[MiMo] Usage percentage truncated to 0% for values under 1%

[Max0633]

Bug Description

MiMo token plan usage percentage is displayed as 0% used in the menu bar when the actual usage is between 0% and 1% (e.g., 0.96%). The Xiaomi MiMo console correctly shows 1% (rounded up), but CodexBar truncates to integer.

Root Cause

In UsageFormatter.usageLine(), the format string "%.0f%% %@" truncates the percentage to zero decimal places. For MiMo's 38B token plan, 0.96% usage (363M / 38B) is displayed as % used instead of 1% used or .96% used.

swift // Current code in UsageFormatter.swift return String(format: "%.0f%% %@", clamped, suffix)

Expected Behavior

• Values under 1% should show at least 1 significant digit
• The Xiaomi console rounds 0.96% to 1%, which is the expected UX

Suggested Fix

Option A: Round to nearest integer:
swift let rounded = clamped.rounded() return String(format: "%d%% %@", Int(rounded), suffix)

Option B: Show one decimal place when under 10%:
swift let format = clamped < 10 ? "%.1f%%" : "%.0f%%" return String(format: "\(format) %@", clamped, suffix)

Environment

• Provider: Xiaomi MiMo
• Cookie source: Manual
• Token plan: Pro (38B tokens/month)
• Actual usage: ~0.96%, displayed as 0%

[clawsweeper]

Codex review: keeping this open for maintainer follow-up; there is still a little grit to resolve. Reviewed June 24, 2026, 3:47 AM ET / 07:47 UTC.

Summary
Keep open: current main still uses zero-decimal percent labels on the MiMo and shared menu/menu-bar paths, so positive sub-1 usage can be hidden as 0%; the reported 0.96% example needs clarification because printf-style %.0f would round that value to 1%.

Reproducibility: Partly: source inspection gives a high-confidence path for positive sub-0.5 values rendering as 0% through the shared %.0f formatters, but the exact 0.96% sample should round to 1% and needs clarification from live MiMo data.

Ways to help us reproduce this

• Add a screenshot or short recording showing the behavior.
• Add expected vs actual behavior.
• Include redacted logs or terminal output.

Next step
This is narrow enough for one focused shared-formatting PR with source-level regression tests and no live provider probing.

Review details

Best possible solution:

Add one shared percent-label rule that keeps true zero as 0% but renders positive sub-1 usage distinctly across menu, menu-card, and menu-bar percent surfaces, with MiMo regression coverage.

Do we have a high-confidence way to reproduce the issue?

Partly: source inspection gives a high-confidence path for positive sub-0.5 values rendering as 0% through the shared %.0f formatters, but the exact 0.96% sample should round to 1% and needs clarification from live MiMo data.

Is this the best way to solve the issue?

Yes: the maintainable fix is a shared percent formatter used by the existing menu and menu-bar paths, not a MiMo-only special case.

AGENTS.md: found and applied where relevant.

Remaining risk / open question:

• The exact 0.96% example is not source-reproduced because %.0f should round that value to 1%, so a live redacted MiMo sample would still help confirm the reporter's observed input value.
• Percent formatting is shared across menu rows, menu-card labels, and compact menu-bar text, so the fix should avoid a MiMo-only fork and should update focused shared-formatting tests.

Codex review notes: model internal, reasoning high; reviewed against 2435c93453fe.

Label changes

Label changes:

• add P3: This is a low-risk provider display correctness issue for very small MiMo usage percentages, not an urgent workflow breakage.
• add impact:other: The issue affects maintainer-visible provider usage display accuracy, which does not fit the more specific owned impact labels.
• add issue-rating: 🐚 platinum hermit: Current issue advisory state selects this label.
• add clawsweeper:needs-live-repro: Current issue advisory state selects this label.
• add clawsweeper:queueable-fix: Current issue advisory state selects this label.
• add clawsweeper:fix-shape-clear: Current issue advisory state selects this label.

Label justifications:

• P3: This is a low-risk provider display correctness issue for very small MiMo usage percentages, not an urgent workflow breakage.
• impact:other: The issue affects maintainer-visible provider usage display accuracy, which does not fit the more specific owned impact labels.

