fix(status): expose verified graph readiness - #760
Conversation
|
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 168a569b2d
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| return Some(if text_generation_present { | ||
| CodeGraphServingReadinessV1::Pending |
There was a problem hiding this comment.
When graph activation fails non-retryably (for example, CodeGraphProjectionError::Corrupt), the worker keeps the text generation but leaves serving_generation empty and schedules no activation retry; this branch therefore reports Pending indefinitely solely because text exists. That hides the terminal failure and makes strict dogfood poll until its global timeout instead of exposing an actionable unavailable/refused state. Preserve the activation failure reason or only return Pending while an activation attempt or retry is actually outstanding.
AGENTS.md reference: AGENTS.md:L167-L169
Useful? React with 👍 / 👎.
168a569 to
f9fdcb7
Compare
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f9fdcb7041
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| } | ||
| }); | ||
| }; | ||
| Some(latest.code_graph_serving_readiness()) |
There was a problem hiding this comment.
When generation A is already seated and reconciliation seals B, the worker installs B as text_generation and drops the reconciliation guard before graph activation, but leaves A in serving_generation until B's later swap. This call therefore returns A's Ready; because the scheduler has already verified B and cleared its hints, status can report fresh/current with code_graph_serving: ready while B's graph is still pending. Compare the serving and current text-generation identities and report Pending on a mismatch so the new typed state describes the current sealed generation rather than its predecessor.
AGENTS.md reference: AGENTS.md:L9-L12
Useful? React with 👍 / 👎.
081d58d
into
ScriptedAlchemy:codex/tracedecay-total-redesign-plan-reopened
2 participants
Summary
Why
graph_statistics.state=observedproves a sealed generation census exists, but it does not prove exact verified graph reads can open. Cold dogfood could therefore declare readiness while activation was still pending or refused, then fail its first context read.Verification
cargo fmt --all -- --checkgit diff --check upstream/pr-707...HEADcontracts:checkcode_graph_serving=readyat 603sStandalone macOS portability still hits the zombie-process regression already isolated in #756; this change does not touch that process probe.
Supports #707.