solidjs / solid / 41317659710

78%

main: 89%

push

github

fix(signals): effects a guess mounts honor defer and run once per write (#4557) (#4567)

* fix(signals): effects a guess mounts honor defer and run once per write (#4557)

A deferred effect born in an optimistic lane's pass voids the first run the
pass queued for the lane's reveal, as a plain or held mount skips it (#4543).

A node that leaves a shown lane marks LEFT_SHOWN so the seam wakes readers
still showing its lane slot (#3892). When the leaving pass already notified
its readers (a change against the committed value), an effect reader had
re-run on the new answer, and the seam's wake ran it again with nothing new.
The seam now skips an effect whose pass started after the node's.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015hPxMkC3jg3Uivh65gvaf2

* fix(signals): skip the seam's left-lane wake once the node notified (#4557)

The seam's LEFT_SHOWN wake re-ran readers the leaving pass had already
notified. laneStage clears the node's notify stamp with the flag; the seam
wakes only if nothing notified since. Replaces the per-reader time check
(-45 B minified on lane rows). isUser drops its double negation.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015hPxMkC3jg3Uivh65gvaf2

* fix(signals): mark a left lane slot on the notify stamp (#4557)

LEFT_SHOWN moves from a pass flag to a sentinel on `_notifiedAt`: insertSubs
overwrites it, so a pass that notified clears the mark by itself and the
seam wakes only readers still showing the slot. The first queued run of a
user effect is an arrow, not a bound function.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015hPxMkC3jg3Uivh65gvaf2

* fix(signals): keep effect()'s first-run runner and isUser as they were (#4557)

Only the defer line changes in effect(); the arrow and isUser tweaks grew
hello world's brotli.

Co-Authored-By: Claud... (continued)

797 of 1098 branches covered (72.59%)

Branch coverage included in aggregate %.

1129 of 1367 relevant lines covered (82.59%)

29.96 hits per line