Evidence reviewed

Acceptance criteria:

• swift test --filter UsageFormatterTests
• swift test --filter StatusItemAnimationTests
• swift test --filter MiMoProviderTests
• make check

What I checked:

• Repository policy read: AGENTS.md was read fully; its focused provider/parser test guidance and no-live-Keychain-probe guidance apply to this MiMo formatting review. (AGENTS.md:1, 2435c93453fe)
• Vision fit: VISION.md explicitly lists bug fixes and small UI/UX tweaks as merge-by-default work while favoring shared provider-driven UI. (VISION.md:1, 2435c93453fe)
• Shared menu formatter uses zero decimals: UsageFormatter.usageLine clamps the selected percent and formats it with %.0f%%, so a positive value below 0.5 renders as 0% in menu rate-window text. (Sources/CodexBarCore/UsageFormatter.swift:81, 2435c93453fe)
• Menu-bar percent path also uses zero decimals: MenuBarDisplayText.percentText uses the same %.0f%% pattern for compact menu-bar labels, so the issue is not isolated to MiMo's native menu row. (Sources/CodexBar/MenuBarDisplayText.swift:5, 2435c93453fe)
• MiMo maps token-plan fractions into usedPercent: MiMoUsageSnapshot multiplies tokenPercent by 100 and stores it as RateWindow.usedPercent, so a tokenPercent of 0.0096 becomes a 0.96 used-percent window before display formatting. (Sources/CodexBarCore/Providers/MiMo/MiMoUsageSnapshot.swift:55, 2435c93453fe)
• Existing tests miss the sub-1 display edge: UsageFormatterTests cover integer percent labels, and MiMoProviderTests cover 5.05% mapping, but there is no regression case for positive sub-1 percent display. (Tests/CodexBarTests/UsageFormatterTests.swift:30, 2435c93453fe)

Likely related people:

• debpramanik: Merged PR history for Add Xiaomi MiMo token plan usage tracking #651 shows Deb Pramanik authored the MiMo token-plan tracking work that added the affected MiMo fetcher, snapshot, docs, and tests. (role: introduced MiMo token-plan behavior; confidence: high; commits: f9a2918afcf1, e636c7e4c1cc, ecfef9598561; files: Sources/CodexBarCore/Providers/MiMo/MiMoUsageFetcher.swift, Sources/CodexBarCore/Providers/MiMo/MiMoUsageSnapshot.swift, Tests/CodexBarTests/MiMoProviderTests.swift)
• steipete: Local blame for the shared percent formatting and MiMo snapshot paths points to the v0.37.2 release baseline, and PR history shows Peter Steinberger carrying MiMo follow-up fixes and release integration. (role: recent area contributor and release integrator; confidence: high; commits: f380287041b8, ca948292c164, f5fa8138092d; files: Sources/CodexBarCore/UsageFormatter.swift, Sources/CodexBar/MenuBarDisplayText.swift, Sources/CodexBarCore/Providers/MiMo/MiMoUsageSnapshot.swift)
• AdrianSimionov: Merged PR history for feat: show Xiaomi MiMo balance breakdown and token usage #1309 shows Adrian Simionov on the recent MiMo balance and token usage rendering work that still routes through the shared percent formatters. (role: recent MiMo token usage contributor; confidence: medium; commits: ca948292c164, 0c940584aed8; files: Sources/CodexBarCore/Providers/MiMo/MiMoUsageFetcher.swift, Sources/CodexBarCore/Providers/MiMo/MiMoUsageSnapshot.swift, Sources/CodexBar/MenuDescriptor.swift)

How this review workflow works

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

[steipete]

Fixed by #1759, landed as c09a3d4.

Proof:

• one shared formatter now keeps true zero at 0% and renders every positive value below one percent as <1%
• regression coverage exercises formatter, menu card, and menu bar labels
• focused suite: 105 tests across 4 suites passed
• full local suite: 512 selections across 43 groups passed
• make check passed; autoreview reported no actionable findings
• fresh CI passed all lint, Linux build, macOS test-shard, aggregate, and security checks (one runner DNS failure was rerun successfully)

Thanks for the report, @Max0633!