tomcat-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Costin Manolache <>
Subject Re: cvs commit: jakarta-tomcat-catalina/catalina/src/share/org/apache/coyote/tomcat5
Date Thu, 22 May 2003 13:45:17 GMT
Remy Maucherat wrote:

> wrote:
>> billbarker    2003/05/21 20:43:37
>>   Modified:    catalina/src/share/org/apache/coyote/tomcat5
>>   Log:
>>   Initial version of the proposal to pass-through all attributes to the
>>   protocolHandler, and let it deal.
>>   Now, only explictly configured attributes are passed through to the
>>   Protocol.  It is the Protocol's responsibility to have the correct
>>   default attributes configured.  This way, the HTTP/1.1 connector can
>>   default the value of "connectionTimeout" to one value, and the
>>   Jk-Coyote connector can default it to another.
>>   For now, I'm leaving the getters and setters for all of the different
>>   attributes, so that they can still be seen from the JMX consol.  In the
>>   transistion, some of the values shown may be wrong because of the
>>   paragraph above.
>>   For now, I'm leaving in the SSL SocketFactory support (for backwards
>>   compatibility).  It should be deprecated.  You can configure SSL
>>   support without it by e.g. specifying
>>   "property[keystore]=/path/to/my/.keystore" on the connector.
> I'm ok with leaving that in for compat.
> For JMX, we need to expose get/setProperty as operations. Is that done
> automagically ? (maybe so, I consider the Costin improved modeler to be
> automagical ;-) )
> Remy

Maybe we need to add more magic for that. For example we could add JMX
attributes that map to setAttribute(name, value ) instead of

I think the real problem is that we are configuring the wrong object - 
it would be better to have the jk, http components ( and whatever else 
is configurable ) directly configured. That would require a bit too much


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

View raw message