Re: Import Wizard dialog box in Xemacs

2003-03-22 Thread Jim Crossley
I had this problem, too.  The solution is here:
http://www.mail-archive.com/[EMAIL PROTECTED]/msg06232.html

-- Jim

Bill Barnard [EMAIL PROTECTED] writes:

 I'm successfully using many of the functions in JDE in Xemacs 21.4
 (linux 2.4.17, Red Hat 7.2). Method name completion and the like work
 fine with menu pop-ups.

 The single thing I've been unable to figure out (so far) is how to use
 the dialog buffer (not a pop-up) when I use the Import Wizard to
 generate my import statements when there are several possibilities for
 the class (Iterator).

 The function jde-import-choose-imports appears to be the one that
 displays the dialog buffer, and to my non-Lisp-enabled brain looks
 correct. I see the dialog buffer displaying radio-buttons for three
 possible completions, with java.util.Iterator selected. The mini-buffer
 shows new-import=javax.swing.text.html.HTMLDocument.Iterator. However
 there are no selection buttons (OK/Cancel presumably) so I'm unable to
 interact with the dialog.

 Has anyone else seen and solved this? Perhaps I should set up the JDE to
 run under emacs instead?

 Thanks,

 Bill
 (I will post the xemacs initialization files if necessary, but I hoped
 this might be something simple so I've put that off...)
 -- 
 Bill Barnard [EMAIL PROTECTED]

-- 
Jim Crossley
http://crossleys.org/~jim
404.409.9085


Open class at point for this.var

2003-03-22 Thread Jim Crossley
Recently, I noticed a complaint about receiving an error when trying
to open class at point when point is over an import statement.  This
is one of two situations I wish would work.  The other one is when
member variables are prefixed with this.

  class X {
String s;
X () { 
  s = ;  // I can open s' class with point on s;
  this.s = ;  // I cannot
}
  }

I use the this. notation a lot as a way to avoid silly naming
conventions for member variables.  Is this very hard to implement?

Thanks,
Jim


Missing buttons

2003-02-22 Thread Jim Crossley
I just built Xemacs 21.4.12 and now I can't see the buttons that
appear when a choice of classes results from the function,
jde-import-find-and-import (C-c C-v C-z).  I see other buttons,
e.g. toolbar and customize, just not those.  What build option might I
have left out to cause that?

Thanks
-- 
Jim Crossley
http://crossleys.org/~jim
404.409.9085


Re: attaching with jdb

2002-01-31 Thread Jim Crossley

Scott Evans [EMAIL PROTECTED] writes:

 Oh, wait.  Looks like JDEE doesn't provide a mechanism to attach to a
 running process if you're using jdb?  Shoot.  If that's true, I guess
 I'm back to the command line for now...

If your servlet container is written in Java (most are), another
alternative is to start your container from within JDEE, so that you
don't need to attach to a process.

To step through a servlet deployed with Tomcat, for example, the
following variables should do the trick.

(jde-set-variables
 '(jde-db-source-directories (quote (/usr/local/jdk/src ./src ~/src/struts 
~/src/tomcat ~/src/servlet)))
 '(jde-db-option-properties (quote ((tomcat.home . /usr/local/tomcat
 '(jde-run-option-properties (quote ((tomcat.home . /usr/local/tomcat
 '(jde-global-classpath (quote (/usr/local/tomcat/lib /usr/local/jdk/lib 
/usr/local/tomcat/webapps/{YOUR_APP_NAME_HERE}/WEB-INF/classes ./lib)))
 '(jde-run-application-class org.apache.tomcat.startup.Tomcat))

This assumes that your classes reside under Tomcat's 'webapps'
directory (or your server.xml refers to their directory if
elsewhere).




Re: Help Needed Debugging Tomcat Servelets

2002-01-26 Thread Jim Crossley

Hi Heather.  

If I may present my experiences debugging Tomcat apps
using XEmacs/JDEE...

1) The JDE debugger, while cooler than the jdb, is not quite ready for
primetime, IMHO.  The jdb, while boring, seems to be more reliable
debugging servlets on Linux.

2) I've tried the attach process thing with Tomcat, and though it
seems to work ok, the debugger is more responsive when I simply run
org.apache.tomcat.startup.Tomcat from within the JDEE after having
deployed my app to $TOMCAT_HOME/webapps.

Here is a typical prj.el file for one of my apps:

(jde-project-file-version 1.0)
(jde-set-variables
 '(jde-db-source-directories (quote (/usr/local/jdk/src 
~/personal/prototypes/receiving/src ~/src/struts ~/src/tomcat ~/src/servlet)))
 '(jde-db-option-properties (quote ((tomcat.home . /usr/local/tomcat
 '(jde-run-option-properties (quote ((tomcat.home . /usr/local/tomcat
 '(jde-global-classpath (quote (/usr/local/tomcat/lib /usr/local/jdk/lib 
/usr/local/tomcat/webapps/receiving/WEB-INF/classes 
/usr/local/tomcat/webapps/receiving/WEB-INF/lib)))
 '(jde-run-application-class org.apache.tomcat.startup.Tomcat))

That last variable is the key to simply typing C-c C-v C-d while
visiting one of the Java source files beneath your source tree.  Once
the debugger is running, type C-x SPC in your source file to set your
breakpoints. 

Good luck!

Heather Marie Buch [EMAIL PROTECTED] writes:

 BTW - At no point to I get the 
 
 step into or step over choices. Shouldn't they be available at some
 point after
 I attach the process?
 
 Thanks,
 
 Heather
 
 Jason Rumney wrote:
  
 I am running a servlet application under tomcat (servlet engine) and
 apache with mod_jk.

