Repository navigation
feat(bench): configure and compare candidate component switches - #404
Merged
Merged
Conversation
…chmark-switches-296
…chmark-switches-296 # Conflicts: # docs/CHANGELOG.md
Haoming9527
marked this pull request as ready for review
October 24, 2026 17:20
This was referenced Oct 24, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Benchmark setup can now select candidate components through repeatable
--switchflags for hide-official, PTC, panel, discovery and shell. It preserves inherited literal settings, refuses conflicting local values without--force, and refuses dynamic or unsafe patch shapes. Baseline profiles receive no component-switch writes. Dry-run prints the plan without launching the host or writing profiles.Real setup verifies composed configuration and requires a successful bounded headless readiness turn before reporting success; READINESS_BLOCKED and other readiness failures point to #207 before any sweep starts. This turn may call the configured model. Effective switch facts are recorded in summaries and sealed
arms.cand.switches, displayed in reports and the offline panel, and compared between candidate runs. Different candidate sets fail comparison with exit 2; baseline's empty set is expected. Unevaluated values remain unknown.Fixes #296. Stacked on #402 (
work/resolve-294-model-matrix) to keep this change under the implementation-size limit; #402 includes #399 and #381. Includes the reviewed #401 workflow change so PRs targeting feature branches run CI. Merge prerequisites first. #303 administrator enforcement and #207 Desktop validation remain open; this PR claims neither.Validation:
8481d91passed every executed check, including Windows, Linux, macOS, Node-floor and CI required. The all-target trigger ran successfully against the stacked feature-branch base. Conditional smoke was skipped as designed.