Skip to content

mind: date the backlog — a Filed: header on all 150 draft prompts - #274

Merged
Jammy2211 merged 1 commit into
mainfrom
claude/pyautomind-task-dates-recent-06v1g7
Aug 23, 2026
Merged

mind: date the backlog — a Filed: header on all 150 draft prompts#274
Jammy2211 merged 1 commit into
mainfrom
claude/pyautomind-task-dates-recent-06v1g7

Conversation

@Jammy2211

Copy link
Copy Markdown
Collaborator

The date convention covered the registries and active/, and missed the largest pool of tasks the Mind holds. draft/ is 150 prompts against a handful of live rows, so the dashboard's Recent feed could reach only 10 of the 50 it holds — most of what had been happening was invisible.

Drafts now carry Filed: YYYY-MM-DD, the day the prompt was written, beside the Issued: an active prompt carries. A prompt keeps its Filed: when it advances and gains an Issued:; the later, more specific event dates the task.

The --follow switch

lifecycle.py dates now covers both state folders, and the two want opposite readings of the same history:

State Reading --follow?
active/ the day it arrived there — being issued is a git mv, so following the rename back would report the wrong day no
draft/ the day it was written, wherever it lived then yes

This matters concretely: the 2026-07-13 lifecycle migration git mv-ed 42 prompts in one commit. Without --follow all 42 date from the migration rather than from themselves — a fact about the repo's plumbing, not about the work. With it, the 150 prompts spread across 53 distinct dates.

The backfill

All 150 dated from git history, none guessed. The Intake trailer (<!-- formalised by the Intake (Conception) Agent on … -->, carried by 53 of them) and the claiming registry entry stand behind git as fallbacks, and every backfilled date still names its source in the file:

Filed: 2026-08-21 (backfilled from git)

Note this required un-shallowing the working clone: a shallow checkout reports its boundary commit as the day everything older appeared, and the guard added in #270 correctly refused those. With full history git dated every prompt on the first source.

Testing

  • 5 new tests, including a real two-commit git fixture proving a bulk git mv does not re-date the backlog, and that active/ takes the opposite reading of the same history.
  • python3 -m pytest tests/ — 181 passed.
  • lifecycle.py check clean; lifecycle.py dates reports OK.

The renderer PR must merge first: this repo's refresh check renders the dashboard using PyAutoBrain's main.


Generated by Claude Code

The date convention covered the registries and `active/`, and missed the
largest pool of tasks the Mind holds: `draft/` is 150 prompts against a handful
of live rows, so the dashboard's Recent feed could reach only 10 of the 50 it
holds. Most of what had been happening was invisible.

Drafts now carry `Filed: YYYY-MM-DD` — the day the prompt was written — beside
the `Issued:` an active prompt carries. A prompt keeps its `Filed:` when it
advances and gains an `Issued:`; the later, more specific event dates the task.

`lifecycle.py dates` covers both state folders, and the two want OPPOSITE
readings of the same history — the switch is `--follow`. An `active/` prompt
dates from the day it ARRIVED there (being issued is a `git mv`, so following
the rename back would report the wrong day); a `draft/` prompt dates from the
day it was WRITTEN, wherever it lived then. The 2026-07-13 lifecycle migration
`git mv`-ed 42 prompts in one commit: without `--follow` all 42 date from the
migration rather than from themselves. With it, the 150 spread across 53
distinct dates.

All 150 backfilled from git history, none guessed — the Intake trailer and the
claiming registry entry stand behind git as fallbacks, and every backfilled
date still names its source in the file.

Co-Authored-By: Claude Opus 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01L4HPWjv5rdzBAkKpfW1SbR
@Jammy2211
Jammy2211 merged commit dd5e502 into main Aug 23, 2026
4 of 5 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants