AW: Problem getting the methods directory to update

[Berndl, Klaus]

>Well,  I byte-compiled the ecb and the jde, and now all files BUT java
>source files get parsed, and the methods and variables show up in the
>methods window.  That includes DOS batch files, C files, makefiles, lisp
>files.   No java.

>JDE users, I attempted to byte-recompile directory, and it compiled 0 of 0
>files.  Any reason why?

Yes, it is in the documentation. This comand only recompiles a file if it´s
*.elc file is older then the *.el file. If no *.elc file exists, nothing is
done!
For ECB you can use either:
- C-u 0 M-x byte-recompile-directory RET  or
- C-u M-x ecb-bytecompile RET

>I tried using the makefile and my build command
>said emacs.exe was too big to run.  Is there a preferred method of compiling
>the jde?

M-x jde-compile-jde RET

>I was using the PVCS command, build.exe.  Do I need to
>byte-compile the utilities elib, eieio, semantic, etc. first?

Don´t know, but here is the sequence which i byte-compile the packages:

1. elib
2. eieio
3. semantic
4. speedbar
5. JDE
6. ECB

Note, that most of these packages must be byte-compiled via Makefile
or from within Emacs via byte-recompile-directory (see above).
If via makefile then note that most of the makefiles needs correct
paths to other packages onto YOUR filesystem (e.g. JDE and ECB to semantic)
and therefore often you must first edit the makefiles before compiling.

But you can make a simple test. Just remove all *.elc files from
all the package elib, eieio, semantic, speedbar, jde, ecb and try it
again with the setup i have tried with.
If then it works then you have a problem with byte-compiling, if not
then i have no clue why :-)

Klaus

AW: Suggestion about jde-complete-at-point* commands.

[Berndl, Klaus]

>I agree!  Although having the parameter types printed out is occasionally
>useful, 
>most of the time I just want the method name inserted, with the cursor
>positioned
>between the parenthesis.

I second this suggestion!

Klaus

Steve Molitor
[EMAIL PROTECTED]

-Original Message-
From: Eric Chastan [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, June 12, 2001 2:27 AM
To: [EMAIL PROTECTED]
Subject: Suggestion about jde-complete-at-point* commands.


Hello all,

I'm using JDE  (with ECB) very intensively and I have a suggestion about
the jde-complete-at-point* commands:
1) First I think that it is not useful to have several times the
same method with different signatures in the menu because this increases
dramaticaly the menu size and then it is not very efficient to scroll in
it.
2) Second writing the parameter types when the method is inserted is
also very unefficient because this implies to go back, delete types and
rewrite parameters. I think that the better is  to insert the opening
"(" and the closing ")" and to set the cursor between them.

On a other hand providing the parameter types is very useful so the
solution could be to have an other popup that raises when the cursor is
between the "()" that shows all possible parameters types (only showing
them without allowing selection).

What do you think about this suggestion ?
Is it possible to have customizable variables that disable method
signatures in menu and suppress parameter types on insertion ?
 Thanks,
Eric.

AW: JDE vs. VAJ

[Berndl, Klaus]

>> Is the purpose of Emacs and JDE mostly to give a nice environment 
>> for those 
>> users who don't have much space, speed, and don't have admin 
>> rights on their 
>> machine?
>In my case, none of the above, 
>All the IDE I have tried(Visual Cafe, JBuilder)
>are incredible slower compare to emacs(I have a pIII 850 btw).
>Do not have the editing power than emacs does,
>the emacs key strokes binding does not do it for me.
>They tend to crash often, are memory hugs, 
>and I could keep going.

>In brief, JDE offers most of the features(or more in some cases)
>of graphical IDE's plus the ability to fully use emacs.

Just combine JDE with ECB (http://home.swipnet.se/mayhem/ecb.html go to
the screenshots-section) and you get even the graphical IDE.

Klaus

AW: Problem getting the methods directory to update

[Berndl, Klaus]

>Well,  I think I found it, although I don't quite understand what's going
>on.  The following code:

>(defun my-java-hook ()
>  "My local settings for Java"
>  (c-set-style "java")
>  (c-set-offset 'substatement-open 0)
>  )
>(setq java-mode-hook 'my-java-hook)

>appeared to disable Senator-minor-mode, which resulted in the java source
>not being parsed (I'm assuming), which in turn resulted in the *ECB Methods*
>window not being populated.

>Can anyone explain this?

Yes, i can. All your assuming go into wrong direction, parsing a buffer
has nothing to do with senator-minor-mode.

This is the wrong statement: (setq java-mode-hook 'my-java-hook)
Change this to (add-hook 'java-mode-hook 'my-java-hook)
and all is working. 
Reason: Semantic and JDE add both also functions to 'java-mode-hook and they
do this before you add 'my-java-hook. Or, let´s wait, "add" is the wrong phrase
for this what you do. With setq you don not add a function to the hook but you
overwrite the old value and set only your function to the hook.

The java-mode-hook must at least contain
(jde-setup-syntax-coloring semantic-default-java-setup), if semantic and JDE
are correct set up. The function 'semantic-default-java-setup for example
does all necessary, that java-buffers will be parsed.

If you add functions to a hook, you MUST always use add-hook instead of setq!!

Hope this helps,
Klaus

RE: Discussion about the ECB-layout - Development needs feedback from the users

[Berndl, Klaus]

>Most of the people who have chimed in have mentioned they could live w/o the
>long compilation window at the bottom, but I've grown to like it.  When
>coding in java, source file path names tend to be long -
>/path/to/source/com/companyname/application/module/subpackage/FileToEdit.jav
>a - just the filename alone takes up 2/3 to 3/4 the width of my screen.
>Because of this I have

>(add-hook 'compilation-mode-hook #'(lambda () (setq truncate-lines nil)))

I have the following:

(defun line-wrap-toggle (&optional arg)
  "Toggle the line-wrap (line-truncate) function.
Covers (and equates) both horizontal and vertical splits.
If ARG > 0 then activate line-wrapping, if ARG < 0 then deactivate it. If
ARG is nil then toggle."
  (interactive "P")
  (if (and arg (numberp arg))
  (setq truncate-lines (if (> arg 0) t nil)))
  (setq truncate-lines (setq truncate-partial-width-windows (not
 truncate-lines)))
  (recenter)
  (if (interactive-p)
  (message (if truncate-lines
   "truncating lines (as '... $')"
 "wrapping lines (as '...\\')")))
  )

(add-hook 'compilation-mode-hook #'(lambda () (line-wrap-toggle 1)))

>in my .emacs so I can see all of the error messages without having to scroll
>(I usually have truncate-lines set to t for all other modes).  (Aside of my
>own - auto-show-mode sometimes doesn't work in the ecb edit window, and
>turning truncate-lines off doesn't make the lines wrap - does anyone else
>see this behavior?)

No, with my setting above i have never seen this behavior.

>So, getting back to my point, I like being able to read more of the error
>message without having to follow the wrapped line off of the right side and
>continuing on the left.  Also other temp buffers, like *buffer-selection*
>view nicely for me with a long compile window.

For temp-buffers and or *Buffer-Selection* you can do:

(add-hook 'help-mode-hook #'(lambda () (line-wrap-toggle 1)))

This works for the command 'list-buffers:
(add-hook 'buffer-menu-mode-hook #'(lambda () (line-wrap-toggle 1)))

IMO each well designed mode which displays buffers like compilation, temp-buffers,
in general lists, where i can choose, click one should/must offer a hook which
is called after preparing the buffer.
In this hook you can always add the line-wrapping code!

In all other points you are right, in compilation-mode buffers and temp-buffers
all text should be readable without scrolling. But even with a window which has
the whole frame-width you will get java-compiler messages (errors and warnings)
which will exceed one line-length.

So, with *every* window-width your lines must be wrapped in such a window and IMHO
it doesn´t then really matter if the window has frame-width or something smaller.
With the three hooks i have described above you should always get line-wrapping
in a compilation output or a temp-buffer.

Therefore IMHO the compilation-window with whole frame-width is not really needed.
But of course this is a question of personal taste!

What do you think?

Klaus

RE: Discussion about the ECB-layout - Development needs feedback from the users

[Berndl, Klaus]

> > in my .emacs so I can see all of the error messages without having to scroll
> > (I usually have truncate-lines set to t for all other modes).  (Aside of my
> > own - auto-show-mode sometimes doesn't work in the ecb edit window, and
> > turning truncate-lines off doesn't make the lines wrap - does anyone else
> > see this behavior?)

> No, with my setting above i have never seen this behavior.

Here i must write some words more:

AFAIK auto-show-mode is obsolete since Emacs 20.X. Use hscroll-mode or
hscroll-global-mode!
At least the doc-string in Emacs 21 says: auto-show-mode: obsolete.

Klaus

Start of the Beanshell

[Berndl, Klaus]

Hi!

I found it very annoying that the first beanshell start takes a long time.
Meens if you do your first completion for example (or help for symbol at point), then 
you must
wait a veeery long time :-)

IMHO there exists a also very simple solution for this problem:
Why not starting the beanshell in the *background* with an asynchronous process the 
first-time
a jde-mode buffer is openend. Then i bet in almost 100% the beanshell is already there 
if a user
needs it first time (completion, help etc.)

What do you think?

Klaus


-- 
Klaus Berndlmailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

RE: Turning off mouse hooks on ECB

[Berndl, Klaus]

First: I will inspect the code carefully where not a cleanup is done because this 
should
be done otherwise it is a bug.
 
But second: Which version of ecb you are using? I suppose something <= 1.32?! Ok, 
latest 1.40beta2 version
which is available via CVS has a lot of improvements: here is the HISTORY-file.
You see, there are menus, keybindings, minor-modes etc. AND there is a 
toggle-visibilty-ecb-windows command.
Therefore you should not activate/deactivate ECB but only toggle the visibility of the 
ecb-windows. But this works
only with latest 1.40beta2. maybe we should realease soon 1.40?!
 
Klaus
 
HISTORY:
Version 1.40:
-
 
- New advice for 'other-window-for-scrolling', so all scroll-functions for the
  "other" window can also scroll the first edit-window if point stays in the
  second one. (Klaus)
 
- ecb-toggle-ecb-windows now preserves the split, the split-amount, the buffer
  contents, the window starts and current point. (Klaus)
 
- Some important changes in the ECB-layout concerning displaying compilation-
  and temp-buffers. (Klaus):
 
  + 'ecb-select-compile-window' has been gone.
  
  + 'ecb-use-dedicated-windows' is not longer a user-customizable option but
it is always set to t, because this is essential for correct working of
the ecb-layout engine!
 
  + 'ecb-compile-window-temporally-enlarge' is now a boolean option.  
 
  + Now ECB does all necessary that your edit-window of ECB seems to be a
normal Emacs frame, especially concerning displaying temp-buffers (like
help- and completion-buffers) and compilation-buffers.

  + The variables 'compilation-window-height' and 'temp-buffer-shrink-to-fit'
(XEmacs) and 'temp-buffer-resize-mode' (GNU Emacs) are now fully supported
by ECB if no durable compilation-window is shown.
 
- New function to toggle visibility of the ECB windows. Now it´s possible to
  hide all ECB windows without deactivating ECB (see 'ecb-toggle-ecb-windows')
  (Klaus).
 
- ECB is now a global minor mode with it´s own menu "ECB" and it´s own keymap
  with prefix "C-c .". New function to (de)activate/toggle ECB with
  'ecb-minor-mode'. (Klaus)
 
- Fixed a bug in highlighting current token in the method-buffer when
  font-lock-mode is nil for the source-buffer. (Klaus)
 
- New option for highlighting the method-header in the edit-window after
  clicking onto the method in the method-window (like Speedbar does). (Klaus)
 
- Function for automatically submitting a problem report to the ECB mailing
  list: ecb-submit-problem-report. (Klaus)
 
- Now not only for method-highlighting an idle delay can be set but also for
  synchronizing the ECB windows with current edit window (see option
  ecb-window-sync and ecb-window-sync-delay; default is 0.25 sec delay)
  (Klaus).
 
- Smarter highlighting of current method (Klaus).
 
- All tree-buffers now have as default-directory the current selected
  directory in the directory buffer. So you can also open files with find-file
  etc. from within the tree-buffers. (Klaus).
 
 
 
 Hi,
I've been using ECB a lot w/ JDE. Typically I activate and deactivate it many 
times during an Emacs session (to take a better look at shell buffers, dired windows, 
etc.), and have even created keybinds for that (below).
 
The problem is, ecb-deactivate don't seem to do a complete clean-up, specially in 
the mouse hooks. After I activate/deactivate ECB I start to have very strange mouse 
(the so called "jump to right destination" behavior, I believe) and frame (popping up) 
behavior.
 
Is there anyway to configure ECB to *really* deactivate the additional mouse/frame 
behavior?
 
Regards,
Nascif
 
;; ECB keybindings
(global-set-key "\C-cea" 'ecb-activate)
(global-set-key "\C-ced" 'ecb-deactivate)
(global-set-key "\C-ceu" 'ecb-update-directories-buffer)
(global-set-key "\C-cec" 'ecb-current-buffer-sync)
(global-set-key "\C-cem" 'ecb-rebuild-methods-buffer)
(global-set-key "\C-cel" 'ecb-redraw-layout)
(global-set-key "\C-cec" 'ecb-clear-history)
(global-set-key "\C-ceh" 'ecb-show-help)

RE: How can I make the JDE smaller ?

[Berndl, Klaus]

>So I have two questions :

>1) Is there a way to disable all the new functinnalities
>(speedbar, eieio, semantic ...) ?

No, latest JDE-version need all these packages!

>2) If it is not possible, where can I find some old JDE 2.1.* 
>versions ?

ftp://ftp.sunsite.auc.dk/projects/jde

>I know the new functionnalities are wonderful (yes, I believe it)
>but I just want :
>- syntax coloring
>- auto indentation
>- compile menu and compile error to source links
>So, please, tell me if you know how I can get a 
>"light JDE version". I know I am not the only one in this case.

Hmm, i think for only these tasks you don´t need any JDE because all
these goals you can reach very well with the standard java-mode of cc-mode and the
standard compile-features (compile, next-error etc.) of Emacs...

Klaus

RE: Method Navigation

[Berndl, Klaus]

>1. use standard emacs search capabilities
>2. use jde classes menu
>3. use ecb methods buffer
>4. use find-tag
>5. use speedbar

>Here's why I don't like any of the above options.  Ecb and the jde classes
>menu require the mouse.  Mouse usage is inefficient.  Find-tag requires
>frequent updating of the tags file and requires knowing and correctly
>spelling the full method name.  Searching usually results in multiple hits
>due to method calls.  The speedbar shows variables and method names and
>takes up too much screen real estate.

I can only speak for ECB: latest ECB-beta (1.40beta2) from the CVS directory
is a minor-mode with it´s own keymap. therefore ECB now is very well usable
only with the keyboard. Example:
Jump with "C-c . m" in the method-window and then use there the special
ECB incremental-search feature (see "Working with the keyboard in the ECB-buffer:"
of the online-help) to jump to a method/variable.

>Are there other ways to more efficiently locate a method?

Yes, for example senator-jump or the senator-incremental-search feature.

>Probably, but what I am looking for is described below.

