  Report audit child-process exits and malformed stdout before JSON parsing, with
  QUnit coverage for both failure paths.

• e0861e7 Merge pull request #54 from FanaticPythoner/fix/issue-53-errors-in-tag-processing-during-a-full-scan

  fix: align reload tag offsets and incremental file order

• 925479f release: v1.1.20