fix(ocap-kernel): report a dead run loop instead of a healthy kernel - #1005

Merged

sirtimid merged 15 commits into

mainfrom

sirtimid/detect-run-loop-death

Aug 6, 2026

Commits on Aug 5, 2026

fix(ocap-kernel): report a dead run loop instead of a healthy kernel

sirtimid
and
claude
committed
6678692
fix(ocap-kernel): roll back the crank the run loop died in

sirtimid
and
claude
committed
4fb9ca6
fix(ocap-kernel): report run loop health sampled after the crank wait

sirtimid
and
claude
committed
ead5e17
fix(ocap-kernel): guard remote ingress at the boundary, not on the queue

sirtimid
and
claude
committed
9c9a0c0
fix(ocap-kernel): call the run loop failure handler off a local, not off this

sirtimid
and
claude
committed
7aab306
docs: Update changelogs
sirtimid
committed
f793d8f
test(ocap-kernel): cover the run-loop rollback and crank bookkeeping gaps

sirtimid
and
claude
committed
d95b710
fix(ocap-kernel): name the real error when an abort rollback fails

sirtimid
and
claude
committed
4d1f5d2
test: cover crank rollback against real SQLite and runLoop on the wire

sirtimid
and
claude
committed
8455163
docs: say that a failed kernel is not recoverable in place

sirtimid
and
claude
committed
be4fb35
fix(kernel-store): close the transaction a failed savepoint rollback abandoned

sirtimid
and
claude
committed
c3eb0a0
fix(kernel-cli): remediate a dead run loop before logging it

sirtimid
and
claude
committed
29affc3
fix(ocap-kernel): guard the last ingress points and carry the cause to the panel

sirtimid
and
claude
committed
0293803
fix(kernel-cli): tear the kernel down when startup aborts, and terminate

sirtimid
and
claude
committed
3af30a6
fix(kernel-cli): log best-effort in front of every fatal exit

sirtimid
and
claude
committed
b5c129d

Navigation Menu

Uh oh!

fix(ocap-kernel): report a dead run loop instead of a healthy kernel - #1005

fix(ocap-kernel): report a dead run loop instead of a healthy kernel#1005
sirtimid merged 15 commits into
mainfrom
sirtimid/detect-run-loop-death

Commits on Aug 5, 2026

fix(ocap-kernel): report a dead run loop instead of a healthy kernel

fix(ocap-kernel): roll back the crank the run loop died in

fix(ocap-kernel): report run loop health sampled after the crank wait

fix(ocap-kernel): guard remote ingress at the boundary, not on the queue

fix(ocap-kernel): call the run loop failure handler off a local, not off this

docs: Update changelogs

test(ocap-kernel): cover the run-loop rollback and crank bookkeeping gaps

fix(ocap-kernel): name the real error when an abort rollback fails

test: cover crank rollback against real SQLite and runLoop on the wire

docs: say that a failed kernel is not recoverable in place

fix(kernel-store): close the transaction a failed savepoint rollback abandoned

fix(kernel-cli): remediate a dead run loop before logging it

fix(ocap-kernel): guard the last ingress points and carry the cause to the panel

fix(kernel-cli): tear the kernel down when startup aborts, and terminate

fix(kernel-cli): log best-effort in front of every fatal exit