>I would like to see the list of methods for the current buffer in the
>minibuffer by pressing some C-[some key].  While the list is presented in
>the minibuffer I can narrow the list of methods names that I am interested
>in by typing a portion of the method name.  Once I have the method name that
>I am interested in at the head of the list I can select it with the enter
>key.  Selecting the method name takes me to the method declaration in the
>current buffer. This is similar to what is provided by ido.el and
>iswitchb.el packages which provide fast buffer and file switching.

>Is there anything like this in existence today?  If not, does anyone think
>this would be a good addition to the jde?

No. this would be a good feature but not to JDE, because such a feature is 
language-independent!

But with ECB you can get something similar like 'list-buffers' but for methods/
variables. See section "Working with small screens:" in the online-help.
I know that you have not asked for small screens but IMHO this solution
fulfills your wishes very well.

But if you want to use the ECB-things i have described above you should download
latest CVS-beta (it´s very stable).

Klaus

RE: jde and ecb

[Berndl, Klaus]

>Please use the following packages:

>- ecb 1.40
>- semantic 1.4beta8
>- latest JDE
>- EIEIO latest beta or 0.16 should work

No, latest eieio-beta works NOT with latest JDE! (see requirements of JDE)

Klaus

/Jesper Nordenberg

RE: jde and ecb

[Berndl, Klaus]

>Well, here's a checklist:

>1. Check that you've added (require 'jde) to your .emacs. 
>2. Check that JDE mode is activated when you load a .java file.
>3. Check that you don't have any old versions of the packages or old .elc files in 
>your load-path.

4. Check what you have into the hooks jde-mode-hook, java-mode-hook, c-mode-common-hook
If you have something in these hooks, try to successive remove all this stuff until ECB
works with java-files.

If nothing works you can do:
1. make a .emacs file which adds only the load-pathes of the packages ECB and JDE 
require
2. Remove all *.elc files from the packages
3. start Emacs with "emacs -q -no-site-file"
4. Load the libraries with:
   - M-x load-library RET jde
   - M-x load-library RET ecb
5. Add some source-pathes to ECB
6. Open your java-source.

If that does not work you have a serious problem :-)

Klaus

>Good luck!

/Jesper Nordenberg

RE: Renaming Proposal

[Berndl, Klaus]

>JIDE sounds great !

At least better than JMode...

Klaus

>Steff

On Wednesday 18 July 2001 17:42, Iain Huxley wrote:
> I'm sure you're sick of naming comments etc. by now, and it may be too
> late, but:
>
> In some ways using the word "mode" understates what JDE can do for you - it
> makes it seem like just some formatting helper; the rich set of features is
> better indicated by "development environment".
>
> so, here are a couple more possibilities which retain that part:
>
>   JIDE [java integrated development environment]
>   eJIDE [emacs ...]
>   j-ide
>   Java-IDE
>
> Iain.
>
> > -Original Message-
> > From: Latchezar M. Dimitrov [mailto:[EMAIL PROTECTED]]
> > Sent: Wednesday, July 18, 2001 2:29 PM
> > To: Paul Kinnucan
> > Cc: [EMAIL PROTECTED]
> > Subject: Re: Renaming Proposal
> >
> >
> > Hi,
> >
> > I like jmode too. How about slight change with big consequencies?:-)
> > That is, jdm (java development mode of course)?
> >
> > Thanks and sorry if it's too late.
> >
> > Latchezar
> >
> > PS. BTW, I've made a suggestion for j-d-e earlier but haven't
> > heard any
> > comments?
> >
> > >Mailing-List: contact [EMAIL PROTECTED]; run by ezmlm
> > >From: William L Anderson <[EMAIL PROTECTED]>
> > >Date: Wed, 18 Jul 2001 12:15:11 -0400
> > >To: Paul Kinnucan <[EMAIL PROTECTED]>
> > >Cc: [EMAIL PROTECTED]
> > >Subject: Re: Renaming Proposal
> > >
> > >
> > >  jmode- is a good choice; could use jm- for an abbreviation if it's
> > >not already in use in another emacs package.
> > >
> > >-WLA
> >
> > Latchezar (Lucho) Dimitrov home: (336) 794-2094
> >
> > mailto:[EMAIL PROTECTED]http://www.wfu.edu/~dimil01g/
> >
> > ... If you think education is expensive, try ignorance.
> > -- Derek Bok, president of Harvard

RE: Method Navigation

[Berndl, Klaus]

>I did download the latest source files from the CVS web page (is that how I
>should obtain the beta release you mentioned)?  I did try it and I think
>this is a good solution, but I ran into a couple of difficulties.

>I could not compile the source with the latest Makefile file. I get the
>following error:

>/tools/emacs-site/ecb-1.40beta] make
>Byte-compiling ECB with LOADPATH=../semantic-1.4beta7 ../jde-2.2.8beta1/lisp
>...
>emacs -batch -no-site-file -l ecb-compile-script --eval '(ecb-byte-compile
>t)'
>Loading wid-browse...
>Symbol's value as variable is void: speedbar-dynamic-tags-function-list
>Signal 127
>make: *** [all] Error 255

Maybe you have a too old speedbar version installed. Which version of
speedbar you use?

>The other problem is that C-c . is bound in jde-mode to c-set-style, which I
>believe prevents using the C-c . as a prefix key.

C-c . is bound to c-set-style in cc-langs.el. I´m using this too but for
me this doesn´t shadow the ECB prefix?! Maybe it related how and when you load
cc-mode?! I have a (require 'cc-mode) before the (require 'ecb).

But you are right. The prefix should be customizable and - what a luck - in the 
latest CVS sources it is (see option ecb-prefix-key).
You can download latest ECB sources from the CVS (available via ECB website)
and then you can configure the ECB prefix.
Current CVS sources are very stable!

Klaus



-Original Message-
From: Berndl, Klaus [mailto:[EMAIL PROTECTED]]
Sent: Friday, July 13, 2001 3:57 AM
To: 'Louis Aloia'; '[EMAIL PROTECTED]'
Cc: 'ECB-Mailing List'
Subject: RE: Method Navigation


>1. use standard emacs search capabilities
>2. use jde classes menu
>3. use ecb methods buffer
>4. use find-tag
>5. use speedbar

>Here's why I don't like any of the above options.  Ecb and the jde classes
>menu require the mouse.  Mouse usage is inefficient.  Find-tag requires
>frequent updating of the tags file and requires knowing and correctly
>spelling the full method name.  Searching usually results in multiple hits
>due to method calls.  The speedbar shows variables and method names and
>takes up too much screen real estate.

I can only speak for ECB: latest ECB-beta (1.40beta2) from the CVS directory
is a minor-mode with it´s own keymap. therefore ECB now is very well usable
only with the keyboard. Example:
Jump with "C-c . m" in the method-window and then use there the special
ECB incremental-search feature (see "Working with the keyboard in the
ECB-buffer:"
of the online-help) to jump to a method/variable.

>Are there other ways to more efficiently locate a method?

Yes, for example senator-jump or the senator-incremental-search feature.

>Probably, but what I am looking for is described below.

>I would like to see the list of methods for the current buffer in the
>minibuffer by pressing some C-[some key].  While the list is presented in
>the minibuffer I can narrow the list of methods names that I am interested
>in by typing a portion of the method name.  Once I have the method name
that
>I am interested in at the head of the list I can select it with the enter
>key.  Selecting the method name takes me to the method declaration in the
>current buffer. This is similar to what is provided by ido.el and
>iswitchb.el packages which provide fast buffer and file switching.

>Is there anything like this in existence today?  If not, does anyone think
>this would be a good addition to the jde?

No. this would be a good feature but not to JDE, because such a feature is 
language-independent!

But with ECB you can get something similar like 'list-buffers' but for
methods/
variables. See section "Working with small screens:" in the online-help.
I know that you have not asked for small screens but IMHO this solution
fulfills your wishes very well.

But if you want to use the ECB-things i have described above you should
download
latest CVS-beta (it´s very stable).

Klaus

RE: Fontification problem

[Berndl, Klaus]

>NTemacs 20.5.1
>JDE 2.2.7.1
>semantic 1.4 beta1

>When I run a compilation from a Java buffer, the compilation buffer is
>not fontified -- same problem with grep. Fontification of the
>compilation buffer is only done the first time I select a buffer which
>is not in jde-mode.

>I have already seen messages about that problem in the mailing list, but
>I don't remember if a solution has been given.

Yes, indeed. I have already posted the same problem but until now - no
solution, even no answer.

If i remember right this problem first occurs with JDE >= 2.2.7. With JDEs
<= 2.2.6 it works. I had posted the suspicion that this can something have
to do with all the fontification-features added to JDE since 2.2.7, but i don´t
know exactly, only a - maybe wrong - suspicion.

This is really annoying and it would be nice if this bug could be fixed.

Ciao,
Klaus

RE: Fontification problem

[Berndl, Klaus]

Yes, thank you, this workaround works!

But nevertheless there is a bug about this within JDE...
I´m using already latest JDE (2.2.8) and Emacs (20.7.1)

Klaus

-Original Message-
From: Louis Aloia [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, August 28, 2001 7:33 PM
To: Berndl, Klaus; 'ABAK Systèmes'; [EMAIL PROTECTED]
Subject: RE: Fontification problem


I encountered this problem a long time ago and I believe I worked around it
by adding these two lines to my .emacs file.

(add-hook 'compilation-mode-hook 'turn-on-font-lock)
(global-font-lock-mode t)

I still have these lines in my .emacs file and I may be that's why I haven't
had this problem since.

You can try this as a quick, but I suggest you upgrade to the latest jde and
NTEmacs 20.7.1

-----Original Message-
From: Berndl, Klaus [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, August 28, 2001 12:13 PM
To: 'ABAK Systèmes'; [EMAIL PROTECTED]
Subject: RE: Fontification problem


>NTemacs 20.5.1
>JDE 2.2.7.1
>semantic 1.4 beta1

>When I run a compilation from a Java buffer, the compilation buffer is
>not fontified -- same problem with grep. Fontification of the
>compilation buffer is only done the first time I select a buffer which
>is not in jde-mode.

>I have already seen messages about that problem in the mailing list, but
>I don't remember if a solution has been given.

Yes, indeed. I have already posted the same problem but until now - no
solution, even no answer.

If i remember right this problem first occurs with JDE >= 2.2.7. With JDEs
<= 2.2.6 it works. I had posted the suspicion that this can something have
to do with all the fontification-features added to JDE since 2.2.7, but i
don´t
know exactly, only a - maybe wrong - suspicion.

This is really annoying and it would be nice if this bug could be fixed.

Ciao,
Klaus

RE: scrolling compile buffer

[Berndl, Klaus]

setting compilation-scroll-output to nil definitly stops scrolling
the output. I have tested it wirh latest JDE, GNU Emacs 20.7 and
some java (compiling and igrep) and elisp (igrep) files.

If it works not for you you must have somewhere some elisp-code
which does something with your compilation-buffer (maybe
added something to the hooks (jde-mode-hook or compilation-mode-hook);
maybe this code sets compilation-scroll-output (temporally) to not nil?!)

You can do a little check; add this to your .emacs:

(add-hook 'compilation-mode-hook (function (lambda ()
 (message
 "Compilation-scroll-output: %s"
 compilation-scroll-output

Then look into the *Messages* buffer after compilation has finished and
check what the output says...

Klaus

-Original Message-
From: Richard Giuly [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, September 04, 2001 7:31 PM
To: [EMAIL PROTECTED]
Subject: Re: scrolling compile buffer


I still haven't been able to stop the scrolling.


compilation-scroll-output is set to nil, the compilation buffer still 
scrolls

>M-x customize-apropos-options
>In response to Customize regexp: type
>compilation-scroll
I get "no matches"

I'm using xemacs 21.4 under win2000 and the jde that came bundled with it.


(C-x `  is really useful, thanks Florian, I'd still like to turn off 
scrolling though)




Richard Giuly
[EMAIL PROTECTED]



_
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

RE: Control structures - differences between emacs and xemacs

[Berndl, Klaus]

As in the JDEE Docu explained(!) you have to set the option 'tempo-interactive' to a 
non nil value.
Seems as if XEmacs has this as default whereas Emacs has nil as default-value.
 
But a hint for Paul: Maybe this option should be placed onto a more eyecatching place 
within the
whole related chapter!
 
here is the docu of this option:
 
-
tempo-interactive's value is 
nil
 
Documentation:
*Prompt user for strings in templates.
If this variable is non-nil, `tempo-insert' prompts the
user for text to insert in the templates
 
You can customize this variable.
 
Defined in `tempo'.
-
 
Hope this helps,
Klaus
 
-Original Message-
From: Ralph Jorre [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, September 05, 2001 3:38 PM
To: [EMAIL PROTECTED]
Subject: Control structures - differences between emacs and xemacs


There seems to be a difference between xemacs and emacs as regards control structures 
such as "if' and "for" etc. 

In xemacs on  after a control structure the control expression can be entered 
in the mini-buffer. 


In emacs both on  and  the default behaviour seems to happen - ie the 
control structure is copied into the file without the possibility of 
including the the control expression. 


Has anyone else noticed this behaviour or found a solution. 


Regards Ralph 

-- 
  

RE: disabling method highlighting in new JDE

[Berndl, Klaus]

>> Hello,
>> 
>> I've just installed and downloaded JDE 2.2.8 and installed it in Emacs 21.1.1.
>> The new JDE highlights a Java method when you make any change.  The highlighting
>> disappears when the .java file is saved.

>My version does not do this?  What did you do to enable this functionality?

> I find this feature annoying.  How is it activated?  So that I can then figure
> out how to disable it.  I couldn't find this in the documentation.

>Actually I kind of think it would be cool :)

This is done if you set
(setq semantic-load-turn-everything-on t).

This call turns on semantic-show-dirty-mode.

Just do (semantic-show-dirty-mode -1) and this feature
will disappear. I have already mailed Eric that IMHO this feature should
be disabled per default!

Klaus

RE: Semantic 1.4beta11

[Berndl, Klaus]

>I installed the new package, and tried it on a C header file.  I have
>several:

>typedef struct
>{
>   blah,
>   blah,
>   blah
>
>} blah_S;

>like definitions in the file.  The *ECB Methods* buffer shows a:

>   ^@

>for each.  It looks like:

>--
>^@
>^@
>^@
>^@
>^@
>[Variables][+]
>--


>Anyone else seen this?

Yes, me too.
It´s not a problem of ECB but of semantic; speedbar shows the same output.

Eric, do you know why this happens?

Klaus

Help->Symbol at point failed with latest beta

[Berndl, Klaus]

Please enter the details of your bug report here

Just upgraded from 2.2.9beta6 to beta7. With beta6 Help->Symbol at point works
but not with beta7.

example:

public void method (String str)
{
   return;
}

Put point onto "String" and call the menu "Help->Symbol at point".
With beta6 the javadoc of class "String" was opened in my IE5.5 and with beta7
i get only the message "Symbols function definition is void: nil"
(unfortunatelly no backtrace available with debug-on-error!!)

