> Ara Abrahamian wrote:
> > A genius guy sent a very interesting patch, see
> >
http://sourceforge.net/tracker/index.php?func=detail&aid=446889&group_id
> > =31602&atid=402706. I think it should work fine. It calculates
> > serialVersionUid based on method signatures of generated interface.
> >
> > I haven't tested it, but if it works fine then I think we can
activate
> > it as default. PK and data-object templates will use it. Is it ok or
it
> > needs something more than adding hashcodes to generate versionID?
> 
> Sorry, but this needs to be used:
>
http://java.sun.com/j2se/1.3/docs/guide/serialization/spec/class.doc6.ht
ml
> #4100
> 
> I wrote something similar in JBoss for the RMI stuff that needed
> identification of methods. See here:
> http://cvs.sourceforge.net/cgi-
>
bin/viewcvs.cgi/jboss/jboss/src/main/org/jboss/ejb/plugins/jrmp/interfac
es
>
/RemoteMethodInvocation.java?rev=1.14&content-type=text/vnd.viewcvs-mark
up

Hi,
I implemented it, based on ObjectStreamClass and by changing it from
using java.lang.reflect api to javadoc api. Unfortunately I had to do
some labor-intensive things, such as providing empty implementations for
javadoc classes, cause they are package protected and I wanted to use
them directly. Anyway, it works and the generated suid is exactly equal
to the one generated from serialver.exe. The only short coming is no
support for static initializers. Unfortunately there's no support in
javadoc api for access to static blocks, but it's no problem for us, our
templates generate files that don't have such blocks. It's implemented
used in entitypk.j currently.

Thanx for the links :-)

Ara.


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com


_______________________________________________
Xdoclet-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/xdoclet-devel

Reply via email to