Hi there,
The problem of usage information gathering is also a pre-requisite to the implementation of refactoring tools (and xref is actually an example of such a tool). So they provide a good place to look for examples on how to solve this problem (as long as you don't mind using Java instead of elisp for the parsing engine).
Below you will find links to open source refactoring tools for Java . FreeFactor has already an Emacs interface.
https://sourceforge.net/projects/freefactor/
https://sourceforge.net/projects/transmogrify/
http://jrefactory.sourceforge.net/csr-commandline.html
Good luck,
Nascif
> -----Original Message-----
> From: Eric M. Ludlam [mailto:[EMAIL PROTECTED]]
> Sent: Friday, October 25, 2002 2:30 PM
> To: [EMAIL PROTECTED]
> Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]
> Subject: Re[4]: missing features in jde?
>
>
> Semanticdb for the next release is much more advanced than
> the current one. It will come with a command line program to
> build a complete set of databases you can reference.
>
> It still won't help with usage information though.
>
> Eric
>
> >>> Andrew Hyatt <[EMAIL PROTECTED]> seems to think that:
> >
> >So far I haven't found a way to get the information I need
> by Semantic
> >only. I think actually it is theoretically impossible,
> unless we also
> >require semanticdb's to be made for each file. We could do
> it with a
> >combination of Semantic and bsh, but that sort of thing has
> been slow
> >for me in the past (when we use completion, for instance),
> so I'm not
> >too confident about it.
> >
> >I think a better way to do it is to parse the class file
> itself. The
> >Java compiler has already done the heavy lifting here, it
> has the exact
> >fully-qualified functions that are called, we just need to
> go and parse
> >out the relevant information. I think this could be very
> fast. Also,
> >this allows us to retrieve usage information for anything
> that compiles
> >to Java bytecode, and gives us access to classes we don't have the
> >source code to. The disadvantage if we don't get line
> numbers if the
> >class file has no debug info, but we could get the class and
> function
> >name each call is in. I think that's good enough.
> >
> >I'm working on the class-file based approach right now, and
> am building
> >a simple parser for it (I don't think semantic can handle
> binary files,
> >but correct me if I'm wrong). Whether I do a full parser of just a
> >quick parse of the stuff we need I haven't decided yet.
> >
> >
> >
> >"Eric M. Ludlam" <[EMAIL PROTECTED]> writes:
> >
> >> It might be possible to do such a thing with semantic 1.4,
> though it
> >> would be very slow. The new bison based parser David Ponce ported
> >> into the semantic 2.0 CVS also has a full java parser. I'm not
> >> exactly sure what it's capabilities are though. As far as I know,
> >> the CVS version is still compatible with the current
> version of JDE.
> >>
> >> Eric
> >>
> >> >>> Andrew Hyatt <[EMAIL PROTECTED]> seems to think that:
> >> >
> >> >You can actually get the "find usages" functionality with
> the xref
> >> >program, a program not only amazing for this feature, but
> also the
> >> >fact that it is an emacs extension that costs money!
> >> >
> >> >See it at:
> >> >http://xref-tech.com/xrefactory/
> >> >
> >> >It costs $29. I bought it and use it, and it's good, but
> I think a
> >> >better solution could probably be designed for free.
> >> >
> >> >I'd actually like to work on this functionality for jde,
> if it was
> >> >possible to do. I don't know enough about semantic to
> tell whether
> >> >it is powerful enough to support this functionality (all we would
> >> >need is to tag lines by what functions of what classes
> are called,
> >> >then create and store a big table of these), but I think
> so. We can
> >> >take it a step farther than xref by adding functionality
> to produce
> >> >call trees, which would be incredibly useful.
> >> >
> >> >Even though it appears possible, I remember reading on this list
> >> >that it would require semantic 2.0 to do? If anyone has
> any further
> >> >information on that, I'd be interested to hear.
> >> >
> >> >
> >> >
> >> >Paul Kinnucan <[EMAIL PROTECTED]> writes:
> >> >
> >> >> Raul Acevedo writes:
> >> >> > I've been using JDE for a while, and while I'm
> generally happy
> >> >> with it, > there is some basic functionality that it
> seems to be
> >> >> missing: > > * Minibuffer completion on class names for
> >> >> jde-open-class-source. >
> >> >> > * "Find usages" functionality: list all
> callers/users for a given
> >> >> > method, field or class. No, I don't want to use
> grep, I want
> >> >> > something 100% reliable.
> >> >> >
> >> >> > * Functionality to list all classes/interfaces
> which implement/extend a
> >> >> > given class/interface.
> >> >> >
> >> >> > JDE is great, but these features are pretty standard
> on most Java IDEs
> >> >> > and they are quite useful. Are they there and I've
> missed them, or will
> >> >> > they be provided in a future release?
> >> >> >
> >> >>
> >> >> Hi Raul,
> >> >>
> >> >> Thanks for the suggestions. Others have requested some of these
> >> >> features as well. I think they'd be interesting to implement. I
> >> >> have other features that are higher priority for me at
> the moment
> >> >> but when I get some time, I'll take a stab at them.
> Meanwhile, I'd
> >> >> be glad to support anybody who'd like to work on them.
> >> >>
> >> >> Regards,
> >> >>
> >> >> Paul
> >> >
> >
>
> --
> Eric Ludlam: [EMAIL PROTECTED],
> [EMAIL PROTECTED]
> Home: http://www.ludlam.net Siege: www.siege-engine.com
> Emacs: http://cedet.sourceforge.net GNU: www.gnu.org
>