Klaus

Emacs  : GNU Emacs 21.1.1 (i386-msvc-nt4.0.1381)
 of 2001-10-22 on buffy
Package: JDE version 2.2.9beta7

current state:
==
(setq
 jde-gen-session-bean-template '("(jde-import-insert-imports-into-buffer "
 "  (list \"javax.ejb.*\""
 "\"java.rmi.RemoteException\"))" "'>"
 "(jde-wiz-update-implements-clause \"SessionBean\")" 
"'>" "(jde-gen-method-signature" "  \"public\"" "  \"void\"" "  \"ejbActivate\"" "  
nil" "  \"RemoteException\"" " )" "'>" "(if jde-gen-k&r " " ()" " 'n)" "\"{\"'>'n" 
"\"}\"'>'n 'n" "(jde-gen-method-signature" "  \"public\"" "  \"void\"" "  
\"ejbPassivate\"" "  nil" "  \"RemoteException\"" " )" "'>" "(if jde-gen-k&r " " ()" " 
'n)" "\"{\"'>'n" "\"}\"'>'n 'n" "(jde-gen-method-signature" "  \"public\"" "  
\"void\"" "  \"ejbRemove\"" "  nil" "  \"RemoteException\"" " )" "'>" "(if jde-gen-k&r 
" " ()" " 'n)" "\"{\"'>'n" "\"}\"'>'n 'n" "(jde-gen-method-signature" "  \"public\"" " 
 \"void\"" "  \"setSessionContext\"" "  \"SessionContext ctx\"" "  
\"RemoteException\"" " )" "'>" "(if jde-gen-k&r " " ()" " 'n)" "\"{\"'>'n" "\"}\"'>'n 
'n" "(jde-gen-method-signature" "  \"public\"" "  \"void\"" "  
\"unsetSessionContext\"" "  nil" "  \"RemoteException\"" " )" "'>" "(if jde-gen-k&r " 
" ()" " 'n)" "\"{\"'>'n" "\"}\"'>'n 'n" "'>")
 jde-gen-beep '("(end-of-line) '&" "\"Toolkit.getDefaultToolkit().beep();\"'>'n'>")
 jde-complete-signature-display '("Eldoc")
 jde-project-name "default"
 jde-which-method-format '("[" jde-which-method-current "]")
 jde-run-classic-mode-vm t
 jde-complete-unique-method-names nil
 jde-javadoc-gen-nodeprecatedlist nil
 jde-which-method-max-length 20
 jde-package-search-classpath-variables '(jde-compile-option-classpath
  jde-global-classpath)
 jde-imenu-include-classdef t
 jde-javadoc-gen-link-online nil
 jde-gen-code-templates '(("Get Set Pair" . jde-gen-get-set)
  ("toString method" . jde-gen-to-string-method)
  ("Action Listener" . jde-gen-action-listener)
  ("Window Listener" . jde-gen-window-listener)
  ("Mouse Listener" . jde-gen-mouse-listener)
  ("Mouse Motion Listener" .
  jde-gen-mouse-motion-listener)
  ("Inner Class" . jde-gen-inner-class)
  ("println" . jde-gen-println) ("beep" .
  jde-gen-beep)
  ("property change support" .
   jde-gen-property-change-support)
  ("EJB Entity Bean" . jde-gen-entity-bean)
  ("EJB Session Bean" . jde-gen-session-bean))
 jde-gen-cflow-else '("(if (jde-parse-comment-or-quoted-p)" " '(l \"else\")"
  " '(l '> \"else\"" " (if jde-gen-k&r "
  "   jde-gen-conditional-padding-3 " "  '>'n)"
  " \"{\"'>'n'>'r'n" " \"}\"" " (if jde-gen-comments "
  "  '(l \" // end of else\"))" " '>'n'>)" " )")
 jde-make-args ""
 jde-jdk-registry nil
 jde-javadoc-gen-destination-directory "JavaDoc"
 jde-mode-line-format '(#("-" 0 1
  (help-echo
   "mouse-1: select window, mouse-2: delete others,
   mouse-3: delete ...")
  )
mode-line-mule-info mode-line-modified
mode-line-frame-identification
mode-line-buffer-identification
#("   " 0 3
  (help-echo
   "mouse-1: select window, mouse-2: delete others,
   mouse-3: delete ...")
  )
global-mode-string
#("   %[(" 0 6
  (help-echo
   "mouse-1: select window, mouse-2: delete others,
   mouse-3: delete ...")
  )
(:eval (mode-line-mode-name)) mode-line-process
minor-mode-alist
#("%n" 0 2
  (help-echo "mouse-2: widen" local-map
   (keymap (mode-line keymap (mouse-2 .
   mode-line-widen
  )
  

JDE does not accept latest semantic beta13

[Berndl, Klaus]

The variable jde-semantic-max-version has the value
"1.4beta12" so that JDE does not accept latest beta13 of semantic
if jde-check-version-flag is not nil.

IMHO the value of jde-semantic-max-version should be changed to "1.4beta13"  because 
latest JDE seems
to work properly with latest semantic beta13.

Ciao,
Klaus


-- 
Klaus Berndlmailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

RE: Speedbar question (slightly OT)

[Berndl, Klaus]

>Two months ago or so I upgraded to the latest JDE beta and along the way I 
>upgraded speedbar to 0.14beta2 (as well as other required components). 

>After the upgrade, I cannot "expand" java files on speedbar - when I click on 
>the "+" to the left or "Test.java(etc)" it changes to "-" but I do not see 
>the fields and methods in "Test.java". I looked through (and played with) 
>customizing the speedbar to get the "expandable" behaviour with no luck.

>Would anyone here know how to "fix" this?

Don't know, but maybe you use a wrong semantic-version?!

Anyway, maybe you are also interested in trying out the ECB (Emacs Code Browser) which 
gives you all what speedbar gives you (ok, not all, but almost :-), but well 
integrated into your Emacs-frame and with better looking browsing-display (IMHO!).

Klaus

RE: JDE-2.2.9beta8 available at

[Berndl, Klaus]

AFAIK semantic-add-minor-mode is added first in beta12 of semantic.
Am i right?

Klaus

>When loading emacs fails to load giving me the following message:

>Error in init file: Symbol's function definition is void:
>semantic-add-minor-mode

>I am using:

>(setq load-path (cons (concat add-on-path "elib-1.0") load-path))
>(setq load-path (cons (concat add-on-path "eieio-0.16") load-path))
>(setq load-path (cons (concat add-on-path "speedbar-0.14beta1") load-path))
>(setq load-path (cons (concat add-on-path "semantic-1.4beta11") load-path))
>(setq load-path (cons (concat add-on-path "jde-2.2.9beta8/lisp") load-path))
>(setq load-path (cons (concat add-on-path "pcl-cvs-2.9.9") load-path))
>(setq load-path (cons (concat add-on-path "xae-1.0beta6/lisp") load-path))
>(setq load-path (cons (concat add-on-path "ecb-1.60") load-path))

RE: JDE-2.2.9beta8 available at

[Berndl, Klaus]

>OK.  I downloaded semantic-1.4beta13, and the JDE-2.2.9beta8 load OK.
>Klaus, do you recommend upgrading to eieio-0.17beta3?
>Jeff

I'm using eieio17beta3 and it works without problems but IMHO
there is no need for an upgrade if you don't need the new features
of eieio itself: AFAIK neither JDE nor ECBN need them, eieio16 is enough.

>BTW, I am still using NTEmacs 20.7.  Seeing some of the issues made me
>reluctant to try 21.1 yet.

Hmm, i don't know what you make reluctant but i was first also reluctant
but after trying it i must say, Emacs 21.1 is really stable (at least the
NT version, with solaris  i have not tested very intensive.) and
some important stuff is simply spoken better than in 20.7:
- much better scrolling especially if there a horizontal splits as in ECB
- the ugly track-mouse stuff of 20.7 is not needed anymore by ECB cause of
  the new help-echo-property of Emacs 21 ehich is a much better mechanism.
- With Emacs 21 ECB can display the ECB-buffers with smaller font than edit-window
  so you see more functions etc. at a glance.
- etc...

Klaus

RE: JDE-2.2.9beta8 available at

[Berndl, Klaus]

>I didn't realize that this release requires semantic-1.4beta12
>and eieio-0.17beta3. This was pointed out to me by Karsten
>Ensinger. Thanks Karsten.

I would strongly recommend using latest beta13 and not beta12 of
semantic, because beta12 has really some drawbacks related to
the new auto. reparsing feature. I have discussed this a long with
Eric and David and the result is beta13 which is much safer with
respect to this point for itself and there is also a new really
helpful hook 'semantic-before-toplevel-bovination-hook.

IMHO all tools needing a semantic-version >= beta12 should insist
in beta13!

Here is what i have first mailed Eric and David about the beta12 version:

;; - currently parsing a non finished "defun" results in complete wrong
;;   parsing in the rest of the file. Therefore the auto-parse feature is
;;   quite unusable because if you code a function/method which is of course
;;   not parsable during coding (in case not all closing parens, braces etc.
;;   are already set correct) and you stop coding (because you have to reflect
;;   some aspects) for a longer time than the idle time you define in
;;   `semantic-auto-parse-idle-time' semantic parses the not finished
;;   function/method, this parsing fails probably and therefore all the rest
;;   of the file is either parsed completely wrong (java) or not at all
;;   (elisp). It would be much better if semantic would recognice the next
;;   beginning of a defun (elisp: (defun...) in column 0; C/C++/java: not so
;;   simple but should be possible) and then parse from the next defun so only
;;   the current coded function/method can't be parsed but the rest of the
;;   file is still parsed correctly.

With beta13 and the new hook you can do something like the following:

;; This prevents reparsing the buffer if not at least full-balanced concerning
;; parens, brackets and braces.
(add-hook 'semantic-before-toplevel-bovination-hook
  (function (lambda ()
  (condition-case data
  ;; Buffer can't have more than (point-max) sexps.
  (not (scan-sexps (point-min) (point-max)))
(error nil)

Eric told me that the next semantic 14 release (> beta13) is much better and
saver with respect to the auto. parsing feature.
But with beta13 and even the workaround above is works also very well now!

Klaus

Open class for symbol at point

[Berndl, Klaus]

Think of an arbitrary symbol in my java-source, like:


  MyOwnClassType aType = new (...);
  ...
  // somewhere in the code (of course with correct scope)
  aType.aMethod();
^  // means point stays here


Suppopse point stays for example where the ^-sign is.
Is there already a mechanism/function in JDE whic i can
call and then auto. the file is opened where class
"MyOwnClassType "" is defined AND(!) point is located
at definition of "aMethod"?

I kknow jde-show-class-source but i doubt that this function
can do all this especially putting point to the right symbol.

What is the state of the feature "Show definition of symbol at point"?

Thanks,
klaus



-- 
Klaus Berndlmailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

Suggestion for path-expanding

[Berndl, Klaus]

Hi,

in current JDE there are a lot of options where we can define path-lists, e.g.
jde-db-source-directories, jde-global-classpath etc. Some of them uses expansion
of environment-variables some of them not (AFAIK).

Here comes a design-suggestion for the JDE-implementation:

Every(!) option which can store pathes should be able to substitute env-vars.

This can easily be done be something like:

(defmacro jde-expand-path-list (pathlist)
  "Expands all pathes of the list PATHLIST and returns a list with the
expanded pathes."
  (` (mapcar (lambda (path)
   (expand-file-name (substitute-in-file-name path)))
 (, pathlist

With this macro each of these options should have a :set-clause in it's defcustom-
definition, e.g.

(defvar jde-global-class-path-internal nil)
(defcustom jde-global-class-path nil
  "..."
  :group 'jde-project
  :type '(repeat (file :tag "Path"))
  :set (function (lambda (sym val)
 (set sym val)
 (setq jde-global-classpath-internal (jde-expand-path-list 
val)

For each path option jde-XXX in JDE there should be an internal var jde-XXX-internal
which is set by the .set-clause of jde-XXX and contains always the full expanded pathes
Every code in JDE uses always and only jde-XXX-internal.

Advantages: 
- Clear design of all path-options in JDE
- Only one call for expanding (in the :set-clause)

Only drawback: Changes of an env-var does not take effect until the path-option is set
again by the user (either by reloading JDE or hitting "Set for current session" in the 
customize-
buffer).

Any thoughts?

Klaus

-- 
Klaus Berndlmailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

RE: Suggestion for path-expanding

[Berndl, Klaus]

> > Ok, then the only suggestion i still have is to add for every path
> > option jde-XXX an internal jde-XXX-internal variable with is set
> > by the :set-clause of the defcustom-form of jde-XXX. Then not every
> > path-using feature of JDE must call itself the normalizing-function.

>How would this approach handle resolution of relative paths? I could
>not think of a way when I considered it before implementing the
>jde-normalize-path approach.

Hmmm, let me think about it(1,3 seconds later)hmmm, this 
approach can't handle relative paths...
Ok, a man must know when he must give up :-))
Ok, not only a joke: You are right Paul, now i think too, that current
implementation of JDEE is necessary and the :set-clause-approach is
not a real good idea.

Well, forget my suggestion...

Ciao,
klaus

RE: JDE config problem?

[Berndl, Klaus]

Gary Lawrence Murphy writes:
> > 
> > On using jde-gen-class-buffer to create a class that implements
> > java.oi.FileFilter, JDE correctly inserts the needed accept() method
> > but also inserts the following message just after that method:
> > 
> > Invalid function: (macro . # > 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
> > ("/usr/lib/xemacs/xemacs-packages/lisp/semantic/semantic-util.elc"
> > . 14006)>)
> > 
> > What is the likely cause of this?  Outdated Semantic?  Config error?
 
>This is a FAQ. You have incorrectly compiled your files. Remove all
>compiled files for JDE, eieio, and semantic and run uncompiled. Or if
>you insist on running compiled (really makes no difference in
>performance), go back and search the JDE mailing list archive for the
>many discussions of this subject and how to cure it.

At least eieio should be compiled because otherwise the performance can
really be very very bad!

Klaus

- Paul

RE: JDE config problem?

[Berndl, Klaus]

>I've run the JDE with eieio both compiled and uncompiled and I've
>never noticed any difference in performance.  Further, I can't imagine
>why it should cause bad performance.  eieio is not used to do any
>computationally intensive tasks.

First of all not only JDE uses eieio but also semantic which in turn
is used by JDE.

Secondly macros & defsubst functions are inlined if byte-compiled
and eieio has a lot of them (if i remember right, Eric has told
me this).
BTW: This argument would also recommend, that we should
semantic byte-compile!

But here comes the original explanation from Eric, the author of eieio:

"Byte compilation converts the lisp program into a stack based program
with integer IDs for all built-in symbols.  The core of EIEIO uses
almost all built in commands (like car, cdr, assoc, etc).  In
particular, the method dispatch mechanism does a lot of work.  When
EIEIO core is not byte compiled, the method dispatch runs between
every single EIEIO method call.  The byte compilation step speeds
this up a lot."

Therefore my strong recommendation: At least eieio and semantic
should be byte-compiled.
And for other packages it is IMHO at least a good way to detect some
errors...

>If you have noticed bad performance, I'd like to know under what
>circumstances so that I can investigate. There may be ways to
>improve the performance.

Yes, just byte-compile eieio and semantic :-)

- Klaus

Field-completion also viy mounted drives?

