Repository navigation
fix(provider): show diagnostics for failed runtime verification - #38185
Closed
primevishnu wants to merge 1 commit into
Closed
primevishnu wants to merge 1 commit into
primevishnu wants to merge 1 commit into
Conversation
Contributor
|
Closing: primevishnu already has 3 or more open PRs here, which is the limit for contributors who are not vouched. Once one of them is merged or closed, open this again. See the open PR limit. |
Contributor
ApprovabilityVerdict: Needs Human Approval Macroscope's review found this PR requires human approval — The change adds bounded, redacted runtime-verification details to provider setup. It also introduces a file-level You can add or adjust custom eligibility rules. Learn more. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
Managed Codex and Antigravity runtime verification can fail or time out while setup shows a generic error and tells the user to check server logs. Antigravity retains startup stderr internally but does not send it with the timeout's installation state. Codex drains and discards validation stderr. A remote client cannot recover output that the server discarded.
For example, a child that writes
Startup library missingand then hangs currently produces a timeout message without that diagnostic. The expected behavior is to retain a bounded, redacted diagnostic for that failed attempt and make it available where setup failed.Change
Capture happens in the environment running the validator. Existing environment-scoped RPCs deliver the bounded result to local, SSH, direct-remote, and relay clients; there is no new arbitrary log-file reader. Older server messages remain accepted because diagnostics are optional on the wire. Runtime platform support is unchanged.
Scope
This follows #37896 by fixing loss of the startup diagnostic behind its verification-timeout message. There is no separate issue; the failure is reproduced below. The server, contract, and client changes all serve the same failed-installation result.
Related work: #31050 adds Antigravity's exit code/signal to its summary message; #35021 preserves readiness failures after CLI updates. This change carries per-attempt managed-runtime verification diagnostics to setup clients. It does not diagnose every provider failure or infer a cause when the runtime supplies no diagnostic.
Verification
Recorded on Windows with Node 26.13.1, at head
cceabac9e1e4f07515dda320918ccb7708c02bec:Before the fix, three installer regressions fail because installation state lacks diagnostics: an Antigravity initialization timeout and Codex cached-update/reverification timeouts. After the fix, startup text survives the timeout, credentials are redacted, retry/cancel clears details, and the previous installation remains active.
Antigravity installer and capture suites: 42 passed, 6 existing platform skips. Selected portable Codex installer cases: 5 passed. Contract/auth compatibility suite: 3 passed. These are 50 focused passing cases; the final targeted installer/capture recheck also passed 11 cases.
Capture tests cover continued drainage after truncation, split UTF-8, oversized lines, incomplete quoted secrets, credential headers, sensitive environment values, home paths, typed spawn failures, and a real child process.
A separate credential-free runtime driver used real Node subprocesses and the actual capture helper:
Server and web typechecks passed independently with exit code 0. Changed-file lint, formatting, and
git diff --checkpassed. The branch merges cleanly with upstreambee9c3b7ingit merge-tree.Verification limits: the browser runner could not connect to its Chromium session, so UI assertions and before/after screenshots were not completed. Native mobile typecheck is blocked by missing Expo/React Native dependencies; no device was exercised. macOS/Linux execution and live remote deployment were not performed. The required
check:changed --concurrency 1 --timeout 30attempt passed formatting, timed out on several workspace steps, and failed affected-workspace typecheck on missing marketing Astro setup; it is not a full workspace pass. The runtime examples use synthetic fixtures, not a production provider outage. UI diagnostic state is not persisted across server restart; the bounded server-log record follows existing logger retention.AI-assisted implementation and verification: GPT-6.1 Sol through Codex in T3 Code.