Skip to content

docs(win-time-sys): confirm WT-D-7 and WT-D-8, and complete the checklist - #165

Merged
MikeGrier merged 4 commits into
mainfrom
mikegrier/wt-3-1-confirm-wt-d-7
Oct 22, 2026
Merged

MikeGrier merged 4 commits into
mainfrom
mikegrier/wt-3-1-confirm-wt-d-7

Conversation

@MikeGrier

Copy link
Copy Markdown
Owner

Records the engineer's review of the two decisions the assistant made while building win-time-sys, and completes the crate's checklist.

  • WT-D-7 (the timeline and clock shapes) is confirmed in full, with the addition that no checked_ method ever panics. That was already true; it is now stated in the decision and in TimePoint's docs, and held by a test that runs every checked method, to_duration, abs_duration and from_duration over the extremes of five tick rates. Making checked_since panic fails it.
  • WT-D-8 (the clock_costs example) is confirmed in full. The engineer ruled that no windows-platform-probes probe and no machine-readable row is wanted, so the decision schedules no work.
  • WT-inf.1 and WT-inf.2 move out of CHECKLIST.md into PARKED.md as a pure relocation, because parked work does not belong in the checklist.
  • The checklist is archived and recorded in COMPLETED-PLANS.md.

Three docs: commits, so merging triggers no release. The only code change is a doc comment and a test in win-time-sys.

Mike Grier and others added 3 commits October 22, 2026 18:04
… hold that no checked method panics

The engineer confirmed WT-D-7 in full on 2026-10-22 ("confirm all, the checked should not panic"). The decision row now records the confirmation, and that it was given knowing that a timeline's tick rate belongs to its type and that truncation in from_duration can leave a timeout less than a tick short.

No checked_ method panics was already true by inspection. It is now stated in TimePoint's documentation and held by a test that runs every checked_ method, to_duration, abs_duration and from_duration over the extremes of five periods, including the largest and smallest a NonZeroU64 allows. Making checked_since panic fails it.

Completed item: WT-3.1: The engineer confirms or amends WT-D-7, from WT-1.2: Clock::now taking &self; a point a u64 tick count and a distance an i64 one; ticks_per_second returning NonZeroU64; conversions to and from Duration truncating towards zero; and panicking operators beside checked_ methods.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Pure relocation of WT-inf.1 and WT-inf.2, unchanged, out of CHECKLIST.md: parked work lives in PARKED.md and never in the checklist a contributor reads to choose the next work. PLANS.md gains the parked row and loses the sentence that described the parked items inside the checklist's row.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
… complete win-time-sys's checklist

The engineer confirmed WT-D-8 in full on 2026-10-22 and ruled that no probe in windows-platform-probes, and no machine-readable row from clock_costs, is wanted ("I don't need to have a probe for any of this; otherwise, confirm all"). The decision's row records both, and that it schedules no work. WT-M3 is archived with its item, its stubs deleted, and the checklist moves to COMPLETED-PLANS.md; the parked items stay in PARKED.md.

Completed item: WT-3.2: The engineer confirms or amends WT-D-8, from WT-1.6: the cost probe shipped as this crate's clock_costs example rather than in windows-platform-probes, and reporting what it observed -- least, median and most cost per read, and the smallest step -- with no ranking, no verdict, and no figure quoted in the crate's documentation.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Copilot AI balanced review requested due to automatic review settings October 22, 2026 22:24

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟢 Approval recommended

Production behavior is unchanged, and the remaining documentation corrections are non-blocking.

3 open issues
What changed in this PR

Records confirmation of the win-time-sys design decisions, documents the non-panicking checked-arithmetic contract, and closes the completed plan.

Changes:

  • Confirms WT-D-7 and WT-D-8, including the decision not to add further probe output.
  • Adds boundary-value regression coverage for checked arithmetic and duration conversions.
  • Moves deferred work to PARKED.md and archives completed review items.
File Description
crates/​win-time-sys/​src/​timeline/​tests.rs Adds non-panicking boundary-value coverage across tick rates.
crates/​win-time-sys/​src/​timeline.rs Documents the checked-method guarantee.
crates/​win-time-sys/​PLANS.md Links completion records and registers parked work.
crates/​win-time-sys/​PARKED.md Holds the relocated deferred items.
crates/​win-time-sys/​DESIGN-NOTES.md Records confirmation of both design decisions.
crates/​win-time-sys/​COMPLETED-PLANS.md Records the completed checklist.
crates/​win-time-sys/​COMPLETED-CHECKLIST.md Archives the engineer’s review and rationale.
crates/​win-time-sys/​CHECKLIST.md Replaces completed items with archive and parked-work pointers.

Review effort: Balanced

Comment thread crates/win-time-sys/COMPLETED-PLANS.md
Comment thread crates/win-time-sys/DESIGN-NOTES.md Outdated
Comment thread crates/win-time-sys/PARKED.md
… at the root, and correct what the step measures

Review of the confirmation found three documentation gaps. The repository-wide PLANS.md still listed the win-time-sys checklist as in progress and had no row for PARKED.md, and COMPLETED-PLANS.md had no row for the completed checklist; they now match the component's own indexes. The working-position section's link to WT-inf.2 still named CHECKLIST.md after the item moved, and now names PARKED.md.

WT-D-8's row, and the archive entry that restated it, called the smallest step the system timer resolution. It is the smallest difference between readings that differ, sampled within a time budget, so it depends on how often a clock updates and on which updates the sampling saw, and a missed update can make it span several ticks. The row now says that, and the archive entry points at the row instead of restating it. Swept the timer-resolution claim: 3 mentions in win-time-sys's markdown, all corrected, none elsewhere.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Copilot AI balanced review requested due to automatic review settings October 22, 2026 22:32

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟢 Approval recommended

The documentation and regression coverage match the existing behavior, prior feedback is addressed, and no blocking issues remain.

No open issues 🎉

3 resolved since last review

Review effort: Balanced

@MikeGrier
MikeGrier merged commit 47d7cfd into main Oct 22, 2026
43 checks passed
@MikeGrier
MikeGrier deleted the mikegrier/wt-3-1-confirm-wt-d-7 branch October 22, 2026 22:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants