solidjs / solid / 41544689872

78%

main: 89%

push

github

fix(store): a projection leaf written by a lane pass is the lane's (#4631, #4629) (#4633)

* fix(store): a projection leaf written by a lane pass is the lane's (#4631)

A projection's commit adopts the derive's result through reconcile, whose
leaf writes went straight to setSignal. On a lane pass only the container
entered the lane (stageOn -> holdWithDerive); a leaf a transaction held
(`length`, after a truth landing staged beneath a newer guess) took the
lane's value into the hold's staging instead, so the screen paired the
held length with the lane's rows: the closed row stayed, or read as
undefined. Adoption leaf writes now route through holdWithDerive like the
draft path's notifyWrites.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0117BoWNSSgtemsSYsaX2hXE

* test(signals): pin #4629's second guess through a keyed projection

The same fix covers #4629: the projection's leaf write in the lane pass is
the lane's, so action 2's guess shows when action 1's re-ask lands, as the
overlay memo does. Test from #4632.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0117BoWNSSgtemsSYsaX2hXE

* chore: mention #4629 in the changeset

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0117BoWNSSgtemsSYsaX2hXE

* docs(signals): regenerate the rules index

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0117BoWNSSgtemsSYsaX2hXE

* test(signals): pin #4629 frame (4) as a D-26 known deviation

The second guess should show at once beside the screen's issues (a plain
read in lane work of a parent-held node is the screen's value, D-26); today
it waits for action 1's re-ask. Pinned it.fails with a D-26 status note.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0117BoWNSSgtemsSYsaX2hXE

--------... (continued)

803 of 1104 branches covered (72.74%)

Branch coverage included in aggregate %.

1131 of 1369 relevant lines covered (82.62%)

30.1 hits per line