> On Jan. 2, 2013, 9:09 p.m., Kenneth Giusti wrote:
> > /trunk/qpid/cpp/src/qpid/broker/Broker.cpp, line 758
> > <https://reviews.apache.org/r/8804/diff/2/?file=244412#file244412line758>
> >
> >     One (potential) issue: the link name is supplied by the user and is 
> > pretty much free-format (with the exception of starting it with the prefix 
> > "qpid").
> >     
> >     Would there be an issue if the user assigned link name collided with an 
> > incoming connection name?

Hmm, yes I think that if an administrator chose a link name that was identical 
to another connection name, then Connection events coming from that Connection 
might be confused with events coming from the real Connection associated with 
the Link. This would be a good reason to do a small amount of "uniquifying" to 
the name we use for the connection.


- Andrew


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/8804/#review15019
-----------------------------------------------------------


On Jan. 2, 2013, 8:25 p.m., Andrew Stitcher wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/8804/
> -----------------------------------------------------------
> 
> (Updated Jan. 2, 2013, 8:25 p.m.)
> 
> 
> Review request for qpid and Kenneth Giusti.
> 
> 
> Description
> -------
> 
> This change removes the ad hoc deconstruction of Connection management ids to 
> discover the end points of the Connection so that they can be used by to look 
> up a Link in the LinkRegistry.
> 
> Instead we allow the federation code to specify the management id of the 
> Connection directly then use this specified name as an index to find the Link 
> directly.
> 
> Currently this makes the Connnection id identical to the Link name that is 
> using the connection, it would be easy to change this to some other 
> convention, but this seems simple.
> 
> 
> This addresses bug qpid-4315.
>     https://issues.apache.org/jira/browse/qpid-4315
> 
> 
> Diffs
> -----
> 
>   /trunk/qpid/cpp/src/qpid/broker/Broker.h 1426152 
>   /trunk/qpid/cpp/src/qpid/broker/Broker.cpp 1426152 
>   /trunk/qpid/cpp/src/qpid/broker/Link.h 1426152 
>   /trunk/qpid/cpp/src/qpid/broker/Link.cpp 1426152 
>   /trunk/qpid/cpp/src/qpid/broker/LinkRegistry.cpp 1426152 
>   /trunk/qpid/cpp/src/qpid/sys/AsynchIOHandler.h 1426152 
>   /trunk/qpid/cpp/src/qpid/sys/AsynchIOHandler.cpp 1426152 
>   /trunk/qpid/cpp/src/qpid/sys/ProtocolFactory.h 1426152 
>   /trunk/qpid/cpp/src/qpid/sys/RdmaIOPlugin.cpp 1426152 
>   /trunk/qpid/cpp/src/qpid/sys/SslPlugin.cpp 1426152 
>   /trunk/qpid/cpp/src/qpid/sys/TCPIOPlugin.cpp 1426152 
> 
> Diff: https://reviews.apache.org/r/8804/diff/
> 
> 
> Testing
> -------
> 
> Passes make check
> 
> 
> Thanks,
> 
> Andrew Stitcher
> 
>

Reply via email to