Skip to content

Release pass 0.11.0: what an operator now schedules, and the roadmap reconciled - #37

Merged
arpanghoshal merged 2 commits into
v0.11/3a-retention-reviewfrom
release/0.11.0
Sep 14, 2026
Merged

arpanghoshal merged 2 commits into
v0.11/3a-retention-reviewfrom
release/0.11.0

Conversation

@arpanghoshal

Copy link
Copy Markdown
Member

Pairs with ctrlrun#206, and stacks on #36.

operations.mdx is the page this milestone most changes and said nothing about either new command. Anchoring and pruning are the two things v0.11 gives an operator to run, and both are runs their scheduler makes. The "there is nothing to run" bullet now says which two commands write, and that no hold expires, which is SPEC-v0.9.md §4's rule about a timer applied unchanged.

The interval is the claim. An anchor freezes a prefix, so the window you are exposed to is (last anchored seq, current head], and that number is what to quote rather than any sentence about tamper-evidence.

Three pages quoted a verify transcript headed ctrlrun 0.10.0. The guard that actually runs verify and compares caught all three.

ROADMAP.md marks v0.11 shipped, with exit criteria stated as what ran rather than what the line promised, and a "Reconciled against what shipped" block for the four sentences that did not survive the build: the anchor detects truncation and not append; coverage comes from receipts and not events, because ACTION_PROPOSED carries an action_hash and nothing that maps it back; CONTROL-MAPPING.md is still not written; and retention shipped twice, because the review the build order made required for the prune found seven defects and the merge did not wait for it.

test_release_documents.py re-pins the README paragraph the kernel rewrapped. Only one of its four lines carries a word the T180 scan looks for, and the allow-list carries only that one, because an entry matching nothing is how a softened allow-list gets caught.

Every generator run against the release branch:

cli reference: matches            api reference: 83 pages, 0 drifted
schemas: both match               capabilities: 26 entries, 0 drifted
cookbook: 19 recipes, 0 drifted   probe page: matches
soak page: matches                badges: 0 drifted
readiness: 6,221 tests, version 0.11.0

CLAIMS.md needed no repointing: scripts/repoint-claims.py re-pointed 0, unresolved 0. git status in the kernel tree afterwards is clean apart from the commit this pairs with.

all checks passed             (ruff, ruff format, 1727 passed)
snippets: 98 runnable block(s) in 198 documents, 0 failed
lint: 198 document(s), 0 finding(s)
links: 192 document(s), 0 broken, 0 planned

Do not merge before #36. The tag is the maintainer's.

🤖 Generated with Claude Code

…the build

Exit criteria against what actually ran, not against what the line promised:
five released wheels rather than fixtures, the released 0.10.0 from PyPI rather
than a store this build wrote and called 0.10.0's, and two prunes racing under
the multi-process standard against a real Postgres server.

Three sentences are reconciled rather than quietly edited. The anchor does not
detect an append, and the bullet said "erased or appended" until today. Coverage
does not come from events, because `ACTION_PROPOSED` carries an `action_hash`
and nothing that maps it back to a name. And `CONTROL-MAPPING.md` is still not
written, which the v0.11 line cited in the present tense.

Readiness re-measured, and the version moves to 0.11.0.

Signed-off-by: arpan <contact@arpanghoshal.com>
…nscripts

`operations.mdx` is the page this milestone most changes and said nothing about
either new command. Anchoring and pruning are the two things v0.11 gives an
operator to run, and both are runs **their** scheduler makes: nothing sweeps,
nothing reaps, and no hold expires, which is the same rule about a timer that
`SPEC-v0.9.md` §4 applied to a budget hold.

The interval is the claim. An anchor freezes a prefix, so the window you are
exposed to is `(last anchored seq, current head]`, and that number is what to
quote rather than any sentence about tamper-evidence.

Three pages quoted a verify transcript headed `ctrlrun 0.10.0`. The guard that
runs verify and compares caught all three.

`ROADMAP.md` gains the line about retention shipping twice: the review the build
order required for the prune found seven defects, and the merge did not wait for
it. `SPEC-v0.11.md` §13.4 has all seven.

Readiness re-measured at 6,221 against a checkout carrying the extras CI
installs, and `CLAIMS.md` needed no repointing.

Signed-off-by: arpan <contact@arpanghoshal.com>
@coderabbitai

coderabbitai Bot commented Sep 14, 2026

Copy link
Copy Markdown

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 5eb7381a-08f6-4efe-9c79-39eda6fa7f36

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@arpanghoshal
arpanghoshal merged commit 3b11975 into v0.11/3a-retention-review Sep 14, 2026
3 checks passed
@arpanghoshal
arpanghoshal deleted the release/0.11.0 branch September 15, 2026 19:15
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.

1 participant