tomcat-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mladen Turk <>
Subject Re: svn commit: r745842 - in /tomcat/connectors/trunk/jk: native/apache-2.0/ native/common/ xdocs/miscellaneous/
Date Wed, 25 Feb 2009 16:10:23 GMT
Rainer Jung wrote:
> On 25.02.2009 13:44, Mladen Turk wrote:
> Concerning host vs. hostname: I would prefer to keep the variable names 
> the same. I can do the change. Do you prefer host in both cases, or 
> hostname? I think host makes somes sense, because it reflects the 
> attribute.

Sure, feel free to rename it to whatever you think
is more appropriate.

>>> i think we should log an error in this case and maybe also reset the
>>> data to the previous working set.
>> I don't like resetting to a previous version.
>> If it's bad, it's bad, just like with init.
> Hmm, but that means any typo when submitting the status form will break 
> the application. Of course we can't fix this in any case, but at least 
> if the address doesn't resolve, we could. Maybe we should add a 
> resolution test when the form gets submitted, and if the address doesn't 
> resolve, we immediately return an error message in the status worker. 
> Are you OK with that? I would add it then. It is kind of validation of 
> form input data.

Think we could make that safe in status worker, and if resolved
then copy to shm. This resolution is already part of status worker,
so it shouldn't be a problem.

> In any case a log message (error) is needed, because the problem is fatal.


>> We could. Nothing prevents that.
> Any idea, why this port limit was chosen? What's bad about allowing to 
> connect to a privileged port?

No idea. Think someone thought that tomcat should only
run with user account (port > 1024), but we can have it
run by a commons-daemon, so that's not the case any more.

Anyhow this limit makes no sense. It should check
just for (port > 0), cause 0 means uninitialized

This is important so we can create multiple nodes
in advance for whom we don't know what the address will be.


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message