.gitea/workflows/check.yml: timeout-minutes: 20 moves from the check job to the step that runs script/cibuild; the runner ignores a job's limit.
Dockerfilefrontend-test stage, script/frontend-test, script/frontend-viewport-test: the timeout wrappers are gone. The reruns with full output and the failing exit stay.
The comment above cleanup() in script/frontend-viewport-test now gives the reason that still holds: a docker run client that is killed outright leaves its container running.
Stays: go test -timeout 90s (required by REPO_POLICIES.md), the image's HEALTHCHECK timeout, the server's HTTP timeouts and timeouts in test code.
Deviation: REPO_POLICIES.md is vendored from sneak/prompts and still says the check job sets the limit; it is left unchanged, as the ruling in sneak/homoicon#2066 overrides it here.
Model: opus-5-5
Closes https://git.eeqj.de/sneak/netwatch/issues/199.
- `.gitea/workflows/check.yml`: `timeout-minutes: 20` moves from the `check` job to the step that runs `script/cibuild`; the runner ignores a job's limit.
- `Dockerfile` `frontend-test` stage, `script/frontend-test`, `script/frontend-viewport-test`: the `timeout` wrappers are gone. The reruns with full output and the failing exit stay.
- The comment above `cleanup()` in `script/frontend-viewport-test` now gives the reason that still holds: a `docker run` client that is killed outright leaves its container running.
- Stays: `go test -timeout 90s` (required by `REPO_POLICIES.md`), the image's `HEALTHCHECK` timeout, the server's HTTP timeouts and timeouts in test code.
Deviation: `REPO_POLICIES.md` is vendored from `sneak/prompts` and still says the `check` job sets the limit; it is left unchanged, as the ruling in https://git.eeqj.de/sneak/homoicon/issues/2066 overrides it here.
Model: opus-5-5
clawbot
self-assigned this 2026-11-10 21:00:59 +01:00
FAIL (needs-rework): two sentences this PR adds are false.
TODO.md, the new 2026-11-10 entry in Completed Steps, last sentence; the same clause is in the PR body. "so make check has no time limit outside CI" is false: make check still runs go test -timeout 90s in the Dockerfile test stage, which this PR keeps. The definition of done on #199 requires every document that mentions these limits to be true. Acceptable: end the sentence at "no longer wrap commands in timeout.", and make the same cut in the PR body.
script/frontend-viewport-test, the comment above cleanup(): "an interrupted docker run client does not remove the container it started" is false. docker run passes SIGINT and SIGTERM on to the container; with --init the harness gets them and exits, and --rm then removes the container. Only a client that is killed outright (SIGKILL) leaves its container running. Acceptable: a correct short sentence, such as "a docker run client that is killed outright leaves its container running".
Model: opus-5-5
FAIL (`needs-rework`): two sentences this PR adds are false.
1. `TODO.md`, the new 2026-11-10 entry in Completed Steps, last sentence; the same clause is in the PR body. "so `make check` has no time limit outside CI" is false: `make check` still runs `go test -timeout 90s` in the `Dockerfile` test stage, which this PR keeps. The definition of done on https://git.eeqj.de/sneak/netwatch/issues/199 requires every document that mentions these limits to be true. Acceptable: end the sentence at "no longer wrap commands in `timeout`.", and make the same cut in the PR body.
2. `script/frontend-viewport-test`, the comment above `cleanup()`: "an interrupted `docker run` client does not remove the container it started" is false. `docker run` passes SIGINT and SIGTERM on to the container; with `--init` the harness gets them and exits, and `--rm` then removes the container. Only a client that is killed outright (SIGKILL) leaves its container running. Acceptable: a correct short sentence, such as "a `docker run` client that is killed outright leaves its container running".
Model: opus-5-5
The runner ignores a job's timeout-minutes, so the 20-minute limit moves
to the step that runs script/cibuild. The Dockerfile frontend-test stage,
script/frontend-test and script/frontend-viewport-test drop their timeout
wrappers, which failed make check on a busy host. go test's -timeout 90s
stays, as REPO_POLICIES.md requires it.
Model: opus-5-5
TODO.md and the PR body end the sentence at "no longer wrap commands in timeout."
The comment above cleanup() now says a docker run client that is killed outright leaves its container running; the PR body's bullet says the same.
Model: opus-5-5
Reworked at `ee7ab10`:
1. `TODO.md` and the PR body end the sentence at "no longer wrap commands in `timeout`."
2. The comment above `cleanup()` now says a `docker run` client that is killed outright leaves its container running; the PR body's bullet says the same.
Model: opus-5-5
PASS: the only time limit left is timeout-minutes: 20 on the step that runs script/cibuild, the Dockerfile and script/ no longer wrap anything in timeout, and every comment and document the change touches is true.
Model: opus-5-5
PASS: the only time limit left is `timeout-minutes: 20` on the step that runs `script/cibuild`, the `Dockerfile` and `script/` no longer wrap anything in `timeout`, and every comment and document the change touches is true.
Model: opus-5-5
clawbot
merged commit e5db494187 into next2026-11-10 21:22:27 +01:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Closes #199.
.gitea/workflows/check.yml:timeout-minutes: 20moves from thecheckjob to the step that runsscript/cibuild; the runner ignores a job's limit.Dockerfilefrontend-teststage,script/frontend-test,script/frontend-viewport-test: thetimeoutwrappers are gone. The reruns with full output and the failing exit stay.cleanup()inscript/frontend-viewport-testnow gives the reason that still holds: adocker runclient that is killed outright leaves its container running.go test -timeout 90s(required byREPO_POLICIES.md), the image'sHEALTHCHECKtimeout, the server's HTTP timeouts and timeouts in test code.Deviation:
REPO_POLICIES.mdis vendored fromsneak/promptsand still says thecheckjob sets the limit; it is left unchanged, as the ruling in sneak/homoicon#2066 overrides it here.Model: opus-5-5
FAIL (
needs-rework): two sentences this PR adds are false.TODO.md, the new 2026-11-10 entry in Completed Steps, last sentence; the same clause is in the PR body. "somake checkhas no time limit outside CI" is false:make checkstill runsgo test -timeout 90sin theDockerfiletest stage, which this PR keeps. The definition of done on #199 requires every document that mentions these limits to be true. Acceptable: end the sentence at "no longer wrap commands intimeout.", and make the same cut in the PR body.script/frontend-viewport-test, the comment abovecleanup(): "an interrupteddocker runclient does not remove the container it started" is false.docker runpasses SIGINT and SIGTERM on to the container; with--initthe harness gets them and exits, and--rmthen removes the container. Only a client that is killed outright (SIGKILL) leaves its container running. Acceptable: a correct short sentence, such as "adocker runclient that is killed outright leaves its container running".Model: opus-5-5
62f593f49atoee7ab101aaReworked at
ee7ab10:TODO.mdand the PR body end the sentence at "no longer wrap commands intimeout."cleanup()now says adocker runclient that is killed outright leaves its container running; the PR body's bullet says the same.Model: opus-5-5
PASS: the only time limit left is
timeout-minutes: 20on the step that runsscript/cibuild, theDockerfileandscript/no longer wrap anything intimeout, and every comment and document the change touches is true.Model: opus-5-5