[ http://issues.apache.org/jira/browse/HTTPCORE-11?page=all ]
Oleg Kalnichevski updated HTTPCORE-11:
--------------------------------------
Attachment: httpclientconn.patch
Here's the third patch that makes some significant changes to the
HttpClientConnection interface in order to address the following issues:
(1) Presently HTTP client connections contain the code intended to take care of
opening a network socket to a target host. This poses a problem for
asynchronous scenarios where the process of establishing a connection is
non-blocking and must be performed by an I/O reactor.
(2) Presently HTTP client connections need to be aware of their target host in
order to be able to re-create the underlying network socket. In my opinion, it
is the responsibility of a connection manager to know which target host a
connection is pointing at, and not of the connection itself.
(3) HttpRequestExecutor should not attempt to re-open the connection in case of
an I/O failure, because the process of re-creating the connection may be more
involved than just opening a new socket (think of connection tunneling or the
non-blocking I/O)
Please take a closer look at the proposed changes and let me know what you think
Oleg
> Provide generic server and client connection primitives that can work with
> arbitrary HTTP data receivers and transmitters.
> --------------------------------------------------------------------------------------------------------------------------
>
> Key: HTTPCORE-11
> URL: http://issues.apache.org/jira/browse/HTTPCORE-11
> Project: HttpComponents Core
> Issue Type: Improvement
> Components: HttpCore
> Affects Versions: 4.0-alpha2
> Reporter: Oleg Kalnichevski
> Assigned To: Oleg Kalnichevski
> Fix For: 4.0-alpha3
>
> Attachments: conn.patch, httpclientconn.patch,
> httpcore-abstractconn.patch
>
>
> Provide generic server and client connection primitives that can work with
> arbitrary HTTP data receivers and transmitters. The default connection
> primitives in their present form are tightly coupled with the java.net.Socket
> class. There are cases when the underlying I/O transport is based on
> different media (MINA transport, NIO channels)
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
