Repository navigation
Recover Agent dependency caches and cut CI environment setup - #1297
Merged
Merged
Conversation
Contributor
|
Maple development preview: https://00523f4d.maple-ca8.pages.dev Commit: Uses development API, billing, flags and PCR configuration. Cloudflare Access applies. |
Contributor
|
Maple Auth development preview: https://994a20c3.maple-auth-dev.pages.dev Commit: Uses development services. Cloudflare Access applies. Provider OAuth is not configured for PR URLs; use the stable Auth Dev host for OAuth rehearsal. |
This branch was successfully deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Recent completed master Agent builds took roughly 32–50 minutes and rebuilt all workspace units because native cache reservations were rejected as read-only while the jobs still concluded success. Increasing configured capacity did not establish persistence. The same audit found about 24 minutes of SDK integration setup spent realizing unrelated interactive developer tooling, and a master cancellation/selection gap that could replace unfinished Agent validation with selector-only green checks.
This change:
#cishell using the same Rust/native/PostgreSQL inputs, without local container/credential tooling. Backend and SDK integration keep their migrations, OAuth and both language integration suites.Noneno longer reject admission; retained history and exclusive transitions remain closed.The audit and acceptance contract are recorded in
docs/agent-ci-performance.md. Signing, publication, PCR approval, runner/billing settings and external branch rules are outside this change. Refs #1224.Validation:
nix flake check --no-update-lock-file --max-jobs 2passed on aarch64-darwin; the checked archive-helper hash matches this commit.#cipassed.An initial unchanged packaging fixture timed out; its isolated retry and the full current Nix check passed without increasing its timeout.
Hosted cache recovery requires the first successful master seed and a subsequent actual consumer. PRs cannot publish the initial seed, so this PR may still build cold. Local fixtures, configured capacity and earlier warm runs do not establish deployed recovery or a promised speedup.
First hosted performance/validation result on exact head
1e32bd94624bad5cc0a3b05cb9ed24e7b2e03a82:Final hosted validation on the unchanged head above: 49 successful checks, 7 intentional skips, no pending or failed checks. All seven Agent lanes, Dev desktop producers and actual downloaded Dev consumers passed; Research desktop/mobile and the complete backend/Nix/SDK workflows also passed. The backend Nix/application package job passed in 12m37s, including nine actual host checks.
Mac/Windows workspace jobs passed in 45m13s / 45m07s, both genuinely cold: all 1,293 / 908 build units compiled. All nine Agent builders exercised the new scope/reader plus read-only native fallback, and all nine correctly skipped dependency-seed publishing on the PR. This validates the PR and its guard behavior; master seed publication and a subsequent actual warm consumer remain unproven. The first master run may still build cold before publishing its seed.