Since the dependency in my deep-clone method is only on SerializationUtils, and the clone method there just serializes and then deserializes the object being cloned, it might be easiest to avoid the dependency by simply re-implementing that functionality in my method. What do you think?
As for the move method I mentioned, I don't think it's as specialized as several have said. I implemented it for a web application that stored the rows in a data list in a List, and I had to make it possible for the user to move a row up or down in the list. I think this is, while not an eveyday requirement, certainly common enough that inclusion of the method has some merit. -Matt On 2/1/06, Torsten Curdt <[EMAIL PROTECTED]> wrote: > > > On 02.02.2006, at 00:06, Jörg Schaible wrote: > > > Torsten Curdt wrote on Wednesday, February 01, 2006 12:25 PM: > > > >>> This could be useful, but would need to be done without a > >>> dependency on commons-lang. > >> > >> We already talked about that at the ApacheCon. > >> > >> Maybe it would useful to loosen such requirements by using tools like > >> "minijar" > >> > >> http://vafer.org/blog/tag/minijar > >> > >> or "proguard + jarjar" > >> > >> http://proguard.sourceforge.net/ > >> http://tonicsystems.com/products/jarjar/ > >> > >> we add the dependency so we don't duplicate code but strip the > >> unnecessary classes and inline the remaining. > >> > >> WDYT? > > > > Looking at proguard, this might be enough. Why do you think, jarjar > > is also necessary ? > > Jarjar is better suited for the inlining. Unfortunately proguards > configuration sucks quite bad and it require all transitive deps > to be analyzed. Both are GPL btw. > > That's why I came up with minijar ...mainly for using it for jci. > > cheers > -- > Torsten > >
