Jean-Luc Makiolaandmakiolaj
17b3753ad3
Step 8 of the simplification review: the three behaviour-sensitive merges. Each lands behind a characterization test written and passed against the old code first. **Month row seating** - `MonthWeek.seating(col, laneCap)` is the one source for which event sits in which lane and what overflows. `laneEvents`, `chipAt`, `overflowEvents` and MonthWeekRow's inline copy all build on it, so dots, bars and pills can't drift apart (#53). - MonthWeekRow computes the seating once per row (`remember(week, laneCap)`) instead of re-filtering every column on each recomposition. - `OverflowDots.total` and `shownLanes` are gone; they always equalled `events.size` and `laneCap`. - MonthSeatingTest pins the four old copies against each other over 400 generated weeks × lane caps 0–6; they agreed everywhere, and `seating()` matches them. **Timeline resize** - TimelineDragController's eight loose `resize*` fields are one `ResizeSession?`, so cancel clears it in one assignment. Moves mutate it in place; no per-frame allocation. None of the fields were Compose state. - Only difference: a resize starting without the previous one ending used to inherit the old held slot; it now starts clean. **Timeline paging** - Week, multi-day and day page through one `DaysGrid` with `weeks(firstDay)`, `days` and `startingAt(date, count)` factories; the PageGrid interface, WeekGrid and DayGrid are gone. Page numbers and grid equality (used by pager keys) are unchanged; PageNumberingTest pins them. - jumpToDate no longer detects the multi-day view by class; it re-frames when the target would open mid-page, which a test checks against the old rule for every in-range date. Not yet checked on a device. Worth a look: month bars/pills/+N dots on crowded days in paged and split styles plus the split morph, chip taps and month drag; resize of both edges incl. across midnight, cancel then plain drag, RTL; paging, view switching, jump to date/today, week-start and multi-day count changes. Co-authored-by: Jean-Luc Makiola <business@jeanlucmakiola.de> Reviewed-on: https://codeberg.org/jlmakiola/calendula/pulls/418