There's going to be confusion no matter what we do.
The thing I care about most is if people run old scripts or follow old
instructions that their teams have written up that say
--controller=XXX:9999 it should just work. We shouldn't force people to
rewrite their scripts and processes just because we decided port
reduction was more important than compatibility.
Remember also that people will use the new CLI binary to manage legacy
servers where HTTP upgrade is not an option. So even if by default WF
isn't using 9999, there will be need for the CLI to deal cleanly with 9999.
(Most of the above wasn't directly in response to your last comment.)
On 7/9/13 11:03 AM, Darran Lofthouse wrote:
On 09/07/13 16:57, Brian Stansberry wrote:
> Sounds fine. I'd be fine with a simple heuristic based on the port as
> well, so --controller=localhost:9990 would default to http-remoting.
I can add that if we want but just thinking if it would cause confusion
if it suddenly stops working once they start trying to connect to a
second instance with offset ports.
If they do not specify any address at all it will default to the address
in the config which is http-remoting by default.
_______________________________________________
wildfly-dev mailing list
wildfly-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/wildfly-dev
--
Brian Stansberry
Principal Software Engineer
JBoss by Red Hat