Skip to content

Sunset luria upgrade statuses at 1.0.0 - #189

Draft
dmarx wants to merge 2 commits into
mainfrom
claude/sunset-the-statuses-upgrade-e6evzh
Draft

dmarx wants to merge 2 commits into
mainfrom
claude/sunset-the-statuses-upgrade-e6evzh

Conversation

@dmarx

@dmarx dmarx commented Sep 6, 2026

Copy link
Copy Markdown
Owner

Draft, pinned to 1.0.0. More records are coming onto luria before the first stable release and each of them predates #181, so the command has work left even though both current records are done with it.

Two commits, and the first one is wanted now:

1. b89cc2a — say when it goes, not that it could have gone already

The sunset condition it shipped with was "once luria's own record and the anthology are both on the release that carries this". That was met about an hour after it was written, so spent-upgrades started reporting a removal that was not going to happen — on every luria lint run, in both records.

A marker stating a condition it does not mean is the failure it exists to prevent, one level up. Repinned to 1.0.0: the version rather than a list of projects, because a list is never finished at the moment you read it.

If you want the false nag to stop before 1.0.0, this commit is the whole fix — say the word and I'll split it into its own non-draft PR.

2. 135439c — the removal itself

Deletes luria/upgrade.py, its CLI registration, the spent-upgrades lint class and spent_upgrades(), and the three tests that covered them. The finding for a scheme declaring no status vocabulary now names the declaration to write rather than a command that no longer exists.

The marker goes with the thing it marked. Its job was to raise the question rather than wait to be remembered, and it did — keeping a check whose only finding is answered would be the same mistake in the layer above.

Before this merges

  • luria — its own record, on 0.10.0
  • dmarx/anthology-of-the-sota#25, on 0.10.0
  • dmarx/mathematics-of-meaning
  • dmarx/strata-g

Each unchecked record predates #181, so each needs luria upgrade statuses (or the equivalent by hand) while the command still exists. Neither repository is in this session's scope yet; both are reachable and I can pick them up on request.

Checks

  • python -m pytest tests -q: 905 passed (908 → 905, the three removed with the feature).
  • luria lint: clean — including the spent-upgrades row, which this removes.

🤖 Generated with Claude Code

https://claude.ai/code/session_01YP4P3m8rzVFb8idnTE4FfT


Generated by Claude Code

The sunset condition it shipped with — both records that predated #181
having run it — was met about an hour after it was written, so the
marker started reporting a removal that was not going to happen. A
marker stating a condition it does not mean is the thing it exists to
prevent, one level up.

Pinned to 1.0.0 instead. More records are coming onto luria before the
first stable release and each of them predates #181, so the version is
the honest gate; a list of projects is never finished at the moment you
read it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YP4P3m8rzVFb8idnTE4FfT
Both existed for #181, which every record that predates it has now run.
An upgrade kept past its purpose is dead code that still has to be read,
tested and explained — which is what `spent-upgrades` was added to
notice, and it noticed about an hour after being written.

So the marker goes with the thing it marked. Its job was to raise the
question rather than wait to be remembered, and it did; keeping a check
whose only finding is answered would be the same mistake in the layer
above.

The finding for a scheme declaring no `status` vocabulary now names the
declaration to write, since the command that used to write it is gone.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YP4P3m8rzVFb8idnTE4FfT
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