* Remove code comments and add the no-comments rule
Remove the comments from first-party code and configuration files:
TS/JS, JSX, CSS, SQL migrations, Swift, Kotlin, Gradle, shell,
Docker, systemd, Nix, YAML, TOML, JSONC and GritQL.
Keep only what a tool reads: license headers, shebangs, compiler
and lint directives, JSDoc types in plain .js files, generated file
headers, GritQL scope lines, the install-server.sh help text and the
openbot-boat-pre-stop marker. Vendored code, fixtures and released
schemas keep their comments.
Add the rule to AGENTS.md, and change the docs that asked for
comments. Add two i18n keys that only comments named to
RELEASED_SOURCE_KEYS.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Keep comments that are data in strings
The benchmark task files in opencode-free-models-bench.ts tell the
model what each function must do. The sample file and HTML page in
the two stories are display content. Put their comments back.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Keep released migrations and mobile native inputs unchanged
check:migrations compares each released migration with v0.73.0 byte
for byte, so restore src/backend/migrations. A change to a native
build input of apps/mobile makes a new runtime version, so restore
those files and the extra fingerprint sources.
List these files and the comment markers that a tool reads in
AGENTS.md. Remove the last requests for comments in
src/renderer/AGENTS.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Stop a server start loop instead of a restart each 10 seconds
A server whose start fails stops at the first failure for a newer database or a failed
migration, and after 3 failed starts in a row for other failures. It writes the reason
to openbot-start-halt-v1.json and exits with 78. openbot.service has
RestartPreventExitStatus=78. A restart, openbot update, or the update timer (new check
step) starts it again, and that start tries once more.
Plan 011 item 3 (#10769): parts 1 and 3 and the reason file. Safe mode is left.
* Start a stopped server again at each run of the update timer
* Start a halted server also when the stage step fails
A failed manifest fetch or a full disk made check exit before the
restart, so a hosted server could stay stopped. stage now runs in a
subshell, and check exits with its status after the restart.
* Keep errexit in the stage that the update timer runs
* Remove a one-line helper and comments that repeat the start loop docs
* Count a migration failure as a failed start, and try again each hour
A server stops its start loop after 3 failed starts in a row, of any kind,
and at once only for data from a newer OpenBot. A failed migration rolls
back and counts as one failed start. DatabaseMigrationError and the
migration_failed reason are removed.
openbot-update.timer runs `check` each hour. A stopped start loop stages
and starts again at each run. While OpenBot runs, `check` stages only
about each 6 hours, so the release rollout to servers does not change.
The E2E runs real migration failures and the real `check` under bash.
* Keep a server stopped that an admin stopped after its start loop stopped
openbot stop now clears the failed state of the unit, and the update timer starts OpenBot only
when the unit is failed with status 78.
* Update hosted servers to the newest release at the next start
A timer stages a newer release while OpenBot runs. At the next boot, before OpenBot starts, a
service copies it into place. The updater reads the manifest of the Linux desktop updater, checks
the SHA-512, and never moves a server to an older version.
boat saves /opt by the changed paths, and it drops the files of a directory that `mv` moves.
So a release gets to /opt only as a copy of each file, with markers for a copy that a stop
interrupts. The updater also waits for boat's lazy restore after a resume.
provision.sh uses the same steps, and the Linux build contains scripts/hosting, so a new release
can update the server scripts, units and packages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Do not stop a hosted server in use for a plan change
A resize stops the server. The Worker now waits until the server sent no
activity report for 7 minutes. A server that stops for no use gets the
new machine at its next resume.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Make hosted updates recover from a failed apply
- apply waits for the disk restore only when an update is staged, so
OpenBot does not start late at each resume.
- The markers stay until the hosting files are installed, so a failure
is done again at the next boot. stage does nothing until then.
- stage checks all hosting files and stops when the installed version is
not known. provision.sh does not downgrade.
- Finish an interrupted dpkg run before apt. Longer apply timeout.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Claim a hosted server resize before the boat stop
The cron reads its rows first and then stops each server in sequence.
Activity that came in after the read did not stop the resize. The
resize now sets the row to stopping only when the server had no use in
the last 7 minutes, and a 409 from boat sets it back to running.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>