You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@thrift.apache.org by "Paulo Neves (Jira)" <ji...@apache.org> on 2019/12/01 16:26:00 UTC

[jira] [Commented] (THRIFT-5034) Allow for Reverse Connection servers

    [ https://issues.apache.org/jira/browse/THRIFT-5034?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16985597#comment-16985597 ] 

Paulo Neves commented on THRIFT-5034:
-------------------------------------

For some reason the github PR has not been associated with the ticket. I have added a test case, which also doubles as an example. I have also added the changes required in the review.

The CI travis is currently failing in a C++ scan unrelated to my python changes, so i guess it is ready from a code point of view.

> Allow for Reverse Connection servers
> ------------------------------------
>
>                 Key: THRIFT-5034
>                 URL: https://issues.apache.org/jira/browse/THRIFT-5034
>             Project: Thrift
>          Issue Type: Improvement
>            Reporter: Paulo Neves
>            Priority: Major
>
> Hello,
> One of the reasons I migrated from gRPC to Thrift was because of the fundamental distinction between RPC Client/Server and Transport Client/Server. This has allowed me to create RPC processors that actually connect to a server(broker), in a similar way to a reverse proxy. This is immensely useful for restricted networks.
> As said, due to the elegant separation between processor and transport this turned out to be trivial and I would like to share my implementation of such a transport "decorator" so that we can have Client transports work in Processor servers.
> I do not have tests yet, besides my own working private implementations but will try soon to prepare such for at least the Socket examples.
> I also do not know very well how the naming should be so i just made it TReverseTunnelServer.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)