TestFiveFailedAsksKeepTheNextRequestFromAskingWAF/no_answer_at_all in internal/rproxy/rproxy_test.go fails some runs of make check and not others. The failure is rproxy_test.go:1439: no question. It happened on a fresh clone whose only change from next was outside internal/rproxy.
Likely cause, not verified: the test sets SWW_RPROXY_WAF_BACKEND_TIMEOUT_MS to 200. It expects each of the five visitor requests to reach the stand-in for waf as a question. Under load, with the race detector on, an ask can time out before the stand-in's handler runs. The ask still counts as a failure, but no question arrives, so standIn.question waits its 10 seconds and fails the test.
The test came in with #393. Until it is fixed, any PR's make check can fail at random, including the one run on a fresh clone before a merge into next.
Definition of done
The test passes whether or not each ask reaches the stand-in within the time limit.
It still proves that five failed asks keep the next request from asking waf.
make check passes.
Model: opus-5-5
`TestFiveFailedAsksKeepTheNextRequestFromAskingWAF/no_answer_at_all` in `internal/rproxy/rproxy_test.go` fails some runs of `make check` and not others. The failure is `rproxy_test.go:1439: no question`. It happened on a fresh clone whose only change from `next` was outside `internal/rproxy`.
Likely cause, not verified: the test sets `SWW_RPROXY_WAF_BACKEND_TIMEOUT_MS` to `200`. It expects each of the five visitor requests to reach the stand-in for `waf` as a question. Under load, with the race detector on, an ask can time out before the stand-in's handler runs. The ask still counts as a failure, but no question arrives, so `standIn.question` waits its 10 seconds and fails the test.
The test came in with https://git.eeqj.de/sneak/smallwebwaf/pulls/393. Until it is fixed, any PR's `make check` can fail at random, including the one run on a fresh clone before a merge into `next`.
## Definition of done
- The test passes whether or not each ask reaches the stand-in within the time limit.
- It still proves that five failed asks keep the next request from asking `waf`.
- `make check` passes.
Model: opus-5-5
The test no longer waits for a question from each failed ask. It checks that the sixth request did not ask waf by matching that request's ID, which the app records, so late questions from the failed asks do not count.
Model: opus-5-5
PR: https://git.eeqj.de/sneak/smallwebwaf/pulls/399, branch `issue-398-breaker-test-without-questions`.
The test no longer waits for a question from each failed ask. It checks that the sixth request did not ask `waf` by matching that request's ID, which the app records, so late questions from the failed asks do not count.
Model: opus-5-5
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.
TestFiveFailedAsksKeepTheNextRequestFromAskingWAF/no_answer_at_allininternal/rproxy/rproxy_test.gofails some runs ofmake checkand not others. The failure isrproxy_test.go:1439: no question. It happened on a fresh clone whose only change fromnextwas outsideinternal/rproxy.Likely cause, not verified: the test sets
SWW_RPROXY_WAF_BACKEND_TIMEOUT_MSto200. It expects each of the five visitor requests to reach the stand-in forwafas a question. Under load, with the race detector on, an ask can time out before the stand-in's handler runs. The ask still counts as a failure, but no question arrives, sostandIn.questionwaits its 10 seconds and fails the test.The test came in with #393. Until it is fixed, any PR's
make checkcan fail at random, including the one run on a fresh clone before a merge intonext.Definition of done
waf.make checkpasses.Model: opus-5-5
PR: #399, branch
issue-398-breaker-test-without-questions.The test no longer waits for a question from each failed ask. It checks that the sixth request did not ask
wafby matching that request's ID, which the app records, so late questions from the failed asks do not count.Model: opus-5-5