Repository navigation
Conversation
sabinem
marked this pull request as draft
September 15, 2026 08:08
CI already started Postgres, Keycloak and the backend to run the seed, then tore the whole thing down without asking it a single question. The suites under internal/** run on in-memory SQLite behind a mock keyfunc over bufconn, so three things production depends on were never exercised anywhere: the Postgres schema the migration actually produces, the token exchange with Keycloak, and the gRPC server on a real port. Six specs now run against that stack before it is discarded: health over the socket, a genuine Keycloak token resolving to the seeded user, the seeded fixture read back out of Postgres, a hackathon written and read back, and two refusals asserted on the code rather than merely on failure — PERMISSION_DENIED is a clean 403 in the frontend, INTERNAL is a 500, and a test that accepts any error cannot tell them apart. The stack boot was already paid for, so this adds well under a second.
Both recipes returned as soon as they had asked for the backend to be scaled back up, so dev could still be starting, or failing to start, when they reported success. _backend-up, the counterpart to _backend-down, now scales up and waits for the rollout, and its failure fails the recipe even from the EXIT trap, which otherwise keeps the exit status it found.
check::smoke defaults to the local stack. cluster::smoke points it at dev: it waits for the backend, port-forwards svc/hackagon-backend, and takes the Keycloak address and realm from the backend's own issuerurl, so tokens always come from the issuer the backend trusts. A failure names the usual causes on dev: an unseeded database, missing fixture users or a different password, or direct access grants switched off.
sabinem
force-pushed
the
test/ci-smoke-suite
branch
from
October 8, 2026 12:38
b2ec2dd to
b2e7cad
Compare
Collaborator
Author
|
close for now. |
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
1 participant
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.
CI already started Postgres, Keycloak and the backend to run the seed, then tore the whole thing down without asking it a single question. The suites under internal/** run on in-memory SQLite behind a mock keyfunc over bufconn, so three things production depends on were never exercised anywhere: the Postgres schema the migration actually produces, the token exchange with Keycloak, and the gRPC server on a real port.
Six specs now run against that stack before it is discarded: health over the socket, a genuine Keycloak token resolving to the seeded user, the seeded fixture read back out of Postgres, a hackathon written and read back, and two refusals asserted on the code rather than merely on failure — PERMISSION_DENIED is a clean 403 in the frontend, INTERNAL is a 500, and a test that accepts any error cannot tell them apart.
The stack boot was already paid for, so this adds well under a second.