From f11dd5c4e70ebbf487205aa72998e87b123d7f9d Mon Sep 17 00:00:00 2001 From: Craig Jennings Date: Fri, 12 Jun 2026 17:05:50 -0500 Subject: chore: file the cancelled archive-done task to Resolved --- ...iority-scheme-convention-and-signal-cleanup.org | 75 ++++++++++++++++++++++ todo.org | 17 +++-- 2 files changed, 83 insertions(+), 9 deletions(-) create mode 100644 .ai/sessions/2026-06-12-17-04-priority-scheme-convention-and-signal-cleanup.org diff --git a/.ai/sessions/2026-06-12-17-04-priority-scheme-convention-and-signal-cleanup.org b/.ai/sessions/2026-06-12-17-04-priority-scheme-convention-and-signal-cleanup.org new file mode 100644 index 0000000..0a69944 --- /dev/null +++ b/.ai/sessions/2026-06-12-17-04-priority-scheme-convention-and-signal-cleanup.org @@ -0,0 +1,75 @@ +#+TITLE: Inbox clear, task review, priority-scheme convention +#+DATE: 2026-06-12 + +* Summary + +** Active Goal + +Startup, then four threads Craig directed in sequence: clear the inbox, run a task review, establish a per-project Priority Scheme convention, then two cleanups — quiet the archive-done no-op report and remove the Signal footprint (page-signal wrapper + the Signal triage plugin) while he rebuilds the Signal client from scratch. + +** Decisions + +- Every project's todo.org opens with a "[Projectname] Priority Scheme" section declaring both the [#A]-[#D] semantics and the tag vocabulary in use. Mandatory; each project defines its own set. Codified in claude-rules/todo-format.md. +- The todo-cleanup --archive-done "0 moved while moving" bug is not a defect — it's a correct second-run report (the wrap runs --archive-done twice). Cancelled the task as cannot-reproduce. +- Still fixed the real ergonomic nit: --archive-done now stays silent on a real-mode no-op. Check mode and missing-section skips still report. +- Deleted the page-signal wrapper + workflow; kept the signal MCP server (separate two-way capability). +- Signal client rebuild (Craig): removed the Signal triage-intake plugin + engine references; KEPT the signal-mcp server (mcp/servers.json) and its README section per his option-3 call. The signal-mcp config (path + +15045173983) will not change in the rebuild. Re-add the plugin when stable. +- Task review batch of 7: all kept, none re-graded or killed. + +** Data Collected / Findings + +- Survey of all 20 .ai projects for the Priority Scheme section: 7 already conform, 3 have no root todo.org (git-gutter, google-contacts.el, yt-sync), 11 missing — broadcast a handoff to each of the 11. +- archive-done double-run mechanism: open-tasks.org Phase A and wrap-it-up both run --archive-done, so the second pass correctly reports 0 next to the first pass's diff. The tc-archived counter was always correct (cl-incf inline with each move). +- Full make test green throughout: 415+54+12 pytest, all ERT (28/28 todo-cleanup incl. 5 new), all bats, exit 0. + +** Files Modified + +- claude-rules/todo-format.md — new "Priority and Tag Scheme Header" rule. +- todo.org — renamed scheme section to "Rulesets Priority Scheme", refreshed 7 review-date stamps, cancelled the archive-done bug task, pager VERIFY removed (Craig's live edit, bundled). +- .ai/scripts/todo-cleanup.el + test-todo-cleanup.el (canonical + mirror) — quiet real-mode no-op + capture helper + 5 ERT tests. +- Deleted: claude-templates/bin/page-signal, page-signal.org, triage-intake.signal.org (+ mirrors); removed the ~/.local/bin/page-signal symlink. +- Edited references: INDEX.org, broadcast.org, mcp/README.org, triage-intake.org. +- Six commits pushed: 81395a7, 3b986e3, f2f5f57, 21ab10e, 13256aa, d8c38b0. + +** Next Steps + +- 11 projects have a Priority Scheme handoff to apply on their next session (archangel, chime, duet, emacs-wttrin, gloss, little-elisper, mcp.el, org-drill, rsyncshot, website, winvm). +- git-gutter, google-contacts.el, yt-sync have no root todo.org — no handoff sent; add a scheme if they grow a task list. +- Re-add the Signal triage plugin (drop a new triage-intake.signal.org back in, re-list Signal in the engine enumeration) once the rebuilt Signal client is stable. +- The page-signal pager number is deregistered and the VERIFY is dropped — no action unless Craig wants paging back. + +KB: promoted 0 / consulted no + +* Session Log + +** 2026-06-12 07:32 CDT — Startup + inbox clear + +Startup ran clean: both repos