fix: prefer five-hour usage in composer indicator - #1201
Merged
Merged
Conversation
Model: gpt-6.1-sol
Model: gpt-6.1-sol
wibus-wee
marked this pull request as ready for review
October 1, 2026 14:45
wibus-wee
marked this pull request as draft
October 1, 2026 14:46
wibus-wee
marked this pull request as ready for review
October 1, 2026 14:46
Member
Author
|
@copilot refine note |
Co-authored-by: wibus-wee <62133302+wibus-wee@users.noreply.github.com>
Contributor
Updated the bilingual note to clarify the reproduced baseline test failure and fixed the Chinese PR link spacing. Commit: |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Related issue
Problem / pressure
When a provider reports weekly usage of 29% and five-hour usage of 11%, the composer shows 29% because its indicator reads the first window after the detail popover sorts by descending duration.
Summary
Select the five-hour window independently for the compact indicator. Keep the existing longest-window fallback, context/compacting priority, and detail order. Add behavioral regressions and the bilingual display Spec and decision note.
Visual explanation
Before / after
Test plan
Validation used an independent clone with initialized ACP submodules because the authoring worktree has no dependencies.
pnpm format, i18n/import/platform/public-boundary guards, andpnpm run docs check --base 93545f01b: passed.pnpm check: stops at one existing component failure,boot-shell.test.tsx→ “falls back to the mark alone when storage is unavailable”; 4592 component tests passed. Reproduced after restoring the original usage component on Node 26.10.0.SessionUsagePopover/QuotaOnlystory supplies both windows.