[rabbitmq-discuss] More RabbitMQ Erlang client woes

[Ben Hood]

Matthias,

On 9 May 2008, at 12:24, Matthias Radestock wrote:

> Ben Hood wrote:
>> On 9 May 2008, at 11:57, Matthias Radestock wrote:
>>> The server is already linking the writer and the channel, though  
>>> the  way that is done may not be immediately obvious.
>> Are you talking about start_link/4 in rabbit_channel? If I'm not   
>> mistaken, this is linking the writing and the channel in the  
>> direct  case, but not in the network case.
>
> I was talking about what the server does. Nothing to do with the  
> Erlang client, other than they share a code base. You appeared to be  
> saying that linking the channel and the writer requires changes to  
> that code base. I want to understand why that is the case, given  
> that the server is establish precisely these kinds of links using  
> that code base. Or, to put it differently, if the server can do it,  
> why can't the Erlang client?

That is right, I did actually say it may need a change to the server  
code. But I was thinking aloud rather than proposing something  
definite because I don't want to change anything in the server without  
good reason.

Actually on closer inspection it may reasonably straight forward to  
do. The Pid of the writer process already being registered with the  
channel process, so it would be a case of either linking the two  
processes together or having the channel send the writer a shutdown  
message when the channel exits.

Ben