[Berndl, Klaus]

I'm wondering if the following is possible:

We have a SUN Solaris system, where our java-sources reside. These sources are
also compiled to *.class files with the SUN-JDK and reside also onto the Solaris 
system.

But we want to use JDE with NTEmacs 21.1. onto Windows NT 4.0 and edit the
java-sources onto the SUN-system via SAMBA mounted drives.

Ok, editing works fine, no problem.
But now i have tried to use the field-completion of the Windows-NT-JDE/Beanshell
with our SUN/Solaris-compiled java-sources.
IMHO i have set CLASSPATH onto the NT-system correct so it should find the
class-files. I have also set jde-db-source-directories to the class-files (is this
necessary? Which variable of JDE must point to the compiles class-files so the
method-completion works?).

But sometimes it finds the class and completes correct, sometimes it don't find
it and asks to attempt to import it (after y it finds it), but sometimes the Beansshell
throws an error in jde.util:
  Expression: 
jde.util.Completion.getClassInfo("de.hvb.ha.adapter.bloomberg.BloombergParams",0);
  Error: // Error: // Uncaught Exception: TargetError : at Line: 30 : in file: 
 : jde .util .Completion .getClassInfo ( 
"de.hvb.ha.adapter.bloomberg.BloombergParams" , 0 ) 
Target exception: java.lang.NoClassDefFoundError: javax/ejb/EJBException
 [2 times]
bsh-eval: Beanshell eval error. See messages buffer for details.

OK: Should it be possible to complete with NT-beansshell/JDE SUN/Solaris sources?
What can be the reason for the beanshell error above?

Thanks in advance!
Klaus


-- 
Klaus Berndlmailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

RE: ecb integration problem

[Berndl, Klaus]

>I recently downloaded ECB version 1.7 and included it in the load path.

Please send a bug-report with the buildin bugreport-generator of ECB
(see menu ECB->Help->Submit-bug report and the related stuff in the
online-help of ECB).

>I want it to use for browsing java source code.
>Its fine and works perfectly at the start.

>But after somtime...electing the source codes doesn't appears in the 
>edit buffer ??

Please describe exactly what you are doing and when the source-code
is gone.

>am I doing something wrong..mean..missingsome configuration..
>or some bug in ECB

Not easy to say because you send us not enough information.

Klaus

RE: jde glitches

[Berndl, Klaus]

Hi Paul,
>...
>You should be sure to byte-compile to minimize the parsing 
>problem.

Some days ago your recommendation was never byte-compile the 
packages because this can cause some problems.
I had written a mail to you and this list, where i strongly
recommended at least byte-compiling eieio and semantic,
exactly because speeds up the parsing.

Now you agree with me and correct your statement you have made
the days before?

What is now your recommendation concerning byte-compiling?
(This is not for me, because i always byte-compile, but i think
some JDE users are interested..)

>You could also try the new disc caching feature of
>semantic so that you don't have to reparse API header files and other
>static files every time yo open them.

Yes, semanticdb should always be turned on!!

Ciao,
Klaus

RE: jde glitches

[Berndl, Klaus]

>Thanks everyone.  I've upgraded to the latest semantic (1.4beta14) and
>speedbar (0.14beta3), hoping that some problem has been fixed.  (I
>coulda
>sworn I tried to do this a few days ago on my home machine, but jde
>complained about the new versions... weird).  As always, I've byte
>compiled all the source.

Switch off jde-check-version-flag!

Klaus

RE: navigation question

[Berndl, Klaus]

>You're right, I realised I didn't read the question fully. But then I'd
>already sent my answer. I seem to remember that the feature you're
>describing has been discussed previously though, and that Paul asked for
>somebody with the time and skills to implement it to go ahead. I'm not
>that
>guy, but I also think it would be really nice.

>> Maybe I am wrong, but it seems to me that C-c C-v C-y doesn't jump to
>> the definition of "whatever" is under the cursor. The cursor needs to
>>be over a class.

I have already sent a first implementation of such a "whatever"-feature to Paul and 
also to this list, which works pretty well. It is completely based on the 
method-completion (mc) feature of JDE because IMHO this mc-feature already has all 
what we need for this and this mc-feature has the most powerful parsing-mechansim 
within JDE (exactly this meachanism is needed for the 
"jump-to-source-whereever-point-stays" feature.

Paul said to me that he want to re-writte my implementation so common code is palced 
in jde-parse.el. I don't know in what stadium this task is.

If you search this list you will find my posting!

Ciao,
klaus

RE: Beanshell + windows setup

[Berndl, Klaus]

Elias Biris writes:
 > PS: By  the way I tried the 2.2.9 current beta to see if this error
>can be alleviated but jde keeps complaining that the speedbar version is
>incompatible (0.14beta4). Trying 0.13 as it suggested causes my ECB
>(Emacs Code Browser) mode to stop working successfuly).

Not JDEE related: But could you please describe how speedbar 0.13 
prevents ECB (which version) to stop working successfully?!
What are the effects?

I'm asking because ECB and Speedbar should not have any dependencies
at least not the official releases of ECB!

Klaus

RE: auto-fill-mode problems while typing javadoc

[Berndl, Klaus]

I'm using filladapt.el which is a much smarter (auto-)filling then the
standard-mechanism of Emacs.

here an example. Suppose the following method:

  public void print(int i, double d)
  {
 return;
  }

point is on the first line. Then i do JDE->Documentation->Add and get

  /**
   * Beschreibung der Methode print :
   *
   * @param i Wert vom Typ int - 
   * @param d Wert vom Typ double - 
   */
  public void print(int i, double d)
  {
 return;
  }

Then i fill out the several parts of this comment like follows:

/**
 * Beschreibung der Methode print : hfdghsd fjghksjdf hg
 * hdfjgh sdjfhg jkhsdfkjg hsdkjf ghjksdhfg jkhsdjfg hksjdfhg jksdfhg
 * jhdfjk hsdkfjg hsdkjfg hjsdhfg jkshdfjkgh sdjfgh
 * + sjdj sdhfjk hasdfjhas kdfjhaskdfh ajksdhf jhasdkfj hasdj fh ajksdhf
 *   jashdfkj hsdjf hasjdhf
 *   - sdjkfh sjkdhf jkashdjf hasjkdfh jashdfkj ahsdjfh aksjdf ahsdkjf
 * haksjdhf jashdf
 * + sdjkfh asjkdfh jashdfjk hasjkdfh jkashdfj hasjdfh jasdhf jkahsdjkf
 *   hajksdhf jhasdfjk hasjdfh
 *
 * @param i Wert vom Typ int - jsdfhgj sdhfgjhsdfjg
 *hjskdfhg jsdhfg jhsdfjgh sdkjfhg jsdfhg jhsdfjghjksdfhg
 *jkhsdfjg hdhfg
 * @param d Wert vom Typ double - djfhgjk sdhfgjk hsdfkg
 *jsdhfgjkshdfkghsdjfhgjdhfg jshdfg hsdfjghsdkfg
 *+ dfhg sjdfhg jksdhfgjk hsdfjkgh sdjkfg hskjdfhg jsdfhgk
 *  jsdfhg jsdhfg jkh.
 */
public void print(int i, double d)
{
   return;
}

Note: I have not made any manual line-breaks except before beginning a new bullet (+). 
Have not inserted any manual * or indendation, but all is
doen automatically by filladapt.el. Note also the indendation of the
javadoc-keywords "param" (works for all other keywords too).

Klaus 

-Original Message-
From: Laurent Mirguet
To: [EMAIL PROTECTED]
Sent: 05.06.02 15:51
Subject: auto-fill-mode problems while typing javadoc

Hello,

I use (auto-fill-mode) as a minor mode for JDE.

When I type the javadoc description for a class, and when this mode 
provoques a carriage-return, I get a * at the beginning of the new line 
which is correctly indented. This is quite nice.

But when I type the javadoc description for a method, I don't get the 
same behaviour : the * doesn't appear automaticaly.

Is there something that I can do to fix that ? Is this the normal 
behaviour ?

I use :
Linux Redhat 7.3
emacs 21.2.1
jde 2.2.9beta9.1
ecb-1.70
eieio-0.17beta4
elib-1.0
semantic-1.4beta14
speedbar-0.13a

Thank you
Laurent Mirguet

RE: Limit posts to members?

[Berndl, Klaus]

It's ok for me or better after reading all postings to this topic i would
say: This is necessary because otherwise this really helpful list is quite
unuseable

Klaus

-Original Message-
From: Paul Kinnucan [mailto:[EMAIL PROTECTED]]
Sent: Thursday, June 27, 2002 7:48 AM
To: [EMAIL PROTECTED]
Subject: Limit posts to members?


Hi all,

Do you have any objections to limiting postings
to this list to subscribers? It may help
to reduce the spam. (It won't stop spammers
disguised as legitimate members.)

- Paul

RE: Javadoc comments

[Berndl, Klaus]

Hi!


Smart ENTER in all cc-mode based major-modes (so also for jde-mode):


Here is what i have bound to RET in my c-mode-common-hook
(jde-mode inherits all this!):

(defun my-cc-mode-return ()
  "Intelligent line breaking in all cc-modes. Handles strings in a smart
way"
  (interactive)
  (cond ((eq (c-in-literal) 'string)
 (if (or (eq major-mode 'java-mode) (eq major-mode 'jde-mode))
 (insert-and-inherit "\"+")
   (insert-and-inherit "\""))
 (c-context-line-break)
 (insert "\" ") ;; cause of strings in c will be concatenated we
;; must add a single space.
 (end-of-line))
(t (c-context-line-break

With this function you will get best and smartest results after hitting
ENTER regardless if in code, if within a string or within a comment.


Some more maybe usefull settings:
=

Here are some of my other cc-mode common settings which should be also
useful for java (all this code is added to c-mode-common-hook!):

  ;; comments should be treated as whitespace in paren-matching, jumping
  ;; etc.
  (setq parse-sexp-ignore-comments t)
  ;; we like auto-newline and hungry-delete
  (c-toggle-auto-hungry-state 1)
  ;; keybindings for C, C++, and Objective-C.  We can put these in
  ;; c-mode-base-map because c++-mode-map and java-mode-map  and
  ;; c-mode-map inherit it.
  (define-key c-mode-base-map "\C-m" 'my-cc-mode-return)
  ;; lines of a c-block-comment should begin with a ""
  (set-variable 'c-block-comment-prefix "")
  ;; we want no autofilling within cpp-makros string-literals and normal code
  (set-variable 'c-ignore-auto-fill '(string cpp code))
  ;; turn on adaptive auto-fill-mode for max. easy writing comments
  (auto-fill-mode 1)
  (c-setup-filladapt)
  (turn-on-filladapt-mode)
  ;; no automatic newlines after ';' if following line is non blank or
  ;; inside oneline inline methods
  (add-to-list 'c-hanging-semi&comma-criteria
   'c-semi&comma-no-newlines-before-nonblanks)
  (add-to-list 'c-hanging-semi&comma-criteria
   'c-semi&comma-no-newlines-for-oneline-inliners)


Maybe someone find these settings also useful.


Block closing comments:
===

Ok, and here comes codes which add automatically after every closing
} auto. after hitting } a block-closing comment so you see for large
blocks which statement is closed by }:

To get the following code below working you have to add

  (block-close . my-c-snug-do-while)

to c-hanging-braces-alist, so e.g. a sensefull setting could be:

(c-hanging-braces-alist . ((substatement-open . my-set-begin-block)
   (brace-list-open)
   (brace-list-close)
   (brace-list-intro)
   (brace-entry-open)
   (block-close . my-c-snug-do-while)))


Here comes the code:

;; 
;; comment block-close

;; The communication-channel between `my-c-snug-do-while' and
;; `comment-block-close': `block-close-data' is set during the call of the
;; ACTION-function `my-c-snug-do-while' and evaluated within the
;; `c-special-indent-hook' (here `comment-block-close', s.b.)
;; If not nil it is a list of the following elements:
;; 1. The text of the comment to be inserted behind the closing brace of
;;the block.
;; 2. The exact position at which the closing brace was inserted
;; 3. The size of the block in lines
(defvar block-close-data nil)

;; define here how many lines a block at least must contain so the
;; block-close-comment will be added.
(defcustom c-block-close-min-size-to-comment 15
  "*Define how many lines a block at least must contain so the
block-close-comment will be added."
  :type 'integer
  :group 'c)

(defun comment-block-close ()
"For all blocks unless do-while-blocks this function adds after the
block-close-brace the block-begin-statement as comment, for example:
for \(i = 1; i < 6; i++\)
\{
cout << i;
...
\} // for \(i = 1; i < 6; i++\)
Already hitting the '\}' inserts the comment! This works only if the
autonewline-feature of cc-mode is turned on!"
  (if (and block-close-data
   (> (nth 2 block-close-data) c-block-close-min-size-to-comment))
  ;; Because this function is hooked to c-special-indent-hook and
  ;; therefore called after EVERY indentation, we do the
  ;; comment-insert only if we have a block and it´s not a do-while and
  ;; block is larger the 'block-close-min-size-to-comment' lines
  (save-excursion
;; frirst we must do some stuff to go to the right position for the
;; block-close-comment:
(goto-char (nth 1 block-close-data))
(if (not (looking-at "[ \t]*}"))
(next-line 1))
(end-of-line)
(forward-sexp -1)
(forward-se

RE: cygwin-emacs and classpath under jdee

[Berndl, Klaus]

Do not know exactly, but JDE checks sometimes the following condition:

(if (eq system-type 'cygwin32)...)

and here is:

(defvar jde-classpath-separator (if (eq system-type 'cygwin32) 
";" path-separator)
  "The separator to use in a classpath.
This is usually the same as `path-separator'")

So, normally it should work i would say. Check the value of
system-type in your cygwin Emacs 21.2. Maybe it's different!?

Ciao,
Klaus

-Original Message-
From: Artur Hefczyc
To: jdee
Sent: 12.09.02 10:47
Subject: cygwin-emacs and classpath under jdee

Hello,

Recently I installed on my WinXP box last cygwin version
with emacs-21.2 with X toolkit.

Jdee translates classpath almost well except that it puts ':'
between directories instead of ';'.

Earlier when I compiled emacs on my own under this environment
it worked well, I could even use both unix-like and windows-like
path style when loeaded files (i mean i could use: "d:/file.java" and
"/cygdrive/d/file.java") and in paths I always got ';' as needed,
however it looks like cygwin-emacs version only sees cygwin 
path and path style but it doesn't know about windows anything.
And probably jdee rely on emacs knowledge about path style.

So my questions is:
does anybody know how to force jdee to put ';' as path/file separator
instead of ':'.
I want classpath like this: "d:/jar1.jar;d:/jar2.jar;d:/jar3.jar"
instead of: "d:/jar1.jar:d:/jar2.jar:d:/jar3.jar" what i currently get.

regards
Artur Hefczyc
-- 
Artur Hefczyc[EMAIL PROTECTED]
Open Source Developer
http://wttools.sourceforge.net/

JDE supporting new cygwin GNU Emasc 21.2

[Berndl, Klaus]

Hi Paul,

to make JDE supporting well the new cygwin port of GNU Emasc 21.2. you should add to 
every code which checks currently for (eq system-type 'cygwin32) the symbol 'cygwin 
because XEmacs use cygwin 32 whereas GNU Emacs use cygwin. So for example you could 
replace the old code with

(member system-type '(cygwin32 cygwin))

Ciao,
klaus


-- 
Klaus Berndlmailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

Small ugly code in jde.el

[Berndl, Klaus]

Hi Paul,

Could you please replase this "hard" setq in jde.el with add-to-list.

