Skip to content

CHANGELOG.md

Latest commit

 

History

History
101 lines (80 loc) · 8.74 KB

CHANGELOG.md

File metadata and controls

101 lines (80 loc) · 8.74 KB
Raw

Changelog

All notable changes to vim-openvox.

[Unreleased] - Phase 4 M1: Core Bugfixes, Hardening & Feature Wiring (Harness Execution)

Fixed (M1)

  • autoload/openvox/align.vim: Removed dead l:target_col (scope leak, unused, used stale l:indent). Replaced rough single-line post-=> heuristic in block() with proper column-based detection (collect safe arrows + compute max_key + compare current arrow_col to expected = len(indent) + max_key per entry for true cross-line misalignment per style guide). Added optional silent param to arrows()/block() (for auto paths). Made BufWritePre auto call use silent=1 (context-aware, no noise on save for non-violations or "not in block"). Only echoerr on real violations (unsafe arrows in strings/comments, unaligned in block); info echoes suppressed in silent. Safety model (IsInStringOrComment + searchpair skip + heredoc/interp) kept/enhanced + documented.
  • indent/puppet.vim: Updated s:IsStringOrComment to match align's precise logic (synID + 'Comment|String|Heredoc|Interpolation' regex + comment; was less complete 'string|comment|heredoc').
  • plugin/openvox.vim: Wired g:openvox_auto_align = 0 default. Set g:openvox_lint_command default to 'openvox-lint' (from legacy 'puppet-lint').
  • autoload/openvox/lint.vim + compiler/openvox_lint.vim: Branding/consistency sweep - all user-facing echoes now dynamic: let l:tool = fnamemodify(get(g:, 'openvox_lint_command', 'openvox-lint'), ':t') (and equiv for puppet_cmd). No more hard 'puppet-lint:' / 'puppet:' in lint/validate paths. Removed no-op YOLO style_checks force loop (and related dead comments). Style now "real" via defaults + explicit --no- handling (openvox-lint supports --only-checks via user args if needed; noted in code). Internal types kept for compat; s:lint_tool / s:puppet_tool for display. Header/comments cleaned for openvox-lint preference + compat.
  • Corrected prior overclaims in docs/CHANGELOG from incomplete YOLO (no evidence of prior harness use, dead code persisted, heuristic not full col, no-op force, defaults not wired, etc.).

Added (M1)

  • g:openvox_auto_align fully wired + documented (README config, doc/openvox.txt, AGENTS.md + .grok/AGENTS.md).
  • Harness compliance: todo_write used throughout M1 (one in_progress at time); all files read before edit (read_file/grep/list_dir); search_replace for changes; git worktree isolation at /tmp/vim-openvox-phase4-m1 (branch phase4-m1, source /Users/jsheets/workspace/Personal/vim-openvox on development); pre-commit followed (this CHANGELOG, docs, verification runs, no version bump needed).
  • AGENTS.md / .grok/AGENTS.md updates for M1 details + new config var.
  • Verification: vint simulation, batch vim/nvim loads for g: defaults, greps for changes, git status/diff in worktree.

Changed

  • Pre-commit discipline enforced for M1 (even in subagent implementer mode). All edits in isolated worktree. No new files created (chose sync indent helper over shared util.vim to obey "NEVER create unless absolutely necessary").
  • Default now prefers openvox-lint everywhere (plugin, compiler, dynamic echoes, docs notes updated for compat).
  • Alignment auto is discoverable/configurable (was in align comments only, undoc'd, default not set).

M1 executed per Phase 4 plan (from memory synthesis + explicit M1 tasks). Parallel reviewer handoff via outputs/todos. M2+ (tests/Makefile/CI/full docs/release) not started. See worktree for patch: git diff in /tmp/vim-openvox-phase4-m1 .

This closes the identified gaps (target_col, heuristic, wiring, branding, no-op, helper inconsistency, docs drift, harness usage).

[Unreleased] - Phase 4 M2: Testing Foundation & Makefile

Added

  • Full automated test suite modeled after vim-grok:
    • tests/vimrc: Isolated minimal config (disable mappings/auto, set openvox-lint).
    • tests/run.vim: Runner with failure tracking, sources test_*.vim.
    • tests/test_core.vim: Plugin load, commands, functions, defaults, indent settings.
    • tests/test_align.vim: Basic alignment, string safety, block().
    • tests/test_indent.vim: Resource/conditional/heredoc indent rules.
  • Makefile:
    • make help, make lint (vint --style-check on Vimscript dirs).
    • make test (vim + nvim via isolated runner).
    • make ci-test (non-interactive, log capture).
    • make clean.
  • Updated tests/README.md with run instructions and how to add tests.
  • Wired into AGENTS.md (Development Commands), CONTRIBUTING.md (new Testing section), README.md (M2 note + make references).