current, .ai/ synced, no crash anchor, no reminders/pending decisions. Two inbox items, both no-action FYIs: the .emacs.d task-audit provenance note (commit fbeef0c already in HEAD) and a script-generated task-review staleness note (lint-followups.org). Craig approved clearing both, then a task review, then a new directive: every project's todo.org should carry a "[Projectname] Priority Scheme" section at the top defining both priorities (A-D) and the tags in use (typically quick/solo/feature/bug/test/refactor for code projects). Each project defines its own, but the section is mandatory. Deleted the two FYIs. + +** 2026-06-12 — Task review (batch 7) + +Ran task-review on the 7 oldest-unreviewed top-level tasks. All 7 kept as-is, stamped :LAST_REVIEWED: 2026-06-12 (page-signal VERIFY got a new PROPERTIES drawer; it was never-reviewed). No re-grades, no kills, no new :quick:/:solo: tags. Staleness >7-day count now 0. Flagged the page-signal VERIFY to Craig as the one item he could clear now (re-register +15045173983 via captcha/SMS). + +** 2026-06-12 — Priority-scheme convention + +Craig's directive: every project's todo.org must open with a "[Projectname] Priority Scheme" section declaring both priorities (A-D) and the tag vocabulary in use. Codified it as a new section in claude-rules/todo-format.md ("Priority and Tag Scheme Header"). Conformed rulesets' own todo.org: renamed "* Priority and Tag Scheme" → "* Rulesets Priority Scheme". Survey of all 20 .ai projects: 7 already follow the naming convention (archsetup, auto-dim-other-buffers.el, .emacs.d, home, pearl, rulesets, work), 3 have no root todo.org (git-gutter, google-contacts.el, yt-sync), 11 are missing the section. Broadcasting a handoff to those 11 (gitignore-broadcast pattern — each adds its own section from its existing task usage, confirms with Craig): archangel, chime, duet, emacs-wttrin, gloss, little-elisper, mcp.el, org-drill, rsyncshot, website, winvm. + +** 2026-06-12 — Commit + push + +Published via the full flow: Step 0 reconcile (0 behind), /review-code --staged (Approve, no blockers), /voice personal on both drafts. Two commits: 81395a7 docs(todo): require a per-project Priority Scheme header (todo-format.md); 3b986e3 chore: conform the task list to the Priority Scheme header (todo.org rename + review-date stamps + lint-followups.org deletion). Pushed origin main 8e18033..3b986e3 — also carried the earlier unpushed fbeef0c (task-audit) out. All authored as Craig. + +** 2026-06-12 — /start-work on todo-cleanup --archive-done 0-moves bug: Phase 0.3 says not-a-bug + +Existence check could not confirm the bug. Evidence: (1) todo-cleanup.el is unchanged since the wrap that logged it (no commits in b6d286f..HEAD touch it), so the code that ran is the code I read; (2) the counter tc-archived is cl-incf'd inline with each move and read straight in the report — no reset, no rebind path; (3) reproduced the exact pre-archive state (b6d286f:todo.org) through the unchanged tool: reports "3 subtree(s) moved", listing all 3, count correct; (4) the wrap commit 8e18033 moved 3 subtrees, not 4 (task body miscount); (5) a SECOND --archive-done run on the already-archived copy reports "0 subtree(s) moved" — correct, nothing left. Conclusion: the observed "0 moved while moving" was a second invocation in the wrap (open-tasks.org Phase A also runs --archive-done) correctly reporting 0, with the git diff coming from the first run. Not a code defect. No task claimed (todo.org claim is deferred past the Justify gate), so nothing to roll back. Also: todo.org has an uncommitted external edit (page-signal VERIFY deleted) — left untouched, flagged to Craig. Cancelled + committed as f2f5f57 (bundled the pager VERIFY removal per Craig). + +** 2026-06-12 — Two follow-ups: quiet archive-done no-op + delete page-signal + +Archive-done ergonomic fix (TDD): added tc-test--archive-report capture helper + 5 ERT tests (real-mode 0-move silent, double-run 2nd pass silent, check-mode still previews 0, moves still reported, skips still reported). Red confirmed on the two silence tests, then fixed tc--emit-archive-report to stay silent in real mode when tc-archived=0 and no archive-skip (check mode + skips always report). Inlined the per-issue dolist for parity with the sibling report emitters rather than a single-caller helper. 