Repository navigation
Fix reset detection and chart readability in 0.8.5 - #32
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5d8cf4e541
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| return (previous.ResetsAt!.Value - previous.RecordedAt - duration).Duration() <= ResetCycleTolerance | ||
| && current.ResetsAt!.Value - previous.RecordedAt - duration >= -ResetCycleTolerance | ||
| && current.ResetsAt.Value - current.RecordedAt - duration <= ResetCycleTolerance; |
There was a problem hiding this comment.
Preserve fixed cycle changes after a full-allowance sample
When a 100% observation is recorded near the start of a real cycle, its fixed reset is also one full window after RecordedAt; if a genuine early rollover or schedule change later moves the reset forward, these one-sided bounds can still classify the pair as an unused moving estimate. IsCycleChange then returns false, so the chart connects both cycles without a reset marker and downstream forecasting can treat them as continuous. Constrain this exception to evidence that the deadline is actually moving with closely spaced idle readings rather than allowing any later reset up to current.RecordedAt + duration.
AGENTS.md reference: AGENTS.md:L27-L27
Useful? React with 👍 / 👎.
Unused weekly quota can report a moving reset estimate. Treating each change as a reset produced hundreds of amber markers and fragmented the observed line. Version 0.8.5 keeps these estimates in one cycle while preserving genuine reset events, prevents overlapping chart labels and duplicate event lines, and shows the forecast basis with an accurate legend and compact token scale.
Validation: GitHub CI passed 457 x64 tests and 22 Pester checks, built both architectures and validated the self-contained Store upload without changing the source manifest. Before the version bump, 457 native ARM64 tests and 125 final targeted tests passed; both local builds and artifact preflights passed. The release-version ARM64 rerun is blocked by Smart App Control (
0x800711C7), and live UI verification requires a trusted signed build.