-0 The setProperty was done way before JMX, and it's quite hacky. I think standardising on JMX-like model is better - i.e. if we make changes to setProperty and connector configuration, it should bring it closer to JMX.
In JMX we have a "void setAttribute(...) throws AttributeNotFound", plus a method to get supported attribute names. So instead of returning the boolean - I think it would be better to add a method to list supported attributes ( and with some changes to modeler this could be reflected in the mbean - right now setProperty is ignored ). It's a bit more complicated - but also more valuable. Costin -0 On 10/19/07, Filip Hanik - Dev Lists <[EMAIL PROTECTED]> wrote: > > Ok, sending a little email to lobby for a useful change. To the contrary > of comments and vetoes to this change, it doesn't change existing > behavior, it simplifies it, and also adds the NIO connector to the > validation of server.xml attributes. > > It'd be nice to get one more vote, so that we can get it in before the > tag. > > From STATUS > * Make server.xml parsing warnings more generic, add support for the NIO > connector > * Connector is no longer an exception case. Any new component can still > use setProperty and return a boolean > * to accept or reject the property > > http://people.apache.org/~fhanik/patches/digester-attribute-warnings.patch > > > Filip > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > >
