--- Comment #6 from Konstantin Kolinko <> ---
(In reply to Remy Maucherat from comment #3)
> Do you imply what the testPost test does is incorrect ?

I do not see any obvious flaw in the design. I think that many clients behave
like that.

The servlet design is a bit unusual for me, as Content-Length of the request is
known in advance.

I wonder whether "Connection reset" is caused by some timeout. The test runs
for 9-10 seconds for me, and other tests in the testsuite complete in about 1

> The test does work for me [so far]. As with the rather equivalent HTTP/2 TLS
> test, it is meant to test large POSTs with TLS since that's what people
> often complain about.
> The test was included in 9.0.33 already (since there was a regression) and
> there are no relevant changes in 9.0.34 (only NIO got a minor change, but
> you are saying it works). Why did it work then ?

The detailed logging of errors was added just recently, by Mark on April 2nd:

SHA-1: f8da18d362b8a773afd557fdda0fe50a3748c90c
* Failed test on Windows. Expand logging to aid debugging of failures.

You are receiving this mail because:
You are the assignee for the bug.
To unsubscribe, e-mail:
For additional commands, e-mail:

Reply via email to