Skip to content

Commit dc33de6

Browse files
ryansolidclaude
andauthored
fix(signals): a landing's re-derived memo re-runs the held runs that read it stale (#4794)
At a hold's landing, a sync memo that read the screen while the hold was up (a verdict reader: `[latest(x), x()]` or `isPending` beside a held write) re-derives as a stale reader. The held runs of its readers were computed on its stale pass, and were released first, so the landing frame showed the old `x` beside the landed value for one frame (F-1, A15). The memo form tore where the inline form did not. The lane engine's landing hook (`adoptParked`) now voids those runs and re-queues the readers, so they re-run on the memo's re-derived value. An async memo's re-ask keeps the runs it reveals with. Core floor 0 B. Claude-Session: https://claude.ai/code/session_0117BoWNSSgtemsSYsaX2hXE Co-authored-by: Claude <[email protected]>
1 parent 9c917ea commit dc33de6

4 files changed

Lines changed: 81 additions & 3 deletions

File tree

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,5 @@
1+
---
2+
"@solidjs/signals": patch
3+
---
4+
5+
At a hold's landing, a memo that read the screen while the hold was up (for example `[latest(x), x()]` or `isPending` beside a held write) no longer shows its stale value for one frame. The runs of its readers that were held with the landing now re-run on its re-derived value.

0 commit comments

Comments
 (0)