|
Posted
6 months
ago
by
Oleg Kalnichevski
Idle connection management features
|
|
Posted
6 months
ago
by
Sanghwa Min
Hi Oleg, Thanks for the clear feedback. My suggestion was on the predictable race condition from a known server idle timeout. I appreciate the clarification on `closeIdle`. My initial thought was to handle this avoidance automatically within the Read more
|
|
Posted
6 months
ago
by
Oleg Kalnichevski
michael-o "...Для Вас, Козлов, подземный переход построили..." davsclaus Usually I would be willing to go to any lengths to help out a fellow Apache project. In this case however I feel that adding one more config knob to tweak would ultimately be a Read more
|
|
Posted
6 months
ago
by
Oleg Kalnichevski
> The validateAfterInactivity setting is very useful for preventing stale connections. However, it does not offer a complete guarantee of avoiding a NoHttpResponseException due to a potential race condition. sanghwa Nothing does. > For instance, Read more
|
|
Posted
6 months
ago
by
Oleg Kalnichevski
gjasinski I believe this defect has been fixed in 5.5 by https://github.com/apache/httpcomponents-client/commit/2f6694fcf3326647aeede50df602806a8ef98d30 Please upgrade to HttpClient 5.5, re-test and if the problem still persists, re-open the ticket. Read more
|
|
Posted
6 months
ago
by
Arturo Bernal
Allow to control HTTP redirect 301/302 to not change POST to GET
|
|
Posted
6 months
ago
by
Arturo Bernal
jonenst Please give the updated branch a spin and tell me if this matches your “opt-out” requirement. No JVM system properties are read unless useSystemProperties() is called. There’s an opt-out for legacy behavior: HttpClients.custom().disableProxyAutodetection().build(); Read more
|
|
Posted
6 months
ago
by
Arturo Bernal
Connection is not released after receiving "execution cancelled"
|
|
Posted
6 months
ago
by
Claus Ibsen
At Apache Camel we have users reporting a situation with using a graphql service that returns a 301 redirect and this causes the call to fail because the payload is removed due to POST becomes GET. This is correct according to HTTP spec. But often Read more
|
|
Posted
6 months
ago
by
Grzegorz Jasiński
Hello, I am working with httpclient async. I had encountered resource leak caused by race condition. The execution of request was cancelled in the very end of call.
20250812
Read more
|