Hi Stefan, Nice work with the patch. Could you read the page on contributing, and make a pull request with this? Thanks!
Pieter On Dec 4, 2012 3:14 AM, "Stefan Radomski" < [email protected]> wrote: > Hi again, > > attached is a patch for socket_base.cpp and socket_base.hpp to be applied > against the current 3.2.2 release (2012/11/23) as well as a test-case. I am > pretty new to the zeromq source, so please review extra carefully. I am not > sure whether we should remove the pipe from pipes array as well. > > The patch solves both LIBZMQ-476 and (by coincidence) LIBZMQ-475 I > reported earlier. > > Best regards > Stefan > > On Dec 3, 2012, at 9:52 PM, Martin Hurton <[email protected]> wrote: > > >> But I cannot figure out, which pipe to terminate as I am having > difficulties finding out, which one is connected to the peer endpoint. > Also, I am not sure how to terminate a pipe - is it actually sufficient to > call terminate on the pipe? If you help me to complete this one, I'll > submit it as a patch. > > > > To terminate a pipe, you can use terminate method. The method takes a > > flag which indicates whether you are to allow the pipe to process > > queued messages. This is used to allow the pipe to process queued > > messages, depending on the linger setting. > > This flag is set to true for now, I am not sure about the eventual > implications for inproc sockets. > > > > > As for endpoint tracking, I would suggest to keep separate container > > for inproc connections. > > > > Would be great to have this covered by a test too. > > > > - Martin > > > _______________________________________________ > zeromq-dev mailing list > [email protected] > http://lists.zeromq.org/mailman/listinfo/zeromq-dev > >
_______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
