Irving Reid
Irving Reid
Looks like your test matrix pairs Redis 4.8 with older versions of the `resque` gem, which trigger deprecation failures. Would you rather change the test matrix, or remove `Redis.raise_deprecations =...
@PatrickTulskie removed the gemspec change here too, because the main branch fixed Redis 5.0 compatibility :-)
After a fair bit of local debugging, I suspect this test failure is flaky? I could only reproduce it when I had another instance of `resque-scheduler` attached to my Redis...
> Would you mind removing the gemspec changes? The gemspec changes aren't required for this fix. @PatrickTulskie done, thanks for reviewing!
Ah OK. We usually do squash merges here, so we don't fuss about the branch being noisy :-)
Sorry about the delay - I got tricked by Github not syncing the `upstream` and `origin` master branches quite the way I expected, and couldn't figure out why the test...
Slid over here from the HTTPWG comment thread, to note that HTTP 1.0 *did* (arguably) specify that when either side half-closes, the server should stop processing the request: the last...