• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

solidjs / solid / 37338342306
76%
main: 89%

Build:
Build:
LAST BUILD BRANCH: fix/single-flight-await-streamed-slices
DEFAULT BRANCH: main
Ran 05 Oct 2026 04:12PM UTC
Jobs 1
Files 8
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

05 Oct 2026 04:07PM UTC coverage: 75.991%. Remained the same
37338342306

push

github

web-flow
fix(signals): two crash-class bugs from the semantic fuzzer — a lane node in a held frame stays the lane's; a verdict reader revealing an orphan flight holds it (#3798)

* fix(signals): a lane's node inside a held frame stays the lane's (fuzzer F7a/F7b)

`holdFrame` holds a parked pass's frame with the transaction (ruling A).
It re-listed a lane's node found there — a verdict reader created in
the frame, broken out into the holder's verdict lane (display-ahead,
L2 rule 6), or a derivation of a guess — on the frame's transaction,
still carrying its lane value (CONFIG_OVERRIDE). At the landing
`dissolveLane` no longer found the node (its `_transaction` was the
transaction's) and `land` nulled the transaction: a lane node with no
lane, and the landing's re-derivation tripped over it (`laneRead` →
`txOf(null)` → TypeError, REACTIVITY_HALTED).

The shapes: a `Loading` whose content reads `latest(x)`, unmounted and
remounted while `x` is held by an action (F7a); a memo of
`isPending(() => x())` mounted in the tick of the write it probes, its
mount toggled off and on while the write is held (F7b). The remounting
gate's write joins the hold (A34 (1)), so its new frame is held with it
— and the verdict-lane nodes inside it were taken along.

`holdFrame` now skips a child listed on a lane, as the seam's own hold
loop already does for this flush's stagings: lane work never makes its
node transaction work (A15, #3698), and a lane pass's frame is the
lane's, revealed and retired at its seam. The lane's end (`dissolveLane`
at the parent's landing) commits the node as before. Every other
`holdNode` site (the seam loop, the store's three) already excluded
lane work; this was the one that did not.

F7a and F7b flip from `it.fails` to `it`. No other pin moved.

Co-authored-by: Claude via Cursor <noreply@cursor.com>

* fix(signals): a verdict reader revealing a flight nobody holds registers it with the frame (fuzzer F12)

A verdict reader (`latest`/`isPending` in its pa... (continued)

649 of 925 branches covered (70.16%)

Branch coverage included in aggregate %.

962 of 1195 relevant lines covered (80.5%)

27.78 hits per line

Jobs
ID Job ID Ran Files Coverage
1 37338342306.1 05 Oct 2026 04:12PM UTC 8
75.99
GitHub Action Run
Source Files on build 37338342306
  • Tree
  • List 8
  • Changed 0
  • Source Changed 0
  • Coverage Changed 0
Coverage ∆ File Lines Relevant Covered Missed Hits/Line Branch Hits Branch Misses
  • Back to Repo
  • Github Actions Build #37338342306
  • b0bad026 on github
  • Prev Build on next (#37287144634)
  • Next Build on next (#37351839761)
  • Delete
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc