Re: [jdee-devel] Status of project and adding 1.5 generics

[Eric M. L. <er...@si...>]

Hi,

  Joakim Verona has started a couple Java related projects with CEDET,
one of which is to use Java as a back-end to a CEDET/Semantic
database.  The API that needs to be supported by the JVM for
semanticdb is small from.  Using such a feature could simplify the
tasks you are looking to accomplish, though I don't know what JDEE
is trying to do in total.

  In the past, I had talked with Paul K. a lot on what he was up to
with the JVM, and I attempted to make some of those concepts more
general in CEDET to support more languages.  Hopefully transitioning
some items to CEDET could make things simpler.

  If anyone want help doing something like that, I'd be happy to
assist, though I don't do any Java programming myself.

Eric

>>> Paul Landes <la...@ma...> seems to think that:
>Over the past few months, I've looked at a lot of the semantic code
>and thought about this problem in general.  I think we still need to
>use a JVM since there's just not a way to compile the information we
>need in semantic for libraries for which we don't have the source.
>
>I've actually thought about writing some Java application specifically
>for the purpose of talking with Emacs.  The single most annoying
>problematic thing I see with this JVM subprocess scripting language
>paradigm is the quoating.  I've tried to do things like send an XML
>buffer to beanshell for transforms and xpath processing and I always
>have problems, in some form, escaping double quotes.  At one point, I
>had to resolve to splitting up the string with single quote char
>representations using appending (i.e. "text " + '"' + "more text").
>
>I really like the idea of using some lisp based language as the
>communication would be much easier.  Currently there is a lot of
>character escaping etc of beanshell script language that floats over
>the pipe to the JVM.  Having something that is a like language,
>particularly another lisp dialect, could really make things easier.
>
>I'm thinking in terms of using eval, read and prin1 type functions
>between the two.  Currently, there is a lot of lisp generating code
>written in Java.  A lot of that goes away using something like armed
>bear.
>
>The concern I have with something like armed bear is what kind of
>community it has or will keep to maintain it.  I think we're all
>concerned about that after what's happened with beanshell.
>
>
>ia...@gr... writes:
> > Perhaps utilizing ABCL (http://common-lisp.net/project/armedbear/), if  
> > possible, would be the best choice.  For those that aren't familiar  
> > with it, it's an ANSI Common Lisp (yay!) that runs on the JVM.  Seems  
> > like a logical choice to me, barring any lurking nasties...
> > 
> > - Ian
> > 
> > Quoting "Phillip Lord" <phi...@ne...>:
> > > But what it does say would be enough, I think. We need some emacs-java
> > > communication. Perhaps it doesn't need to be scripting based like BSH
> > > was; mostly all JDEE uses it for is query-answer (eg  what's the FQN for
> > > JFrame?) or commands (run the compiler and then get back to me).
> > >
> > > Phil
> > 
> > ------------------------------------------------------------------------------
> > Stay on top of everything new and different, both inside and 
> > around Java (TM) technology - register by April 22, and save
> > $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
> > 300 plus technical and hands-on sessions. Register today. 
> > Use priority code J9JMT32. http://p.sf.net/sfu/p
> > _______________________________________________
> > jdee-devel mailing list
> > jde...@li...
> > https://lists.sourceforge.net/lists/listinfo/jdee-devel
>
>