org.apache.http.impl.client.InternalHttpClient.getParams() throws UnsupportedOperationException [SPR-12530]

[spring-projects-issues]

Stephen More opened SPR-12530 and commented

Method getParams() in org.apache.http.impl.client.InternalHttpClient is no longer supported.

org.springframework.remoting.httpinvoker.HttpComponentsHttpInvokerRequestExecutor.setConnectTimeout( int timeout )

and

org.springframework.remoting.httpinvoker.HttpComponentsHttpInvokerRequestExecutor.setReadTimeout( int timeout )

are trying to call getParams() from httpclient.

---

Affects: 4.1.3

Issue Links:

• Upgrade HttpComponentsHttpInvokerRequestExecutor to require Apache HttpComponents 4.3 [SPR-11113] #15739 Upgrade HttpComponentsHttpInvokerRequestExecutor to require Apache HttpComponents 4.3 ("is duplicated by")

[spring-projects-issues]

Stephen More commented

HttpParams is deprecated: http://hc.apache.org/httpcomponents-core-ga/httpcore/apidocs/org/apache/http/params/HttpParams.html?is-external=true

[spring-projects-issues]

Stephen More commented

#707

[spring-projects-issues]

Stéphane Nicoll commented

Thanks Stephen but we gave it a try already in #15739 (see the candidate fix here). Your patch is somewhat similar to what I've done except that I am doing my best not to remove the ability to customize the read and connection timeout.

The current fix is scheduled for 4.2.RC1, maybe we should reconsider and move that back to the 4.1.4 bucket. Juergen Hoeller, thoughts?

[spring-projects-issues]

Stephen More commented

I would vote for 4.1.4....According to http://mvnrepository.com/artifact/org.apache.httpcomponents/httpclient 4.3 was released Sep, 2013
#15739 recommends waiting 6 months after 4.3 is released. 4.3 branch has been out for well over a year. I could understand waiting for 4.2 but getParams() now throws Unsupported which is a much bigger impact than simply being deprecated.

I can revise my pull request to simply wrap getParams() in a try catch block for 4.1.4 to support 4.3.6. My issue is a BUG that should be addressed in 4.1.4 #15739 is an improvement.

Just my 2 cents.

[spring-projects-issues]

Stéphane Nicoll commented

It's a grey area: calling a setter that has no effect isn't a great improvement over calling a setter that throws an exception. I can't accept your patch as it breaks the API in a point release. That being said, I don't think we were aware that the current class was breaking with 4.3 so I'll bring #15739 to the 4.1 bucket. Thanks.