 I modified the tomcat startup file to contain the debug stuff:

 it starts up:

 java -classic -Xrunjdwp:transport=dt_socket,address=,server=y,
 suspend=n -Xdebug -Dtomcat.home= $MAIN
  
  Presumably you already use a number in place of  above, or it
  would fail on startup as well.
  
 problem #1
 tomcat starts ok, but when I shut it down, it complains:

 err:: Address already in use
 Transport dt_socket failed to initialize, rc = -1
  
  I would recommend setting the debug options in TOMCAT_OPTS in
  startup.sh, not in tomcat.sh. Otherwise the shutdown command tries to
  use the same port for debugging, which results in the above error.
  
  The other idea to use ant would also work, but seems like an
  unnecessary complication to me.
  
 the jdebug documentation tells me what to do if I have solaris +
 jdk1.3, or linux + jdk1.2.2, but it doesn't specifically say what
 to do with linux + jdk1.3. So I follow the instructions for
 solaris + jdk1.3
  
  I think that is the right thing to do.
  
 Execute JDEbug-Processes-Attach Process-Local Host
 I then enter my socket number (which may or  may not be good)
  
  This is the  number in the debug options, yes? Not one of the
  other sockets that Tomcat is configured to use.
  
 Then I start my app. in the browser.

 Then I come back to jdee and nothing happens. I try to do:
  
  Did you set any breakpoints? If not, I wouldn't expect anything
  visible to happen. I have found that you need to set the breakpoints
  after attaching the debugger to Tomcat, or they won't work.
  
 Display - local variable, and I get an error
  
  You can only display local variables when you are stepping through
  (once the breakpoint is hit).
  
  --
  Jason Rumney
 
 -- 
 http://www.mit.edu/~hmbuch/
 
 Heather Marie Buch  [EMAIL PROTECTED]



Re: JDE vs. java-mode

2002-01-24 Thread Jim Crossley

Paul Kinnucan [EMAIL PROTECTED] writes:

 Scott Evans writes:

[...]

   imports to work though; I get an error as such:
   
   Invalid function: (macro . #compiled-function (token
   streamorbuffer optional search-parts search-includes) ...(15)
   [search-includes search-parts streamorbuffer token
   semantic-find-nonterminal-by-function lambda (tok) eq
   ((semantic-token-token tok))] 6 (c:\\Program
   Files\\XEmacs\\xemacs-packages\\lisp\\semantic\\semantic-util.elc
   . 14003))
  
 Classic error that occurs when you compile lisp file A, which
 requires macros defined by lisp file B, without first ensuring that
 lisp file B has been loaded. Either delete the compiled versions of
 the semantic lisp files or use the included makefile to compile
 them.

I get this error, too.  And when I remove the semantic *.elc files,
the error turns to this:

Invalid function: (macro lambda (token streamorbuffer optional
search-parts search-includes) Find all nonterminals with a token
TOKEN within STREAMORBUFFER.  TOKEN is a symbol representing the type
of the tokens to find.  Optional argument SEARCH-PARTS and
SEARCH-INCLUDE are passed to `semantic-find-nonterminal-by-function'.
(backquote (semantic-find-nonterminal-by-function (lambda (tok) (eq
(\, token) (semantic-token-token tok))) (\, streamorbuffer) (\,
search-parts) (\, search-includes

Any other ideas?



Missing jdebug key bindings

2002-01-15 Thread Jim Crossley

I'm running JDEE 2.2.8 on Xemacs 21.4.4.  Everything is working fine
except my JDebug key bindings.  The variable, `jde-bug-key-bindings',
looks correct and the function, 'jde-bug-keys', reports what I would
expect.

But when debugging, I get a C-c C-z not defined error when trying to
step, etc.  Any advice?



Re: Comment formatting

2001-06-18 Thread Jim Crossley

Make sure the value of 'c-syntactic-indentation' is t.

Molitor, Stephen wrote:
 
 I'm using cc-mode 5.28 (which I verified by running 'c-version), and JDE
 2.2.8beta1.  No luck.  Is there something I can customize in JDE or cc-mode?
 
 The problem is that when I run 'indent-region on a Javadoc comment block, it
 makes all the '*''s flush left, which isn't quite what I want.
 
 Thanks!
 
 Steve Molitor
 [EMAIL PROTECTED]
 
 -Original Message-
 From: Mark Bucciarelli [mailto:[EMAIL PROTECTED]]
 Sent: Monday, June 18, 2001 8:54 AM
 To: Molitor, Stephen; [EMAIL PROTECTED]
 Subject: RE: Comment formatting
 
 Try upgrading to the latest cc-mode 5.28.  Note that you need to use
 JDE 2.2.7beta8 (?) or higher to work with 5.28.
 
 --
 Mark BucciarelliPeregrine Systems, Inc.   RD, Boston
 (413) 253-0020  [EMAIL PROTECTED]Yahoo: m_bucciarelli
 http://www.peregrine.com
 
  -Original Message-
  From: Molitor, Stephen [mailto:[EMAIL PROTECTED]]
  Sent: Monday, June 18, 2001 9:43 AM
  To: [EMAIL PROTECTED]
  Subject: Comment formatting
 
 
  I would like JDE to format javadoc comments with a leading
  space before the
  '*' on the 2-n lines, like this:
 
  /**
   *
   *
   */
 
  Instead, it formats it like this:
 
  /**
  *
  *
  */
 
  with everything flush to the left.  It wouldn't bother me
  that much, except
  that we have a coding standard here at work that says to do
  it the first
  way.
 
  Any ideas?  Thanks!
 
  Steve Molitor
  [EMAIL PROTECTED]