test(sdk): wait for configured providers - #46614
Closed
kitlangton wants to merge 1 commit into
Closed
Conversation
kitlangton
force-pushed
the
sdk-provider-readiness
branch
from
September 1, 2026 15:44
7cc7462 to
b11674c
Compare
kitlangton
force-pushed
the
sdk-provider-readiness
branch
from
September 1, 2026 19:39
b11674c to
b177cde
Compare
Contributor
Author
|
Superseded: #46639 rewrote the same generate/websearch tests to wait for the configured provider, so this adaptation is no longer needed. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
Existing SDK generation and websearch tests assume catalog operations wait for plugin initialization. Upstream made these operations nonblocking, so the old forced supervisor gate and immediate query can fail before the configured provider is available.
What Changes
The two generation cases wait for their exact enabled inline model, then still assert explicit/default selection, the generated response, and exactly one model request. The websearch case waits for its provider in the same Location before asserting the query result.
Remove the obsolete supervisor gate rather than restoring blocking production behavior. Readiness waits are bounded and observe actual provider inventory, not a fixed sleep.
Scope
Test-only adaptation extracted from #46496, following the same nonblocking-catalog change addressed for Server fixtures in #46567. This does not depend on session-selected SDK instances or strict Core activation waits.
The former shared build prerequisite landed independently in
43d09b9d75; this PR now targetsv2directly.Verification
Using Bun 1.4.0, from
packages/sdk:All three affected cases failed at runtime before the change. Afterward they passed 15/15 repeated runs. Full SDK suite: 25 passed, 1 failed. The failure is the unchanged cancellation assertion at
test/promise.test.ts:73: the aborted iterator completes instead of rejecting withClientError/Transport. That assertion was also reproduced before these edits.These results were reproduced after rebasing onto
v2ate2e82f18e2. SDK typechecking and diff checks passed; range-diff confirms that the fixture patch is unchanged by the rebase. The normal pre-push workspace typecheck passed all 33 tasks.At the rebased head, run
33550743501passed Windows CI and the Linux unit suites. Linux failed only the generated-documentation check: the website OpenAPI copies in basev2omit the diagnosticrefadded in #46594. This test-only PR does not change those documents or their generator inputs.