You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 0847122
Browse filesBrowse the repository at this point in the historyBrowse files
fix(signals): mounts under one boundary keep their own holds across flushes (#4302)
* fix(signals): keep separate-flush mounts under one boundary in independent holds
A boundary's tree and output are structure between the content and the
render effect that shows it, not a user derivation of the content (D-3).
They now carry CONFIG_FRAME, the bit render effects get at creation, and
read a hold they are not part of as committed (re-derived at its landing);
a re-pass outside its hold's flush publishes mainline instead of
re-joining that hold. Two <Show>s under one <Errored>/<Loading>, mounted
in separate flushes over independent holds, no longer merge them
(S-5 rows 1'-3', D-4). An `on` boundary's output keeps joining (D-13).
Fixes#4283
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_015hPxMkC3jg3Uivh65gvaf2
* fix(signals): no floor cost for the boundary frame bit; keep `on` boundaries independent
Render effects carry CONFIG_FRAME from creation (`type << 23`), so the
three frame checks in core test one bit instead of `_type` plus a bit:
the floors get 6 B smaller minified and pay nothing for boundaries.
A frame re-pass published mainline clears CONFIG_HELD whichever frame
reader it is (render effect, boundary structure, verdict memo): its
readers read the value it publishes.
A boundary's output is structure except while a re-arm's swap is the
change's work: `arm` clears the bit, and the output's pass sets it again
once it is neither armed nor held. `<Loading on>` around two mounts made
in separate flushes now keeps their holds independent too (D-13 re-arm
tests unchanged). The regression test covers both landing orders and an
`on` boundary that was re-armed earlier.
Refs #4283
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_015hPxMkC3jg3Uivh65gvaf2
---------
Co-authored-by: Claude <[email protected]>
Two mounts under one boundary, made in separate flushes, keep their holds independent (S-5 rows 1′–3′, D-4, #4283). A boundary's tree and output are structure, not a user derivation of its content: like the render effect that flattens the same children without a boundary, they read a hold they are not part of as committed and are re-derived at its landing, and a re-pass of one outside its hold's flush publishes mainline instead of re-joining that hold and merging it with the next flush's. `<Errored>` and `<Loading>` (with or without `on`) around two `<Show>`s, each over its own hold, now reveal each branch at its own hold's landing. A re-arm's swap is still the change's work: while it is, the output joins the change (D-13).
0 commit comments