… both clearer to the end-user and more consistent with RPC::XML::Server.
…ists of signatures, to catch cases in which a newly-added (or deleted?) signature could cause the re-hashing of the table to fail. Added apropos docs to the man page.
…d added the documentation (including a big chunk from RPC::XML::Server, the XPL stuff).
RPC::XML::Method man page. Changed the absolutisms of RPC::XML::Method in the server code to allow a user-specified class. Some tweaks to other doc areas.
…ince this is no longer a "reference implementation", there is no longer the need for the extra characters being sent down the pipe.
…egate almost all method-related functionality to here, except for the actual invocation (and I'm still debating that one).
…tten to use the RPC::XML::Method class instead. Makes for much cleaner code at this level, particularly around the validity tests and such.
… new object from an old, pointing to the same code-ref but copying the signature array.
…oding until XML stringification, thus it isn't done for usual value() retrieval. This has the added bonus that the class can be a subclass of RPC::XML::simple_type and the new() and value() methods can be inherited from there.
… the string before pasting the value into the XML. This was causing & and related characters to be transmitted unencoded.
…ge::Size, this uses an external template rather than an in-line spec file from __DATA__.
…values filled in by MakeMaker.
…ntry in the list of modules if mod_perl is not available. Otherwise, the resulting test failure may confuse people.
completeness as possible.