;;;###autoload
(setq auto-mode-alist
  (append
   '(("\\.java\\'" . jde-mode))
auto-mode-alist))

The reason is that i want to auto-load my latest JDEE-installation with

(add-to-list 'auto-mode-alist '("\\.java\\'" . jde-mode))
(autoload 'jde-mode "jde")

In my .emacs-file.

With this JDEE is first loaded after opening a java-file. BUT:
Loading JDEE then adds jde-mode again to the auto-mode-alist,
cause of this setq instead of add-to-list which would check if
it is already contained in auto-mode-alist!

Not really a problem but ugly...

Ciao,
Klaus

AW: cedet.sourceforge.net refusing connections

[Berndl, Klaus]

>>> John Collins <[EMAIL PROTECTED]> seems to think that:
>>I'm trying to set up JDEE on a new machine, and I can't get at the 
>>emacs
>>tools eieio, semantic, etc. Any news?
  [ ... ]

>Every once in a while the sourceforge download system fails.  I don't know why, but 
>it >
>usually comes back in a little while.

I tried it and it seems that it is back ;-)

Klaus

Eric

-- 
  Eric Ludlam: [EMAIL PROTECTED], [EMAIL PROTECTED]
   Home: http://www.ludlam.netSiege: www.siege-engine.com
Emacs: http://cedet.sourceforge.net   GNU: www.gnu.org

AW: JDEbug error...

[Berndl, Klaus]

>I want to use JDEBug and tried to set it up but without success. The operating system 
>is >Linux-debian with IBMJava2-13.

>This is my setup:
>1) set jde-db-debugger as debugger
>2) jde-bug-vm-includes-jpda-p  On
>   jde-bug-jdk-directory   /usr/local/share/java/IBMJava2-13/
>3) jde-db-source-directories   /usr/local/share/java/IBMJava2-13/ 
>   /home/development/test/src

>So far all ok, but when I now lauch the process: Processes->Launch Process emacs 
>responds >with:
>  jde-dbs-cmd: Process *JDEbug* not running
>  No response to command 1. (process = 1; timeout = 30 sec.)

AFAIK the option jde-db-debugger has switched meaning; now you have
to use jde-debugger.
In general your option settings above seem to be pretty out-dated...

IMHO now it's time to give JDEE an autom. Option upgrading tool which autom. Upgrade 
renamed, type-changed options to valid values for current release. The user gets a 
display where all upgraded options are listed with old and new values!
For example ECB (Emacs Code Browser) has already such a tool and i have sent Paul the 
library which is pretty independent and can be used for every big Emacs package.

With such a big package like JDEE two often two many option change either name, 
semantic or even both! The package should (and could!) handle this itself.

Ciao,
Klaus

RE: JDEbug error...

[Berndl, Klaus]

>I want to use JDEBug and tried to set it up but without success. The 
>operating system is >Linux-debian with IBMJava2-13.

>This is my setup:
>1) set jde-db-debugger as debugger
>2) jde-bug-vm-includes-jpda-p  On
>   jde-bug-jdk-directory   /usr/local/share/java/IBMJava2-13/
>3) jde-db-source-directories   /usr/local/share/java/IBMJava2-13/ 
>   /home/development/test/src

>So far all ok, but when I now lauch the process: Processes->Launch 
>Process emacs responds >with:
>  jde-dbs-cmd: Process *JDEbug* not running
>  No response to command 1. (process = 1; timeout = 30 sec.)

AFAIK the option jde-db-debugger has switched meaning; now you have to use 
jde-debugger. In general your option settings above seem to be pretty out-dated...

IMHO now it's time to give JDEE an autom. Option upgrading tool which autom. Upgrade 
renamed, type-changed options to valid values for current release. The user gets a 
display where all upgraded options are listed with old and new values! For example ECB 
(Emacs Code Browser) has already such a tool and i have sent Paul the library which is 
pretty independent and can be used for every big Emacs package.

With such a big package like JDEE two often two many option change either name, 
semantic or even both! The package should (and could!) handle this itself.

Ciao,
Klaus

AW: Semantic 1.4.1 is released!

[Berndl, Klaus]

Hello Eric and David!

I just downloaded semantic 1.4.1 and had noticed the following:

1. Bug in semantic-texi.el: It is not enough to add ("defopt" . 1) to
   semantic-texi-name-field-list but is also necessary to add opt to
   semantic-texi-super-regex:
   
(defvar semantic-texi-super-regex
  "^@\\(chapter\\|\\(sub\\)*section\\|unnumbered\\(\\(sub\\)*sec\\)?\\|\
\\(chap\\|\\(sub\\)+\\|major\\)?heading\\|appendix\\(\\(sub\\)*sec\\)?\\|\
centerchap\\|def\\(var\\|un\\|fn\\|opt\\)x?\\)"
  "Regular expression used to find special sections in a Texinfo file.")

2. Problems with shipped semantic-util-modes.el: With 1.4.1 there is after
   every autom. (partial) reparse a recenter of the source-buffer which is
   pretty annoying. I have replaced the version of 1.4.1 with that one shipped
   with 1.4 and all is working fine, i.e. no recentering of the source-buffer.
   
   I'm wondering why the version of semantic-util-modes.el in 1.4.1 is
   X-RCS: $Id: semantic-util-modes.el,v 1.17 2001/12/08 03:12:39
   whereas the version of 1.4 seems to be much newer:
   X-RCS: $Id: semantic-util-modes.el,v 1.17.2.2 2002/05/07 02:40:50

   There are not many differences between these versions of
   semantic-util-modes.el but nevertheless they seem to be the reason for the
   annoying recenter behavior of the source-buffer after autom. reparsing.

Thoughts?

Ciao,
Klaus



-Ursprüngliche Nachricht-
Von: David PONCE [mailto:[EMAIL PROTECTED]] 
Gesendet: Mittwoch, 11. Dezember 2002 10:16
An: [EMAIL PROTECTED]; [EMAIL PROTECTED]; 
cedet-devel; [EMAIL PROTECTED]
Betreff: ANN: Semantic 1.4.1 is released!


 Semantic 1.4.1 is released!

* What's new

  This is a bug fix release only, without new features added.

* Requirements

  Semantic works with Emacs 20.x, 21, XEmacs 20.x, and 21.x.

  Semantic's speedbar interface requires the latest version of
  speedbar (1.4beta4).  The Semantic Database, Semantic Analysis, and
  semantic charts tools require the latest version of EIEIO (0.17).

* Availability

  You can download Semantic, speedbar and EIEIO from the CEDET project
  home page at: .

  The home page of Semantic is at:
  .

* Patches, bug reports & feature requests

  Please use the various "Tracking systems" provided on the CEDET
  project page at: .

* Mailing list

  The cedet-semantic mailing list is available at:
  .

* Changes since 1.4

  2002-11-27 09:53  ponced

 * semantic.el:
 
 (semantic-version): Updated to 1.4.1.

  2002-11-27 09:52  ponced

 * Project.ede:
 
 (ede-proj-project "semantic"): Updated :version to 1.4.1.

  2002-11-27 09:52  ponced

 * NEWS:
 
 Added entry for Semantic 1.4.1.

  2002-11-21 10:58  ponced

 * semantic-c.el:
 
 (semantic-toplevel-c-bovine-table): New c.bnf.

  2002-11-21 10:56  ponced

 * semantic-texi.el:
 
 (semantic-texi-name-field-list): Add `defopt' as a parsable field
 entry.

  2002-11-21 10:55  ponced

 * c.bnf:

 (opt-class-declmods): New.
 (typeformbase): Support templates after :: operator.
 (variablearg): Support *'s in nameless arguments.

  2002-11-18 11:39  ponced

 * semantic-util.el:

 (semantic-change-function-mark-dirty): Fixed bug #513650, "Stack
 overflow in equal" in `add-to-list' when comparing complex tokens.

  2002-05-15 21:21  ponced

 * semantic-java.el:
 
 (semantic-default-java-setup): Re-generated from updated grammar.

  2002-05-15 21:21  ponced

 * java.bnf:

 (setup): take into account new
 `semantic-symbol->name-assoc-list-for-type-parts' option.

  2002-05-07 02:40  zappo

 * semantic-util-modes.el:
 
 Fixed bug #619758 "semantic-map-buffers void, only ver. 1.4"

ECB 1.90 is released!

[Berndl, Klaus]

ECB 1.90 is released!
=

>From this release on ECB has a new official maintainer: Klaus Berndl
<[EMAIL PROTECTED]>.

Also the homepage of ECB has moved to http://ecb.sourceforge.net.
The old homepage at http://home.swipnet.se/mayhem/ecb.html is not longer
supported.



How to get it:
--

If you are using ECB 1.80 then you can just call "M-x ecb-download-ecb" if you
are online. ECB will then download latest ECB (1.90) and install it for you.

Or go to the new homepage at http://ecb.sourceforge.net and download it from
there.



What is new in 1.90:


ECB 1.90 has a lot of new features and fixed some annoying bugs. Here is only
an excerpt from the HISTORY file (which contains the complete list of
enhancements). Here are some of the news in ECB 1.90:

* Fixed an annoying bug which results in an error "Wrong type argument:
  integer-or-marker-p nil)" after a full buffer reparse. ECB 1.80 has repaired
  its internal state but nevertheless the user had to reclick on the same
  token after this error to really jump to a token. With ECB 1.90 this bug has
  been gone.

* Now ECB displays at every start a "Tip of the day" in a dialog-box. This can
  be switched off with option `ecb-tip-of-the-day'.
  
* New feature: Now the methods buffer is auto. expanded if the node related to
  the current token in the edit-window is not visible (probably because its
  parent is collapsed).
  
* New command `ecb-expand-methods-nodes' which allows precisely expanding
  tokens with a certain indentation-level.

* Rewritten the mechanism for storing customized window-sizes. See option
  ecb-layout-window-sizes and the command ecb-store-window-sizes and
  ecb-restore-window-sizes. Now the sizes are always saved as fractions of the
  width (resp. height) of the ECB-frame, so the stored sizes are always
  correct regardless of the current frame-size.

* Rewritten cache-mechanism for directories and sources:
  This results in a speed-boost for big-size directories.

* Added a complete new section "The layout-engine of ECB" to the online-help
  which describes in detail how to program new layouts and new special
  windows.

* Fixed some bugs concerning the eshell-integration.

* Speedbar is now integrated in ECB and can be used instead of the standard
  ECB-directories buffer.

* Naming and managing of layouts has been changed! Now a layout is not longer
  identified by an integer but by an arbitrary string! Example: The layout
  with index 0 in ECB <= 1.80 is now named "left1" in ECB 1.90.

  Therefore the name of the option 'ecb-layout-nr' has changed to
  'ecb-layout-name'! See the docstring of 'ecb-layout-name' for the names of
  all buildin layouts. ECB autom. upgrades the old-option to the new one!

* A lot of new hooks

* Adding a new layout type "left-right" which allows the ECB-tree-windows to
  be located at the left and the right side of the ECB-frame.

* Now ECB can be "autoloaded".

* Now ECB offers a command `ecb-create-new-layout' for interactively creating
  new layouts "by example".


/Klaus Berndl

RE: Introspection based Java class browser....also demonstrates alittle bit of java and lisp integration...

[Berndl, Klaus]

Hello,

attached is a small library ecb-jde.el which enables you to display
the contents of class at point in the method-window of ECB:

(defun ecb-jde-display-class-at-point ()
  "Displays in the ECB-methods-buffer the contents \(methods, attributes
etc...) of the class which contains the definition of the \"thing\" under
point \(this can be a variablename, classname, methodname, attributename).
This function needs the same requirements to work as the method-completion
feature of JDE \(see `jde-complete-at-point')!. The source-file is searched
first in `jde-sourcepath', then in `jde-global-classpath', then in $CLASSPATH,
then in current-directory."
  (interactive)
  ...
)

Just do:

1. Save ecb-jde.el somewhere in your load-path - probably best in the
   ecb-directory ;-)
2. Activate ECB
3. Open a java-source 
4. Load ecb-jde.el, e.g. with M-x load-library RET ecb RET
5. Try it.

Differences to Sandips code:

- uses ECB method-window to display the class-contents
- uses semantic for getting the contents of a class
- needs the source-file of the class to work

The first point can be seen as advantage or disadvantage ;-)

Especially the latter point is a disadvantage because it prevents
from working for classes which are only available as *.class-file or
within a jar-file...

But it demonstrates another elegant way of displaying things of a class
under point without needing java-introspection.

Klaus

-Original Message-
From: Sandip Chitale [mailto:[EMAIL PROTECTED]] 
Sent: Wednesday, February 19, 2003 7:32 AM
To: [EMAIL PROTECTED]
Subject: Introspection based Java class browseralso demonstrates alittle bit of 
java and lisp integration...


>To use :
 
>1. unzip the attached zip at in the jde directory
>2. Just put the following in your .emacs

>(require 'jde-introspect)

>Now put the point anywhere in Java buffer where you would
>normally invoke jde-complete functions. Then type
>(control c) (control v) (/) to see the java typeinfo
>buffer.
 
>For example:
 
>With point in the 'System' below :
>public class Foo {
> public static void main(String[] args) {
>  System.out.println(args.length);
> }
>}
>here is what you get in a temp buffer. All the java class names are
>hyperlinks (activated by
>mouse or  key).
>class java.lang.System




ecb-jde.el
Description: Binary data

RE: auto-delete compile buffer

[Berndl, Klaus]

>I created this function:

I think it has to be:

(defun jde-compile-finish-kill-buffer (buf msg)
  "Removes the jde-compile window after a few seconds if no errors."
  (save-excursion
(set-buffer buf)
(if (null (or (string-match ".*exited abnormally.*" msg)
  (string-match ".*BUILD FAILED.*" (buffer-string
;;no errors, make the compilation window go away in a few seconds
(progn
  (run-at-time
   "2 sec" nil 'delete-windows-on
   (get-buffer-create "*compilation*"))
  (run-at-time
   "2 sec" nil 'kill-buffer
   (get-buffer-create "*compilation*"))
  (message "No compilation errors"))
  ;;there were errors, so jump to the first error
  (next-error

Klaus

ECB 1.92 released!

[Berndl, Klaus]

ECB 1.92 is released!
=
 
Reminder:
 
The homepage of ECB has moved to http://ecb.sourceforge.net 
 .  The old
homepage at http://home.swipnet.se/mayhem/ecb.html 
  is not longer supported.
 

How to get it:
--
 
Sorry, it's too big to post the sources here...
 
If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if you
are online. ECB will then download latest ECB (1.92) and install it for you.
 
Or go to the homepage at http://ecb.sourceforge.net   and 
download it from
there.
 

