Hmm, yeah. If you look at the GitHub actions history, it shows when the
builds started failing.

Gary

On Thu, Jun 19, 2025, 11:37 Ryan Schmitt <rschm...@apache.org> wrote:

> At this order of magnitude? I think it's unlikely. It could be a noisy
> neighbor issue, or a bug in the older version of macOS. The fact that
> the actual timeout is almost exactly double what it's supposed to be
> makes me think it's an issue with the TCP state machine, though,
> similar to how TLS handshake timeouts on JDK11 incur a second blocking
> read while closing the socket. (Socket linger?) I'd need to be able to
> look at it in Wireshark to tell, which requires a local reproducer I
> don't have.
>
> On Thu, Jun 19, 2025 at 8:27 AM Gary Gregory <garydgreg...@gmail.com>
> wrote:
> >
> > Could this be a clock granularity issue?
> >
> > Gary
> >
> > On Thu, Jun 19, 2025, 11:17 Ryan Schmitt <rschm...@apache.org> wrote:
> >>
> >> There's something weird going on that's only affecting (macos-latest,
> >> 11). I can't reproduce it locally, even using the same JRE (Temurin
> >> 11.0.27+6) and architecture, although the environment macOS is an
> >> older version (14.7.6). The fact that the timeout is taking twice as
> >> long as it ought to on JDK11 is reminiscent of the bug I documented
> >> here:
> >>
> >>
> https://github.com/apache/httpcomponents-client/blob/fd2870edec85c0a37429fb357f13f844ff755c0c/httpclient5-testing/src/test/java/org/apache/hc/client5/testing/sync/TestTlsHandshakeTimeout.java#L105
> >>
> >> However, I've seen both Http and Https tests fail, and the behavior is
> >> totally inconsistent.
> >>
> >> I'm going to relax the assertion, so that the tests pass more
> >> reliably. We really just want to verify that the socket timeout during
> >> the TLS handshake is indeed the TLS handshake timeout, not the usual
> >> socket timeout or connection timeout.
> >>
> >> On Wed, Jun 18, 2025 at 7:01 AM Oleg Kalnichevski <ol...@apache.org>
> wrote:
> >> >
> >> > Hi Ryan
> >> >
> >> > One of the socket timeout tests failed in one of the environments.
> >> > Could you please take a look?
> >> >
> >> >
> >> > ```
> >> >
> org.apache.hc.client5.testing.sync.TestSocketTimeout$Https.testReadTime
> >> > outs(String)[6]
> >> > ```
> >> >
> >> >
> https://github.com/apache/httpcomponents-client/actions/runs/15734544257/job/44343633924
> >> >
> >> > Oleg
> >> >
> >> > ---------------------------------------------------------------------
> >> > To unsubscribe, e-mail: dev-unsubscr...@hc.apache.org
> >> > For additional commands, e-mail: dev-h...@hc.apache.org
> >> >
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: dev-unsubscr...@hc.apache.org
> >> For additional commands, e-mail: dev-h...@hc.apache.org
> >>
>

Reply via email to