Skip to content

fix(ci): declare the e2e shard inputs on the no-codeql orchestrator - #714

Merged
0xPlayerOne merged 1 commit into
mainfrom
fix/no-codeql-e2e-inputs
Oct 3, 2026
Merged

0xPlayerOne merged 1 commit into
mainfrom
fix/no-codeql-e2e-inputs

Conversation

@0xPlayerOne

@0xPlayerOne 0xPlayerOne commented Oct 3, 2026 •

Copy link
Copy Markdown
Owner

What

validation-no-codeql.yml declares and forwards e2e-shard-list / e2e-total-shards, matching the main orchestrator. The receipt tests now pin that parity for both orchestrators.

Why

Since v1.42.0, the sync renderer always passes the two e2e shard inputs to the reusable validation call (the #697 sharding defaults). The with-codeql orchestrator declares them; the no-codeql variant does not — and GitHub rejects unknown reusable-workflow inputs at startup. Every codeql: false consumer therefore fails every ready-PR validation with startup_failure:

consumer caller evidence
adea-ai/agent-sim validation-no-codeql.yml@v1.42.0 8/8 non-skipped Validation runs since 10-02 are startup_failure; zero successes ever
pinkbinder/blog-pipeline validation-no-codeql.yml@v1.44.3 startup_failure on the branch-caller run during today's upgrade
NiftyLeague/nifty-world validation-no-codeql.yml@v1.44.3 same render path
NiftyLeague/nifty-playfab-configs validation-no-codeql.yml@v1.44.3 same render path

The failure class was masked while those repositories were billing-paused (jobs gate on CI_BILLING_PAUSED, and startup failures don't surface in a paused rollup). Consumers need their caller uses: ref bumped to the release that carries this fix.

Rendered callers always pass e2e-shard-list and e2e-total-shards (the
v1.42.0 sharding defaults), but the no-codeql orchestrator never
declared them, so every codeql:false consumer's pull-request validation
failed at startup with an unknown-input error. GitHub reports that as
startup_failure, which the billing-pause era hid: agent-sim's
Validation lane has never once succeeded since its v1.42.0 pin.

Declare both inputs and forward them to the test lane, matching the
main orchestrator, and pin the parity in the receipt tests.
@0xPlayerOne 0xPlayerOne changed the title fix: no-codeql-e2e-inputs fix(ci): declare the e2e shard inputs on the no-codeql orchestrator Oct 3, 2026
@0xPlayerOne
0xPlayerOne marked this pull request as ready for review October 3, 2026 20:21
@0xPlayerOne
0xPlayerOne merged commit 7500208 into main Oct 3, 2026
28 checks passed
@0xPlayerOne
0xPlayerOne deleted the fix/no-codeql-e2e-inputs branch October 3, 2026 20:27
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