Nice timing on that commit, Richard. :)
Refactoring this stuff is easy (though it makes for big commit logs), so I'd like to just start from where you ended up and make the changes I suggest below. --G > -----Original Message----- > From: Glen Daniels [mailto:[EMAIL PROTECTED]] > Sent: Wednesday, May 29, 2002 11:27 PM > To: '[EMAIL PROTECTED]' > Subject: RE: [axis] : Code change across the board > > > > Hi Richard! > > Thank you for posting this stuff before committing. > > In general, this is a fine direction, but I think these need > another round of edits before they're ready for primetime. > Some comments in no particular order: > > NS_URI is used to mean "namespace URI". Some of the > constants are not actually ever used as namespaces > (NS_URI_WSDD_HANDLER springs to mind), and should, I believe > just be URI_ (URI_WSDD_HANDLER). Other examples of this > include the actor URIs and the SOAP HTTP binding URIs. I'd > actually prefer URI_*_NS over NS_URI_* for namespaces, so > that all URI constants are URI_*. URIs used ONLY as > namespaces would get the _NS suffix, but I'd also be fine > simply punting the "NS" entirely, so you'd have > "URI_2001_SCHEMA_XSD", which makes fine sense. > > NS_*_NS is redundant. NS_URI_SOAP12_FAULT_NS, etc. It > should either be NS_URI_SOAP12_FAULT or URI_SOAP12_FAULT_NS > (I prefer the latter, or just URI_SOAP12_FAULT). > > There is no need for "CURRENT_" constants when the value in > question only has one possibility (i.e. > NS_URI_CURRENT_SOAP_UPGRADE, NS_URI_CURRENT_WSDL_*, etc.). > They're just clutter. > > We should, I think, change "CURRENT" to "DEFAULT" in the > cases that do have multiple values, since that's really what > it means to have a preferred constant. "Current" implies > "the one we're actually using in the code at a particular > time", which, since we can support multiple SOAP/Schema > versions, is much more dynamic and is expressed by things > like the SOAPConstants interface. > > ELEM is used to mean "XML element", and thus the PROVIDER_* > constants don't really fit with ELEM_ prefixes (provider is > an attribute value). I'd prefer to leave this off for > strings not used as XML elements, so perhaps just PROVIDER_*? > > The namespaceURI for what you have as QNAME_JAVARPC_PROVIDER > (WSDDConstants) should be NS_URI_WSDD_JAVA. > > What's NS_PREFIX_WSDD? > > The ALT constants for various URIs - I am under the > impression that "http://www.w3.org/1999/XMLSchema"; is > correct, and "http://www.w3.org/1999/XMLSchema/"; is just > invalid. Same for the rest. Why do we want these in there at all? > > --Glen > > > -----Original Message----- > > From: Richard Sitze [mailto:[EMAIL PROTECTED]] > > Sent: Wednesday, May 29, 2002 5:16 PM > > To: [EMAIL PROTECTED] > > Subject: Re: [axis] : Code change across the board > > > > > > It's been suggested that I submit for review the "new" Constants, > > WSDDConstants, and XMLType files before I commit them (and > > corresponding > > changes to many other files): > > > > > > (See attached file: Constants.java) > > > > (See attached file: WSDDConstants.java) > > > > (See attached file: XMLType.java) > > > > > > ******************************************* > > Richard A. Sitze [EMAIL PROTECTED] > > CORBA Interoperability & WebServices > > IBM WebSphere Development > > >
