Walter,
please be so kind to use my name rather than "the library maintainer".
I'm sorry, that message was somewhat more confrontational in tone than
was really warranted.
First of all, let's make clear whether this problem is actually with
overloading. For example, the method which was picked in that case was
the *second* method. I would assume, that the first method would be
choosen. Next, we do not even know, whether his problem is with version
2, or version 3.
True enough. I was mainly intending to point out that this could be one
possible cause of the issue, but you are correct that the code he
provided would run under version two.
Then, I wasn't aware, that overloading was possible by
default with version 2. If that's the case, I might rething my opinion.
Interesting; I had thought that you were aware of this. The design of
version two (exactly one Java class per XML-RPC type) meant that any set
of XML-RPC parameters has exactly one correspoding typed method
signature. Thus, the full-signature version of the 'getMethod' call can
be used on method invocation to retrieve the matching overloaded method
for a particular method call, and this is how version 2 of the library
operates. I suppose it's not completely obvious from the code that
overloading is supported, but I've used it personally and can verify
that it works. In fact, my personal use of overloading with version two
is what prompted me to add support in version three, since I need such
support to continue using the library.
Note that this former method won't work for version three because a
getMethod("foo", {List}) call won't find a foo(Object[]) method.
However, given that the non-overloading version 3 currently always
passes Object[]'s (I think), methods taking Lists are broken on the
server side anyway. Whether this is acceptable is your call, but the
code to deal with it is pretty simple (it's in at least one of my
patches.) I'm not sure about method taking double vs. Double etc., but
IMHO bo should probably be supported if possible as well.
> And finally, please note, that I am not the only developer. This is a
> community and I would never reject a discussion.
Even so, in most smaller and many larget F/OSS projects there is a
"lead" developer who end up making the final decisions, and to the
degree than anyone fills such a role for this project, it's currently
you. I don't begrudge you that role -- in fact, I'm very happy that
someone is willing to put forth the time and effort to work on the code
(and to deal with abrasive users/contributors like myself.) And up to
this point, you've been (from my point of view) quite determined to keep
overloading out of version three, despite its usability in version two.
--
Walter Mundt
[EMAIL PROTECTED]
begin:vcard
fn:Walter Mundt
n:Mundt;Walter
org:Cole Engineering Services Inc.
adr:3361 Rouse Road;;St 230;Winter Park;FL;32792;USA
email;internet:[EMAIL PROTECTED]
title:Engineering Specialist
tel;work:407-384-9956
tel;fax:407-281-7011
x-mozilla-html:FALSE
version:2.1
end:vcard
