mirror of
https://github.com/systemd/systemd.git
synced 2026-11-07 13:57:10 +00:00
A token refresh after HTTP 401 counts as a retry, but previously did not set `retry interval`. If a 503 or a connection error followed, `context_schedule_retry()` kept doubling the initial interval of 0, so all remaining retries ran without delay. Simplify and robustify this: Compute the interval from the number of retries instead of keeping it as separate state. Add a test case with a 401 followed by 503s. Signed-off-by: Martin Pitt <martin@amutable.com> (cherry picked from commit068054acef) (cherry picked from commite92a09c97a)