28/28 ERT green; end-to-end double-run on b6d286f:todo.org shows run 1 reports 3, run 2 silent. Synced canonical→mirror. + +page-signal deletion (narrow): removed claude-templates/bin/page-signal, the page-signal.org workflow (canonical + mirror), the INDEX.org entry, the broadcast.org example mention, and the page-signal wrapper references in mcp/README.org (kept the signal MCP two-way path — separate capability). Removed the dangling ~/.local/bin/page-signal symlink. Historical mentions in sessions/, PROCESSED inbox, notes.org log line, and resolved todo.org tasks left as record. Full make test green (415+54+12 pytest, all ERT, all bats, exit 0). Two commits next. + +Published both: /review-code --staged (Approve, no findings), /voice personal on both drafts. 21ab10e fix(todo-cleanup): keep --archive-done silent on a real-mode no-op (4 files, canonical+mirror); 13256aa chore: delete the page-signal pager wrapper (workflows, bin, mcp/README, canonical+mirror, symlink removed). Pushed origin main f2f5f57..13256aa. All authored as Craig. + +** 2026-06-12 — Signal triage plugin removed (client rebuild) + +Craig is rebuilding the Signal client from scratch; swept the repo for Signal-the-messenger footprint. Per his option-3 ruling: KEEP the signal-mcp server (mcp/servers.json) and the README setup section; remove only the triage plugin + engine references. Deleted claude-templates/.ai/workflows/triage-intake.signal.org (+ mirror), de-listed Signal from triage-intake.org's plugin enumeration (line 51), and fixed the "signal and telegram plugins" rationale ref to just telegram (kept the 2026-06-10 incident war-story). Left telegram plugin, broadcast.org's signal-cli example, and history untouched. Committed d8c38b0 chore: drop the Signal triage-intake plugin, pushed 13256aa..d8c38b0. Re-add the plugin when the client stabilizes. diff --git a/todo.org b/todo.org index 3ad97a2..a456c26 100644 --- a/todo.org +++ b/todo.org @@ -851,15 +851,6 @@ Triggers that would prompt "let's build it now": Upstream reference (do not vendor): ComposioHQ/awesome-claude-skills =content-research-writer/SKILL.md=. -** CANCELLED [#D] todo-cleanup =--archive-done= reports 0 moves while moving subtrees :bug: -CLOSED: [2026-06-12 Fri] -:PROPERTIES: -:CREATED: [2026-06-12 Fri] -:END: -Observed at the 2026-06-12 wrap: the pass relocated closed subtrees from Open Work to Resolved while printing "todo-cleanup --archive-done: 0 subtree(s) moved". - -CANCELLED 2026-06-12 — cannot reproduce. =todo-cleanup.el= is unchanged since the wrap that logged this, and =tc-archived= is incremented inline with each move and read straight in the report, so no move can go uncounted. Running the exact pre-archive state (=b6d286f:todo.org=) through the tool reports the right count (3 moved, all listed). The "0 moved" was a correct second-run report: =open-tasks.org= Phase A runs =--archive-done= after wrap-it-up already archived, so the second pass finds nothing to move and prints 0 next to the first pass's git diff. Not a code defect. - ** TODO [#D] Revisit =c4-*= rename if a second notation skill ships :chore: :PROPERTIES: :LAST_REVIEWED: 2026-06-10 @@ -2688,3 +2679,11 @@ CLOSED: [2026-06-11 Thu] Cancelled 2026-06-11: Craig confirmed the decision — one todo queue with a single Open Work / Resolved pair. Home reshapes its consolidated file to that form, and the existing single-pair tooling works unmodified. No code change needed. +** CANCELLED [#D] todo-cleanup =--archive-done= reports 0 moves while moving subtrees :bug: +CLOSED: [2026-06-12 Fri] +:PROPERTIES: +:CREATED: [2026-06-12 Fri] +:END: +Observed at the 2026-06-12 wrap: the pass relocated closed subtrees from Open Work to Resolved while printing "todo-cleanup --archive-done: 0 subtree(s) moved". + +CANCELLED 2026-06-12 — cannot reproduce. =todo-cleanup.el= is unchanged since the wrap that logged this, and =tc-archived= is incremented inline with each move and read straight in the report, so no move can go uncounted. Running the exact pre-archive state (=b6d286f:todo.org=) through the tool reports the right count (3 moved, all listed). The "0 moved" was a correct second-run report: =open-tasks.org= Phase A runs =--archive-done= after wrap-it-up already archived, so the second pass finds nothing to move and prints 0 next to the first pass's git diff. Not a code defect. -- cgit v1.2.3