{{ message }}
Repository navigation
Conversation
Both are inspection tests red on main with their guarded wiring intact: - test_parallel_slots_per_load pinned llama_backend.load_model, but _load_model_impl loads through its backend-dispatch local since the MLX split, so the ordering chain never found its last link. Pin backend.load_model; the resolve < dedupe < guard < load ordering it proves is unchanged. - test_video_attachment_part scanned a 400-char window for the video and audio modality votes; added comments and the widened _needs_image derivation pushed them past the window. Widen to 600 — the votes are still exactly there.
Contributor
Author
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Last two of the Backend CI red families on main (inventory in #9949; siblings #9956, #9957). Both are source-inspection tests whose guarded wiring is intact — the pins just point at pre-refactor spellings:
test_parallel_slots_per_load.py::test_route_resolves_slots_once_before_dedupe_guard_and_load— pinnedllama_backend.load_model, but_load_model_implloads through its backend-dispatch local (backend.load_model) since the MLX split in studio: serve installed mlx models through the openai-compatible api #8768, so.index()throwsValueError: substring not found. The ordering chain the test proves (resolve < fast-dedupe < intent < resolved-dedupe < training-guard < load) is unchanged, verified by position against the current source.test_video_attachment_part.py::test_video_joins_the_projector_requirement_before_switching— scanned a 400-char window after_needs_image = bool(_pre_parsed[2])for the audio/video modality votes; explanatory comments and the widened_needs_imagederivation pushed the votes to ~430+ chars. Widened to 600; both votes are exactly where the test expects, feeding the same_needs_vision.Test-side only.
Verification
Both files green locally: 41 passed, 1 skipped (was 2 failed). Ruff clean.