Changed

  • Pre-commit now includes make lint && make test (or ci variant) as automated step.
  • Estate alignment: Full harness use (todos, pre-commit, AGENTS updates).

M2 completes testing/CI foundation for Phase 4.

[Unreleased] - Phase 4 M3: CI Expansion, Release Workflow Skeleton & Pre-Commit Automation

Added

  • Expanded .github/workflows/ (evolved from single ci.yml):
    • lint.yml: On push/PR to development; ubuntu, setup-python, pip vint, make lint.
    • test.yml: Matrix (editor: [vim, neovim]; Vim primary as user works exclusively in Vim, Neovim for compat); uses rhysd/action-setup-vim; make ci-test; failure artifact upload of /tmp/*.log.
    • release.yml: Skeleton on v* tags; generates notes from CHANGELOG, creates draft GH Release (manual finalize per policy; no auto in commit flow).
  • Deprecated old ci.yml (content updated to note; logic evolved to new files).
  • Pre-commit automation mandated in CONTRIBUTING.md and AGENTS.md: "Run make lint && make test (or ci-test) before any push/PR". Tied to estate Harness pre-commit checklist.
  • "CI Status" notes in README.md (workflows, matrix, Vim priority, badge).
  • Updated AGENTS.md (Development Commands + CI Status), CHANGELOG (this section), CONTRIBUTING (expanded testing + new Pre-Commit & CI section).

Changed

  • CI now separate lint + test (matrix + artifacts) + release skeleton. Matches vim-grok patterns.
  • Branch: workflows on development (project convention).
  • Pre-commit now automated/enforced via Makefile + CI (beyond docs).

M3 completes CI/release skeleton and automation for Phase 4.

[Unreleased] - Phase 4 M4: Documentation Sync, Full Branding Cleanup, Lifecycle & Discoverability

Added

  • RELEASE_PROCESS.md (modeled on vim-grok; version in plugin/openvox.vim + README, CHANGELOG, pre-commit via make + harness, tag on development, draft GH release, deprecations 2 minor versions min).
  • g:openvox_plugin_version in plugin/openvox.vim (referenced in README/docs).
  • Full docs sync in DOCUMENTATION.md: refreshed inventory (now 30+ files including AGENTS, .grok/, tests/, workflows/, Makefile, RELEASE_PROCESS.md), architecture, detailed sections for align.vim (auto/silent/column/violation/safety), lint.vim (dynamic branding/style), plugin (new g: vars), testing/Makefile. Noted "Phase 4 complete".
  • Updated doc/openvox.txt: added g:openvox_auto_align (config + alignment cross-ref), cleaned "puppet-lint"/"puppet:" to prefer openvox-lint + compat note, updated status/intro/differences/requirements/linting/FAQ for Phase 4.
  • README.md: updated YOLO/Phase 4 notes to "complete", added g:openvox_auto_align example, testing/CI/RELEASE_PROCESS notes, requirements (openvox-lint default).
  • AGENTS.md: marked YOLO issues addressed, "Phase 4 complete", updated Development Commands/CI Status, YOLO Rules to "structured with harness".
  • .grok/AGENTS.md: added pre-commit require make lint && make test; reference plan/harness.
  • CHANGELOG: this M4 section (docs sync, branding cleanup, lifecycle, Phase 4 complete note).

Fixed/Changed

  • Branding cleanup: remaining "puppet-lint"/"puppet:" in docs/README/AGENTS/DOCUMENTATION/doc/openvox.txt updated to prefer openvox-lint + compat (dynamic in code from M1/M3).
  • Pre-commit now includes docs sync for M4 changes.
  • Version bumped to 1.1.0 in plugin/docs.

M4 completes docs/lifecycle for Phase 4. See M5 for estate closeout. Full verification: docs match code, "Phase 4 complete" noted, RELEASE_PROCESS followed for future releases.

[Unreleased] - Phase 4 M5: Estate Alignment, Final Verification & Closeout

Added/Changed

  • Estate plan.md updated with M4 status + "Phase 4 progressing; M5 next".
  • Todos: M4 complete; phase4 in_progress (M5 closeout/harness use).
  • Full verification: local make lint && make test; docs/CHANGELOG/AGENTS synced; pre-commit followed; worktree/pruning from prior M's cleaned.
  • Phase 4 complete: M1-M5 done. See AGENTS.md "Phase 4 Complete", PHASE4_PLAN.md, estate plan.md. Project at modern maturity (9-10/10 per subagent reviews).

This addresses all identified gaps from YOLO review + subagent evaluations (alignment, artifacts, style catching, tests/CI, docs, lifecycle, estate). Next: M5 closeout or estate items (e.g., itsys seeding).

See full history in git.

[1.0.1] - Previous

(Keep existing from original README/CHANGELOG if present; assume prior version notes.)

See full history in git.