What's new in 1.92:
---
 
This release fixes some small bugs and introduces some new features for
layout-handling. Here are the entries of the NEWS (was HISTORY) file:
 
** Fixed bugs:
 
*** Fixed small bugs in the popup-menu of the history-buffer.
 
*** Fixed some not working links in the FAQ.
 
** ECB now includes a file "ecb-autoloads.el" which contains all available
   autoloads of ECB. Therefore loading this file is enough.
 
** Enhancements to the layout
 
*** The sizes of the ecb-windows can be fixed even during frame-resizing
This new feature is only available with GNU Emacs 21.X. See new option
`ecb-fix-window-size'.
Suggested by Gunther Ohrner <[EMAIL PROTECTED]  >
 
*** The command `ecb-store-window-sizes' can now also store fixed sizes.
If called with a prefix arg then fixed sizes are stored instead of
fractions of frame-width resp. -height.
 
*** The command `split-window' is now also adviced to work properly.
 
** Enhancements for the tree-buffers
 
*** "Create Source" in the popup-menues now work also for non-java-files.
 
*** ecb-truncate-lines can now be set different for each ECB-tree-buffer.
 
*** Easy horizontal mouse-scrolling of the ECB tree-buffers. See new option
`ecb-tree-easy-hor-scroll'.
 
** Added a command `ecb-jde-display-class-at-point' which displays the contents
   of the java-class under point in the methods-buffer of ECB.
 
** Renamed previous HISTORY file to NEWS and reformate it for outline-mode.
 

Here are some of the new features of ECB beginning with version 1.90:
 
* Fixed an annoying bug which results in an error "Wrong type argument:
  integer-or-marker-p nil)" after a full buffer reparse. ECB 1.80 has repaired
  its internal state but nevertheless the user had to reclick on the same
  token after this error to really jump to a token. With ECB 1.90 this bug has
  been gone.
 
* Now ECB displays at every start a "Tip of the day" in a dialog-box. This can
  be switched off with option `ecb-tip-of-the-day'.
  
* New feature: Now the methods buffer is auto. expanded if the node related to
  the current token in the edit-window is not visible (probably because its
  parent is collapsed).
  
* New command `ecb-expand-methods-nodes' which allows precisely expanding
  tokens with a certain indentation-level.
 
* Rewritten the mechanism for storing customized window-sizes. See option
  ecb-layout-window-sizes and the command ecb-store-window-sizes and
  ecb-restore-window-sizes. Now the sizes are always saved as fractions of the
  width (resp. height) of the ECB-frame, so the stored sizes are always
  correct regardless of the current frame-size.
 
* Rewritten cache-mechanism for directories and sources:
  This results in a speed-boost for big-size directories.
 
* Added a complete new section "The layout-engine of ECB" to the online-help
  which describes in detail how to program new layouts and new special windows.
 
* Fixed some bugs concerning the eshell-integration.
 
* Speedbar is now integrated in ECB and can be used instead of the standard
  ECB-directories buffer.
 
* Naming and managing of layouts has been changed! Now a layout is not longer
  identified by an integer but by an arbitrary string! Example: The layout
  with index 0 in ECB <= 1.80 is now named "left1" in ECB 1.90.
 
  Therefore the name of the option 'ecb-layout-nr' has changed to
  'ecb-layout-name'! See the docstring of 'ecb-layout-name' for the names of
  all buildin layouts. ECB autom. upgrades the old-option to the new one!
 
* A lot of new hooks
 
* Adding a new layout type "left-right" which allows the ECB-tree-windows to
  be located at the left and the right side of the ECB-frame.
 
* Now ECB can be "autoloaded".
 
* Now ECB offers a command `ecb-create-new-layout' for interactively creating
  new layouts "by example".
 

Enjoy,
Klaus
 
Klaus Berndl mailto:   [EMAIL PROTECTED]
sd&m AG   http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

 

RE: ECB + JDE window focus bug.

[Berndl, Klaus]

>When I use 'jde-import-find-and-import' in ECB with a compile window, the
>prompt window that asks for clarification pops up in the compilation buffer. 
>This is fine.  But then, it doesn't have an idea where to put the import
>statement.  Is there a setting in JDE/ECB that I can tweak to correct this?

Hmm, not easy i think. After taking a short look into the efc.el library of
JDEE i saw that thsi library is not very flexible, i.e. it calls for example
`delete-window' after calling efc-dialog-ok or efc-dialog-cancel.

JDEE expects not running in another window-layout than the typical Emacs-layout
where things like completion, help, compile etc...opens a new window at the bottom,
display their stuff and are intended to be delete after use. efc.el seems to expect
this too.

IMHO not very flexible but on the other side i must admit i currently do not
know any other and better and more flexible solution which allows correct working
with every window-layout and therefore also with layouts which contain a durable
window where all stuff like compile-output, help and also efc-dialogs are displayed.
Such a window must not be deleted after efc-dialog-ok or efc-dialog-cancel.

But how to achieve this goal??

If you use ECB without a compile-window all should work well but this is of course
not the optimal solution.

ECB advices currently already delete-window and other window-function so they work
correct with ECB. Maybe ECB should try to be smart here: If delete-window is called
for the compile-window then it does nothing, means it does not delete the window but
behaves to the caller of delete-window as if the window has been deleted

I must try thismaybe this could be a way to make ECB and JDEE working better in
combination


Klaus

>Recreate Problem:

>1. emacs -q

>2. setup loadpaths

>3. load jde, ecb

>3. find asdf.java

>4. customize ecb-layout to use a compilation window.

>5. ecb-activate.

>6. With "adsf.java" in edit window, and focus in same window, do
>`jde-import-find-and-import'

>7. Import "ParseException" class, choose any of the choices, click "ok"

>8. Now the import statement is in my scratch buffer.

RE: ECB + JDE window focus bug - suggestion for solution (was: RE: EC B + JDE window focus bug)

[Berndl, Klaus]

Hi Paul,

yes, indeed, this sounds smarter then the hook-solution! But nevertheless IMHO
a package can not have enough hooks where other packages can "plugin" - IMHO
there is no disadvantage for efc.el offering such hooks.

But back to your suggestions: Yes, a variable efc-dialog-window which can be set
by other packages would be great. In this case efc must not delete this window but
has to restore the original buffer (like you describe).

A word to your other suggestion about using extra frames: Maybe we should go
one step more forward and generalizing efc.el so it is independed from using
buffer, windows, frames at all. Maybe efc.el should offer a well-defined interface
for

a) displaying all the stuff and an implementation has only to satisfy this interface
   but can display all the stuff in any arbitrary suitable manner - for exmaple XEmacs
   has a really powerful dialog-implementation (make-dialog) so for example "real" 
   windows-dialogs could be used to display the efc-contents.
b) hitting ok
c) hitting cancel
d) closing the display-"medium" (buffer, window, frame, OS-dialog, popup-menu etc...)

efc.el itself could then offer standard-implementations of these interfaces which uses
buffer/windows and buffers/frames for display.

Hmm, maybe this is not easy but would be great...

But for the beginning your suggested ecb-dialog-window would be great (maybe you could
then also add some hooks - is not much effort :-))

Thanks,
Klaus

-Original Message-
From: Paul Kinnucan [mailto:[EMAIL PROTECTED] 
Sent: Thursday, February 27, 2003 7:11 PM
To: Berndl, Klaus
Cc: 'Le Wang'; [EMAIL PROTECTED]; 'Paul Kinnucan'
Subject: ECB + JDE window focus bug - suggestion for solution (was: RE: EC B + JDE 
window focus bug)


Hi Klaus,

Wouldn't a simpler solution be for efc to define
either a variable whose value is a window for
displaying the dialog or a function that returns
a window for displaying the dialog. For example,

efc-dialog-window

If efc-dialog-window is nonnil, efc would display the
dialog buffer in the specified window and then
restore the buffer that it was originally displaying.
Otherwise, efc would create a temporary window and
delete the temporary window after the user dismisses 
the dialog buffer.

Fixed layout tools like ECB could set this variable
to the window in which they would like the dialog
to be displayed.

Another or additional solution would be for the
efc to define the following variable:

efc-frame-dialog-p

If this variable is nonnil, efc would display the
dialog buffer in a separate frame centered on the
user's screen. If this variable is nil, it would
display the dialog buffer in the window specified
by efc-dialog-window.

Your thoughts?

- Paul

 Berndl, Klaus writes:
> 
 > But i would vote for the following enhancement of efc.el:
 > 
 > Offering 4 new hooks:
 > A) ecb-show-dialog-before-hook
 > B) ecb-show-diaolg-after-hook
 > C) ecb-close-dialog-before-hook
 > D) ecb-close-dialog-after-hook
 > 
 > What would be the advantages of these hooks:
 > 
 > A tool like ECB (which has such durable windows whcih prevent the efc-dialogs
 > to work correct) could temporary delete its durable compile/etc/stuff-window
 > temporary in the new hook ecb-show-dialog-before-hook so the efc-mechanismus
 > described at beginning (point 1, 2) can work correct. And could also recreate
 > its durable window after the efc-dialog has closed in the hook
 > ecb-close-dialog-after-hook.
 > 
 > You see really necessary are only hooks a) and D) but B) and C) could also be
 > senseful (maybe)...
 > 
 > Thoughts? IMHO it would not be a big deal for JDEE/efc but would offer big
 > advantages for tools like ECB to work better with JDEE under all circumstances.
 > 
 > In general such hooks would be good on every place JDEE creates and deletes windows.
 > 
 > Please give me feedback!
 > 
 > Ciao,
 > klaus
 > 
 > 
 > 
 > -Original Message-
 > From: Le Wang [mailto:[EMAIL PROTECTED] 
 > Sent: Monday, February 24, 2003 5:03 PM
 > To: [EMAIL PROTECTED]
 > Subject: ECB + JDE window focus bug.
 > 
 > 
 > Hi,
 > 
 > When I use 'jde-import-find-and-import' in ECB with a compile window, the
 > prompt window that asks for clarification pops up in the compilation buffer. 
 > This is fine.  But then, it doesn't have an idea where to put the import
 > statement.  Is there a setting in JDE/ECB that I can tweak to correct this?
 > 
 > Recreate Problem:
 > 
 > 1. emacs -q
 > 
 > 2. setup loadpaths
 > 
 > 3. load jde, ecb
 > 
 > 3. find asdf.java
 > 
 > 4. customize ecb-layout to use a compilation window.
 > 
 > 5. ecb-activate.
 > 
 > 6. With "adsf.java" in edit window, and focus in same window, do
 > `jde-import-find-and-import'

ECB 1.95 released!

[Berndl, Klaus]

  ECB 1.95 is released!
 
 
 

What's new in the new release:
--
 
* Changes for ECB version 1.95
 
** ECB now displays the expand- and collapse symbols in the tree-buffers with
   image-icons - the same icons as speedbar uses. See the new option
   `ecb-tree-use-image-icons'.
 
** Adding hideshow to the popup-menu of the Methods-buffer.
   This popup-menu now offers two entries for hiding resp. showing the block
   of that token in the Methods-buffer for which the popup-menu was openend.
   This is done with the hideshow.el library.
   Thanks to Christoff Pale <  [EMAIL PROTECTED]> for 
suggestion.
 
** Horizontal scrolling of the tree-buffers by clicking the edges of the
   modeline with mouse-1 or mouse-2. I.e. if you click with mouse-1 onto the
   left (resp right) egde of the modeline you will scroll left (resp. right)
   with the scroll-step defined in `ecb-tree-easy-hor-scroll'. This is only
   for GNU Emacs because XEmacs has hor. scrollbars.
   
** Changed default keybindings:
   - C-c . r:  `ecb-rebuild-methods-buffer'
   - C-c . lc: `ecb-change-layout'
   - C-c . lr: `ecb-redraw-layout'
   - C-c . lt: `ecb-toggle-layout'
   - C-c . lw: `ecb-toggle-ecb-windows'
 
** Starting ECB is now possible via the "Tools"-menu
 
** ECB now requires speedbar version 0.14beta1 or higher. An automatic
   requirements-check is done by ECB. This is because now ECB has three
   dependencies to speedbar-code: Integrating whole speedbar in the ECB-frame,
   using speedbar-logic to parse files with imenu or etags and using the icons
   of speedbar for the ECB-tree-buffers.
 
** Fixed bugs
 
*** Now the "goto-window-..." menu-entries in the "ECB"-menu are working correct
if a speedbar is integrated into the layout.
 
*** Preventing etags-supported sources from being parsed (and saved) too often.
 
 
 
How to get it:
--
 
Sorry, it's too big to post the sources here...
 
If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if
you are online. ECB will then download autom. latest and newest ECB and
install it for you.
 
Or go to the homepage at   http://ecb.sourceforge.net and 
download it from
there.
 
 
 
General description of ECB:
---
 
ECB stands for "Emacs Code Browser" and is a source code browser for (X)Emacs.
It is a global minor-mode which displays a couple of windows that can be used
to browse directories, files and file-contents like methods and variables. It
supports source-code parsing for semantic-supported languages like Java, C,
C++, Elisp and Scheme as well as for source-types supported "only" by imenu or
etags (e.g. perl, TeX, LaTeX etc.).
 
Here is an ascii-screenshot of what ECB offers you:
 
--
|  | |
| Directories  | |
|  | |
|--| |
|  | |
| Sources  | |
|  | |
|--| Edit-window(s)  |
|  |(can be splitted in two windows) |
| Methods/Vars | |
|  | |
|--| |
|  | |
| History  | |
|  | |
--
||
| Compilation-window (optional)  |
||
--
 
This is only one example-layout - ECB offers a lot of different layouts.
 
For more details go to   http://ecb.sourceforge.net/
 

Enjoy,
Klaus
 

ECB 1.95.1 released!

[Berndl, Klaus]

  ECB 1.95.1 is released!
 
 
 

What's new in the new release:
--
 
* Changes for ECB version 1.95.1
 
** Now the ecb-windows can be "maximized", means all other ecb-windows are
   deleted so only the edit-window(s), and this maximized ecb-window are
   visible (and maybe a compile-window if active). This can be done either via
   the popup-menus of the ecb-windows or via calling `delete-other-windows'
   (bound to [C-x 1]) in one of the ecb-windows or via the new commands
   `ecb-maximize-window-*' where * stands for {directories, sources, methods,
   history, speedbar}. There is also a command
   `ecb-cycle-maximized-ecb-buffers'.
 
   *Note*: If there are byte-compiled(!) user-defined layouts - either created
   interactively by `ecb-create-new-layout' or programmed with the macro
   `ecb-layout-define' - then the file where these user-defined layouts are
   saved (see option `ecb-create-layout-file') must be re-byte-compiled with
   latest ECB version >= 1.95.1! If the user-defined layouts are not
   byte-compiled then there is nothing to do.
 
** Some default keybindings have changed; see RELEASE_NOTES for details.
 
** All popup-menus of the ECB-tree-buffers can be openend via [Meta-m]
 
** Fixed bugs
 
*** Clicking onto the image-icons of an integrated speedbar has failed for
XEmacs. This comes from a bug in speedbar which is now fixed locally in
ECB by advicing `dframe-mouse-set-point' if speedbar is integrated in the
ECB-frame.
 
 
 
How to get it:
--
 
Sorry, it's too big to post the sources here...
 
If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if
you are online. ECB will then download autom. latest and newest ECB and
install it for you.
 
