solidjs / solid / 39705717517

78%

main: 89%

push

github

fix: optimistic release without torn frames (sink slot records, keyed guess fold) (#4210)

* fix(web): a sweep's slot re-emissions join the sweep's unit

Each arg binding of a slot occurrence re-sends the whole record as it
sweeps, so the first copy carried the other args' previous values and
went out alone, ahead of the holes the same commit moved. The chat's
outbox dropped a pending message before the thread showed its row.
Inside a sweep the record now joins the sweep's unit like hole and attr
records (C13): the pass's last re-emission stands and leaves with them.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015SnZadCVdDSeDzxBUhcMU9

* fix(signals): a keyed arrangement guess folds a correction instead of voiding it

Two overlapping actions pushing onto one keyed optimistic store: when the
stream echoed the first write while the second was open, the correction
dissolved the lane. For that window the store read torn (a hole at the
first row's index, `length` from the guess, a raw target served for a row),
and at the commit the confirmed row remounted twice, which destroyed an
open editor and its draft in the board example.

A keyed arrangement guess is rows by key, so a current truth that differs
now folds into the hold beneath it (D-34, amended): the guess stays the
lane's value and the screen's until the hold commits, and rows match by key
there. Deriving passes outside the lane read the hold's world (each guess's
truth, else what it covered) and join the hold, as for any held truth
(#3164); `latest()` shows it; `until`/`resolve` keep their postures. Keys
born under a folded guess are held with it. Lane work in a frame that
already has a transaction reads that frame's stagings as the screen, as a
verdict reader does, so a list pass never pairs the truth's length with the
guess's rows.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://... (continued)

762 of 1055 branches covered (72.23%)

Branch coverage included in aggregate %.

1105 of 1349 relevant lines covered (81.91%)

27.53 hits per line