activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Max-Julian Pogner <max-jul...@pogner.at>
Subject Re: question regarding TcpTransportFactory.java
Date Fri, 16 Mar 2018 22:07:54 GMT
Ah, now i see!


But i wonder: is there a conflict when using the URI like this, because
there might be already another use of the path-part w.r.t to the
active-mq configuration. However, it could be argued that this use is
restricted to the "tcp" scheme and as long as there is no conflicting
using _within_ the "tcp" scheme it'll work out.

However, what if the "auto" schema was specified in the original config
file, and maybe the "auto" scheme passes the uri-path along to the
detected actual scheme, and then two possible actual schemes have
surprisingly different interpretations of the uri-path.


*) the URI location passed as parameter to the createTransport function:
is the location.getScheme() always "tcp"? even if the original
configuration is something like "auto://localhost:61616"?

*) if the path part of location indeed has no conflicting purpose, i
would propose - as a first idea - a patch similar to the diff attached.
(Note: i didn't compile yet, let alone test)


with regards,

Max



> [ w.r.t TcpTransport ]
>
>
> OK, try a url like this:
>
> tcp://localhost:61616/localhost:56565
>
>
> You'll find it connects to localhost:61616 and binds to localhost:56565
>
> I don't see any documentation on the website for this feature though.
>
> Art
>


Mime
View raw message