Or go to the homepage at   http://ecb.sourceforge.net and 
download it from
there.
 
 
 
General description of ECB:
---
 
ECB stands for "Emacs Code Browser" and is a source code browser for (X)Emacs.
It is a global minor-mode which displays a couple of windows that can be used
to browse directories, files and file-contents like methods and variables. It
supports source-code parsing for semantic-supported languages like Java, C,
C++, Elisp and Scheme as well as for source-types supported "only" by imenu or
etags (e.g. perl, TeX, LaTeX etc.).
 
Here is an ascii-screenshot of what ECB offers you:
 
--
|  | |
| Directories  | |
|  | |
|--| |
|  | |
| Sources  | |
|  | |
|--| Edit-window(s)  |
|  |(can be splitted in two windows) |
| Methods/Vars | |
|  | |
|--| |
|  | |
| History  | |
|  | |
--
||
| Compilation-window (optional)  |
||
--
 
This is only one example-layout - ECB offers a lot of different layouts.
 
For more details go to   http://ecb.sourceforge.net/
 

Enjoy,
Klaus
 
Klaus Berndl mailto:   [EMAIL PROTECTED]
sd&m AG   http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

 

ECB 1.95.1 released! - Plain text announcement

[Berndl, Klaus]

Sorry, my previous announcement was in HTML-format...and some users are
complaining about this...therefore here plain text:


  ECB 1.95.1 is released!




What's new in the new release:
--

* Changes for ECB version 1.95.1

** Now the ecb-windows can be "maximized", means all other ecb-windows are
   deleted so only the edit-window(s), and this maximized ecb-window are
   visible (and maybe a compile-window if active). This can be done either via
   the popup-menus of the ecb-windows or via calling `delete-other-windows'
   (bound to [C-x 1]) in one of the ecb-windows or via the new commands
   `ecb-maximize-window-*' where * stands for {directories, sources, methods,
   history, speedbar}. There is also a command
   `ecb-cycle-maximized-ecb-buffers'.

   *Note*: If there are byte-compiled(!) user-defined layouts - either created
   interactively by `ecb-create-new-layout' or programmed with the macro
   `ecb-layout-define' - then the file where these user-defined layouts are
   saved (see option `ecb-create-layout-file') must be re-byte-compiled with
   latest ECB version >= 1.95.1! If the user-defined layouts are not
   byte-compiled then there is nothing to do.

** Some default keybindings have changed; see RELEASE_NOTES for details.

** All popup-menus of the ECB-tree-buffers can be openend via [Meta-m]

** Fixed bugs

*** Clicking onto the image-icons of an integrated speedbar has failed for
XEmacs. This comes from a bug in speedbar which is now fixed locally in
ECB by advicing `dframe-mouse-set-point' if speedbar is integrated in the
ECB-frame.



How to get it:
--

Sorry, it's too big to post the sources here...

If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if
you are online. ECB will then download autom. latest and newest ECB and
install it for you.

Or go to the homepage at http://ecb.sourceforge.net and download it from
there.



General description of ECB:
---

ECB stands for "Emacs Code Browser" and is a source code browser for (X)Emacs.
It is a global minor-mode which displays a couple of windows that can be used
to browse directories, files and file-contents like methods and variables. It
supports source-code parsing for semantic-supported languages like Java, C,
C++, Elisp and Scheme as well as for source-types supported "only" by imenu or
etags (e.g. perl, TeX, LaTeX etc.).

Here is an ascii-screenshot of what ECB offers you:

--
|  | |
| Directories  | |
|  | |
|--| |
|  | |
| Sources  | |
|  | |
|--| Edit-window(s)  |
|  |(can be splitted in two windows) |
| Methods/Vars | |
|  | |
|--| |
|  | |
| History  | |
|  | |
--
||
| Compilation-window (optional)  |
||
--

This is only one example-layout - ECB offers a lot of different layouts.

For more details go to http://ecb.sourceforge.net/


Enjoy,
Klaus

Klaus Berndl mailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Thomas-Dehler-Str. 27, 81737 München, Germany
Tel +49 89 63812-392, Fax -220

ECB 1.96 released!!

[Berndl, Klaus]

  ECB 1.96 is released!




What's new in the new release:
--

* Changes for ECB version 1.96

** ECB can work together with the window-managers escreen and winring
   This allows to run applications like Gnus, VM or BBDB in the same frame as
   ECB! See the new lib ecb-winman-support.el and read the section
   "Window-managers and ECB" in the chapter "Tips and tricks" in the
   info-manual of ECB.
   
   Thanks to Johann Myrkraverk <[EMAIL PROTECTED]> who has
   reported some bugs related to escreen and has therefore motivated a lot to
   add support for the well known window-managers escreen and winring.

** ECB can display the window-number in the modeline of the special windows.
   The left-top-most window in a frame has the window-number 0 and all other
   windows are numbered with increasing numbers in the sequence, functions
   like `other-window' or `next-window' would walk through the frame. This
   allows to jump to windows by name as described at the Emacs-Wiki
   http://www.emacswiki.org/cgi-bin/wiki.pl?SwitchingWindows.

   See the new option `ecb-mode-line-display-window-number'. Currently this
   feature is only available for GNU Emacs 21.X, because neither GNU Emacs <
   21 nor XEmacs can dynamically evaluate forms in the mode-line.

   Thanks to Johann Myrkraverk <[EMAIL PROTECTED]> for
   suggesting this.

   In addition the already existing options `ecb-mode-line-data' and
   `ecb-mode-line-prefixes' are now more flexible and are now able to define
   special modelines not only for the 4 builtin ECB-tree-buffers but for every
   special ECB-buffer. ECB will autom. upgrade the old values of these options
   as best as possible to their new types!
   
** Much better support of the ECB-compile-window

*** General overhaul for displaying buffers in the ECB-compile-window.
The whole mechanism has been rewritten so now using a durable
compile-window is much safer and works as well as without one. Now all
buffers displayed with `display-buffer' are checked if they are
"compilation-buffers" in the sense of the function
`ecb-compilation-buffer-p' and then displayed in the compile-window of
ECB. Especially temp-buffers displayed with `with-output-to-temp-buffer'
(e.g. *Help*-buffers) are now much better supported.

Simply said: All buffers for which `ecb-compilation-buffer-p' says t are
now *always* displayed in the compile-window (if there is any) with
complete consideration of Emacs-options like `compilation-window-height',
`temp-buffer-max-height', `temp-buffer-resize-mode' (GNU Emacs),
`temp-buffer-shrink-to-fit' (XEmacs) etc. All other buffer not.

*** New layout-option `ecb-compile-window-width'.
Thanks for suggestion to John S. Yates, Jr. <[EMAIL PROTECTED]>

*** New layout-option `ecb-compile-window-prevent-shrink-below-height'

*** Better and more reliable eshell-integration. Now the compile-window fits
always autom. to the output of the last eshell-command.

*** All options related to the ECB-compile-window have been moved to the new
customization-group "ecb-compilation" which is a subgroup of "ecb-layout".

** Safer handling of all adviced functions of ECB
   All ECB-advices are now active if and only if ECB is running. Just loading
   the ECB-library doesn't activate any ECB-advice! Also if there is an error
   during the activation-process of ECB then all advices are always disabled
   automatically.

** ECB can visit tokens in a smarter way

*** Narrowing via popup-menu of the Methods-buffer includes now the documentation
of the token if located outside of the token (e.g. Javadoc for Java).

*** More flexible definition what to do after visiting a token via the
Methods-buffer. See the new option `ecb-token-visit-post-actions' which now
replaces the options ecb-highlight-token-header-after-jump',
`ecb-scroll-window-after-jump' and `ecb-token-jump-narrow'.

** ECB selects autom. the current default-directory after activation even if
   no file-buffer is displayed in the edit-window. This is useful if ECB is
   autom. activated after startup of Emacs and Emacs is started without a
   file-argument. So the directory from which the startup has performed is
   auto. selected in the ECB-directories buffer and the ECB-sources buffer
   displays the contents of this directory. See new option
   `ecb-display-default-dir-after-start'. 

** ECB preserves the split-state of the frame after activation rsp. deactivation
   This is done if the already existing option `ecb-split-edit-window' has the
   new value t.
  
** New command `ecb-toggle-window-sync' for fast toggling between autom.
   synchronization of the ECB-buffers is switched on and off.

** The popup-menu of the Directories-buffer allows opening a dir with Dired.
   The Sources-Buffer offers the same for the current selected directory. In
   the History-buffer the directory of the select

RE: ECB 2.01 released!

[Berndl, Klaus]

Berndl, Klaus wrote:
>>>Cool Stuff. I tried setting  ecb-tree-buffer-style to image and it did
>>>not change from Ascii with guide-lines. I am using GNU Emacs 21.2.1 on
>>>Windows 2K
> 
> 
>> Of course, because GNU Emacs 21.2 does not support images on windows
;-)
>> The image-code for Windows is only in the cvs.
>> So, the image-style works only for XEmacs, GNU Emasc 21 for Unix/linux
or GNU Emacs 21 cvs.

>And you can get GNU Emacs 21 cvs for windows from here :

>http://www.crasseux.com/emacs/

Yes, works very stable - i use this precompiled version already some weeks for my 
tests with next GNu Emacs release.

Klaus

RE: ECB 2.01 released!

[Berndl, Klaus]

bzip2 is contained in the cygwin-suite...

Klaus

-Original Message-
From: Suraj Acharya [mailto:[EMAIL PROTECTED] 
Sent: Friday, November 07, 2003 11:46 PM
To: James Cox
Cc: 'Jayakrishnan Nair'; [EMAIL PROTECTED]
Subject: Re: ECB 2.01 released!


Bzip2 http://sources.redhat.com/bzip2/

try tar -jxcf instead of tar -zxvf

Suraj

James Cox wrote:

> I had the same problem.  Cygwin tar didn't grok it either.  I downloaded the
> trial version of powerarchiver, that seemed to do the trick.  Maybe someone
> can clue us in on an open source tool that can do the same...
> 
> 
> 
> 
>>-Original Message-
>>From: Jayakrishnan Nair [mailto:[EMAIL PROTECTED]
>>Sent: Friday, November 07, 2003 4:39 PM
>>To: [EMAIL PROTECTED]
>>Subject: RE: ECB 2.01 released!
>>
>>
>>I downloaded the binary from the site and the file is
>>emacs_bin_cvs_2003_09_20.tar.bz2. WinZip cannot open it. Is there any
>>tool to open .bz2 files ?
>>
>>
>>-Original Message-
>>From: Suraj Acharya [mailto:[EMAIL PROTECTED] 
>>Sent: Wednesday, November 05, 2003 1:11 AM
>>To: Berndl, Klaus; [EMAIL PROTECTED]
>>Subject: Re: ECB 2.01 released!
>>
>>
>>
>>Berndl, Klaus wrote:
>>
>>>>Cool Stuff. I tried setting  ecb-tree-buffer-style to image 
>>
>>and it did
>>
>>>>not change from Ascii with guide-lines. I am using GNU 
>>
>>Emacs 21.2.1 on
>>
>>>>Windows 2K
>>>
>>>
>>>Of course, because GNU Emacs 21.2 does not support images on windows
>>
>>;-)
>>
>>>The image-code for Windows is only in the cvs.
>>>So, the image-style works only for XEmacs, GNU Emasc 21 for 
>>
>>Unix/linux
>>or GNU Emacs 21 cvs.
>>
>>And you can get GNU Emacs 21 cvs for windows from here :
>>
>>http://www.crasseux.com/emacs/
>>
>>Suraj
>>
> 
> 
> 
> This E-mail message is for the sole use of the intended recipient(s) and may
> contain confidential and privileged information.  Any unauthorized review,
> use, disclosure or distribution is prohibited.  If you are not the intended
> recipient, please contact the sender by reply E-mail, and destroy all copies
> of the original message. 
> 
> 
> 
> 

ECB 2.11 released!

[Berndl, Klaus]

  ECB 2.11 is released!




What's new in the new release:
--

* Changes for ECB version 2.11

** Using semanticdb to jump to type-tags defined in other files.
   In OO-languages like CLOS, eieio and C++ there can be type-tags in the
   method-buffer which are somehow virtual because there is no definition in
   the current source-file. But such a virtual type collects all its outside
   defined members like methods in C++ which are defined in the *.cc file
   whereas the class-definition is defined in the associated header-file.
   ECB uses semanticb to open the definition-file of such a tag and to jump to
   the definition of this tag. Same for parent-tags in the methods-buffer.
   This feature can only work correctly if semanticdb is well configured!

** New option `ecb-ignore-special-display'
   The options `special-display-buffer-names', `special-display-regexps' and
   `special-display-function' work as expected. Per default this in only true,
   when no durable compile-window is used (see option
   `ecb-compile-window-height'), i.e. with a durable compile window the
   special-display-options are ignored per default. But this behavior can be
   changed with the new option `ecb-ignore-special-display'.

   Thanks to Rob Walker <[EMAIL PROTECTED]> for a first suggestion.

** Better reporting if there are errors during ECB-startup

** Fixed bugs

*** Clicking onto nodes with positionless semantic-tags as data doesn't fail

*** Non-semantic sources (e.g. Perl, TeX) work when cedet 1.0 is loaded

*** Fixed the versioning of ECB so the autom. upgrade works



How to get it:
--

Sorry, it's too big to post the sources here...

If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if
you are online. ECB will then download autom. latest and newest ECB and
install it for you.

Or go to the homepage at http://ecb.sourceforge.net and download it from
there.



General description of ECB:
---

ECB stands for "Emacs Code Browser" and is a source code browser for (X)Emacs.
It is a global minor-mode which displays a couple of windows that can be used
to browse directories, files and file-contents like methods and variables. It
supports source-code parsing for semantic-supported languages like Java, C,
C++, Elisp and Scheme as well as for source-types supported "only" by imenu or
etags (e.g. perl, TeX, LaTeX etc.).

Here is an ascii-screenshot of what ECB offers you:

--
|  | |
| Directories  | |
|  | |
|--| |
|  | |
| Sources  | |
|  | |
|--| Edit-window(s)  |
|  |(can be splitted in two windows) |
| Methods/Vars | |
|  | |
|--| |
|  | |
| History  | |
|  | |
--
||
| Compilation-window (optional)  |
||
--

This is only one example-layout - ECB offers a lot of different layouts.

For more details go to http://ecb.sourceforge.net/


Enjoy,
Klaus

Klaus Berndl mailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Carl-Wery-Str. 42, 81739 Muenchen, Germany
Tel +49 89 63812-392, Fax -220

ECB 2.20 released!

[Berndl, Klaus]

  ECB 2.20 is released!



There is no restriction to max. 2 edit-windows. Now you can split the
edit-area in as many edit-windows as you like and need! Neither the ecb-windows
nor the compile-window will be deleted when you call C-x 1 in an edit-window!
The split-state of the edit-area will be fully preserved when hidding or showing
the ecb-windows!




What's new in the new release:
--

* Changes for ECB version 2.20

** New customization group "ecb-most-important"
   This group centralizes the most important options of ECB in one group which
   is also offered by the "Preferences" submenu ob the ECB-menu. These are
   options you should at least know that they exist.

** The keybinding for the online-help has changed from [C-c . o] to [C-c . h]
   This has been done because the key [C-c . o] is needed for the new command
   `ecb-toggle-scroll-other-window-scrolls-compile'.

** The options `ecb-major-modes-activate' and `ecb-majors-mode-deactivate' have
   been replaced by one new option `ecb-major-modes-show-or-hide'. The purpose
   of the old option is now quite out-of-date because:
   - ECB supports window-managers like winring.el or escreen.el very well,
   - as of version 2.20 ECB has no restrictions about the number of
 edit-windows and
   - the new option simplifies things a lot.
  
** The restriction of only two edit-windows has been gone!

*** Beginning with this version 2.20 there are no restrictions about the
window-layout in the edit-area of ECB. This means you can split the
edit-area of ECB in as many windows as you like. The split-state of the
edit-area will be preserved when toggling the visibility of the
ECB-windows and even between activation and deactivation of ECB. Deleting
of certain edit-windows will never destroy the special ECB-windows or the
compile-window. Just work with the edit-area of ECB as if it would be an
extra frame!

*** Option `ecb-split-edit-window' has been renamed in
`ecb-split-edit-window-after-start' because this new name reflects much
better the purpose of this option. In addition there is offered a new
value 'before-deactivation which is also the new default value. This new
value preserves the full state between activations of ECB, i.e. the
visibility of the ECB-windows, the visibility of a compile-window and also
the full split-state of the edit-area. ECB auto. upgrades your setting!

*** Compile-window can now be displayed even when the ECB-windows are hidden.
So now you can have the same compile-window functionality when the
ECB-window are hidden as when the ECB-windows are visible. The state of
the compile-window will be preserved when toggling the ecb-windows or when
maximizing one ecb-windows! So you have the advantage of one special
window for all help-, grep or compile-output also when the ecb-windows are
hidden - a window which will not be deleted if you call
`delete-other-windows' (bound to [C-x 1]) for one of the edit-windows.

*** The option `ecb-primary-mouse-jump-destination' has been renamed to
`ecb-mouse-click-destination' and has also changed its default value to
'last-point. ECB autom. upgrades the old settings.

*** New keybinding [C-c . g l] for selecting the last selected edit-window.
Also available via the ECB-menu. The command is
`ecb-goto-window-edit-last'.

*** Option `ecb-other-window-jump-behavior' has been renamed to
`ecb-other-window-behavior' and has also two new allowed values: 'smart
(the new default) and an arbitrary function-symbol. With the latter one
the user can define his own other-window-behavior and the former one
chooses in a smart and intuitive way the "other window" for going to it
via `other-window' or for scrolling it via one of the "scroll another
window"-functions (e.g. `scroll-other-window'). Thanks to John S. Yates,
Jr. <[EMAIL PROTECTED]> for suggesting the new smart behavior. ECB
autom. upgrades the old value of `ecb-other-window-jump-behavior' to the
new option-name.

*** New option `ecb-scroll-other-window-scrolls-compile-window' and new
command `ecb-toggle-scroll-other-window-scrolls-compile' (bound to [C-c .
o]). If this new option is nil then ECB chooses very smart and intuitive
the window which will be scrolled by commands like `scroll-other-window'
(see documentation of the the option `ecb-other-window-behavior'). But
sometimes the user wants to scroll the compile-window from another window.
With this new command the user can fest and easy toggle the behavior ECB
chooses another window for scrolling.

*** Higher compatibility of ECB with other packages
Without the 2-edit-window-restriction ECB is now more compatible w

ECB 2.21 released!

[Berndl, Klaus]

  ECB 2.21 is released!




What's new in the new release:
--

* Changes for ECB version 2.21

** Advice for `balance-windows' so only the edit-windows are balanced.
   Thanks to David Ishee <[EMAIL PROTECTED]> for suggestion.

** New option `ecb-ignore-display-buffer-function'
   Per default the adviced version of `display-buffer' ignores the value of
   `display-buffer-function' when called for the ECB-frame. If this variable
   should not be ignored then the function of `display-buffer-function' is
   completely responsible which window is used for the buffer to display - no
   smart ECB-logic will help to deal best with the ECB-window-layout! You can
   define if and when `display-buffer-function' should be ignored: Always
   (default), when a compile-window is used or never.

** Compatibility enhancements
   In general the current layout-engine of ECB is so flexible that there
   should be no - or at least very few - conflicts between ECB and any other
   elisp-library - even when another lib is running during the ECB-windows are
   visible.

*** Commands like `bs-show' of the library bs.el now work correctly with ECB

*** Xrefactory works even when all ECB-windows are visible (see Fixed Bugs).

*** Applications like Gnus or BBDB run withing the ECB-frame without conflicts -
even when the ECB-windows are visible.

*** Commands using `Electric-pop-up-window' now work correctly with ECB.
This ensures that the electric-* commands (e.g. `electric-buffer-list' or
`electric-command-history') work well with ECB. If the related
"display-buffer" of such an electric command is a "compilation-buffer" in
the sense of `ecb-compilation-buffer-p' then this buffer will be displayed
in the compile-window of ECB - if there is any shown.

** For XEmacs the package fsf-compat is no longer required by ECB.
   But it is still necessary to check if fsf-compat is required by the
   packages semantic, eieio and speedbar which in turn are required by ECB.

** Fixed Bugs

*** Fixed a fatal bug which prevents `other-window' from working with
arguments < 0. This bug has also prevented Xrefactory from working correct
when ECB is active - see compatibility enhancements above.

*** If point stays in the current source-buffer on a function argument then still
ECB highlights the function in the Methods-buffer.

*** ECB now uses `compilation-window-height' correctly when set buffer-local
as possible with latest CVS-version of GNU Emacs 21.

*** Fixed a bug in `ecb-sources-filter' and `ecb-history-filter'.



How to get it:
--

Sorry, it's too big to post the sources here...

If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if
you are online. ECB will then download autom. latest and newest ECB and
install it for you.

Or go to the homepage at http://ecb.sourceforge.net and download it from
there.



General description of ECB:
---

ECB stands for "Emacs Code Browser" and is a source-code-browser for (X)Emacs.
It is a global minor-mode which offers a language-independent and complete IDE
(Integrated Development Environment) within one Emacs-frame. It displays a
couple of windows that can be used to browse directories, files and
file-contents like methods and variables. It supports source-code parsing for
semantic-supported languages like Java, C, C++, Elisp and Scheme as well as
for source-types supported "only" by imenu or etags (e.g. perl, TeX, LaTeX
etc.). In addition it offers (optional) a durable "compile-window" at the
bottom of the frame which is used to display all help-, grep-, compile- and
etc.-output. The rest of the frame is called the "edit-area" which can be
divided in several (no limit) edit-windows which are used for editing of
sources. Deleting some of the edit-windows does neither destroy the
compile-window nor the browsing-windows.

Here is an ascii-screenshot of what ECB offers you:

--
|  | |
| Directories  | |
|  | |
|--| |
|  | |
| Sources  | |
|  | |
|--|   Edit-area |
|  |(can be splitted in several edit-windows)|
| Methods/Vars | |
|  | |
|--| |
|  | |
| History  |

ECB 2.22 released!

[Berndl, Klaus]

  ECB 2.22 is released!




What's new in the new release:
--

* Changes for ECB version 2.22

** New nifty feature for filtering the tags displayed in the Methods-buffer
   You can filter by regexp, protection, tag-class etc. See the new command
   `ecb-methods-filter' (bound to [C-c . f m] and the new option
   `ecb-methods-filter-replace-existing' for getting all details.

   Filters can be layered filters, i.e. you can apply filters on top of
   already applied filters: An example of 3 combined filters would be: Display
   only all public methods having the string "test" in its name. The modeline
   of the Methods-buffer always displays the topmost filter. But you can see
   all currently applied filters by moving the mouse over this section of the
   modeline: The filters will be displayed as help-echo.

   The new functionality is also completely available via the popup-menu of
   the Methods-buffer.

   Currently tags of class 'type can not be filtered out, because especially
   in OO-languages like Java or C++ this makes only less sense because if the
   class is filtered out then no filter can be applied to its members like
   methods or attributes.

   These tag-filters can also applied to sources which are not supported by
   the semantic-parser but "only" by imenu or etags. But because for these
   sources not all information are avaiable the protection- and tag-class
   filter can not used with such "non-semantic"-sources.

** All tree-buffer filters are now guarded against filtering out all tags.
   Each filter which a user can apply to the sources-, methods- or
   history-buffer is now guarded against filtering out all nodes by accident,
   ie. if a filter would empty the whole tree-buffer then this filter will be
   autom. unwinded.
   
** Much smarter mechanism to highlight the current tag in the methods-buffer.
   Previous versions of ECB have always fully expanded the whole tree in the
   Methods-buffer if the current tag in the source-buffer was not visible in
   the current tree - e.g. because the variables-bucket was collapsed or the
   containing type of a tag (e.g. the class of a method) was collapsed. So in
   most cases much more was expanded as needed to make the current tag
   visible.

   The mechanism of ECB 2.22 only expands the needed parts of the tree-buffer
   to make the related node visible: First we try to highlight the current tag
   with current expansion-state of the Methods-buffer. If the node is not
   visible so the tag can not be highlighted then we go upstairs the ladder of
   type-tags the current tag belongs to (e.g. we expand successive the nodes
   of the whole class-hierachy of the current method-tag until the related
   node becomes visible). If there is no containing type for the current tag
   then the node of the tag is probably contained in a toplevel-bucket which
   is currently collapsed; in this case we expand only this bucket-node and
   try highlighting again. Only if this has still no success then we expand
   the full tree-buffer and try to highlight the current tag.

   There is also a new option `ecb-auto-expand-tag-tree-collapse-other': If
   this option is set then auto. expanding the tag-tree collapses all not
   related nodes. This means that all nodes which have no relevance for the
   currently highlighted node will be collapsed, because they are not
   necessary to make the highlighted node visible. This feature is switched
   off by default because if always collapses the complete Methods-tree before
   the the following highlighting of the node for the current tag expands the
   needed parts of the tree-buffer.
   
** The popup-menus can now be nested into 4 levels of submenus. In general there
   could be an infinite depth of nesting but it makes no sense - if possible
   at all - to define infinite nested defcustom-types. So there is a limit of
   4 levels until a user asks for more - then this wish can be satisfied
   during 1 minute: Just increase the value of the `ecb-max-submenu-depth'
   before loading ECB. This enhancement affects the four options with name
   `ecb-*-menu-user-extension'.
   
** The indication for ECB in the minor-mode is now hidden when the ECB-windows
   are visible because in this case the activity of ECB is quite obvious. When
   the ECB-windows are hidden the the value of `ecb-minor-mode-text' is
   displayed in the modeline.

** Compatibility enhancements

*** Cause of the fixed bugs the compatibility for `bs-show' has increased when
the buffer-selection should be displayed in the permanent compile-window.

** Fixed Bugs

*** Fixed a bug preventing the native Windows-port of XEmacs from working.

*** Makes the behavior of adviced `display-buffer' fully compatible to the
original version for all buffers which are neither ecb-tree-buffer nor
compilation-buffers in the sense of `ecb-compilation-buffer-p'. These
"source-buffers" are now 

RE: automatic adding to hs-special-modes-alist??

[Berndl, Klaus]

I second this - ECB already recommend this in its manual, so the hideshow-integration 
into ECB can work with JDEE too...

Klaus


>I have this snippet in my .emacs, would it be a good idea if jde would
>do something like this internally? It is just the same definition as for
>java-mode, copied from hideshow.el.



>(add-to-list 'hs-special-modes-alist '(jde-mode "{" "}" "/[*/]" nil
>hs-c-like-adjust-block-beginning))


jtl

AW: auto newline, indent and close brace on open brace, return

[Berndl, Klaus]

Here is a better way to determine the syntactiy context of current point, means if 
point stays within a line-comment, block-comment or within a string. IMHO this is more 
robust than jde-line-has-incomplete-string...and uses well proved Emacs-concepts for 
this instead of fiddling with some regexps and increasing numbers ;-)

(defun syntactic-context ()
  "Check in which syntactic context point is. Return nil if no special context
meaning, otherwise 'string if within a string, 'comment if within a
line-comment and 'block-comment if within a block-comment."
  (let* ((beg (save-excursion
(beginning-of-defun)
(point)))
 (state (save-excursion
  (parse-partial-sexp beg (point)
(cond
 ((nth 3 state) 'string)
 ((nth 4 state) (if (member major-mode
'(c-mode c++-mode java-mode jde-mode))
(if (nth 7 state) 'comment 'block-comment)
  'comment))
 (t nil

This function returns 'string if point is within a string - so also when point is at 
the end of an unterminated string.
In that situation a newline-command should insert a (java)string terminator and so on 
... As already done by the code of this thread. This has the side-effect that when 
point stays within a terminated string a newline-command breaks this string by adding 
a new terminator behind the break...so the smart newline-command does not only for 
unterminated strings the right thing but also for terminated.

Thoughts?

BTW: I do not want to upset someone who contributes to this thread and code, but i 
have the strong feeling, that the most things could be done more generally with 
concepts and mechanism of the underlying cc-mode. I'm think into the direction of 
`c-hanging-braces-alist' and all this stuff...IMHO JDEE and also this contributed new 
code reinvents the wheel in some aspects...

Off topic: JDEE does this also with all the template-stuff - where IMHO somehow 
cumbersomely is specified if a newline after a brace or not etc... all this could be 
done more nifty with mechanism of cc-mode and tempo, so the user specifies with 
cc-mode when he wants newlines before or after praces etc. and tempo uses this 
informations instead of introducing new options by JDEE so the user has to customize 
the same thing at differrent places.

Klaus

-Ursprüngliche Nachricht-
Von: Paul Kinnucan [mailto:[EMAIL PROTECTED] 
Gesendet: Freitag, 12. März 2004 23:04
An: Suraj Acharya
Cc: [EMAIL PROTECTED]; Paul Kinnucan; Petter Måhlén; [EMAIL PROTECTED]
Betreff: Re: auto newline, indent and close brace on open brace, return


Hi All,

I've been saving the email on this thread and will review it over the weekend. My plan 
is to include the latest and greatest version that comes out of this discussion.

Thanks,

Paul

ECB 2.23 released!

[Berndl, Klaus]

  ECB 2.23 is released!




What's new in the new release:
--

* Changes for ECB version 2.23

** New cedet1.0beta2 is supported.

** Enhancements for the Methods-buffer

*** The look&feel of the Methods-buffer is now much nicer because it has been
polished with a lot of new icons. ECB contains now a lot of new icons for
displaying the Method-buffer as pretty and professional as possible. The
new default-value of the option `ecb-tree-image-icons-directories' points
autom. to the new icon-images. Thanks a lot to Markus Gritsch
<[EMAIL PROTECTED]> for providing these new icons.

To use this you should reset the option `ecb-tree-image-icons-directories'
to the new default-value if you have customized this option!

You can disable this icon-display via the new option
`ecb-display-image-icons-for-semantic-tags' if you do not like the new
icons. 

*** New feature for applying default tag-filters for certain files.
The new option `ecb-default-tag-filter' allow to define default
tag-filters for certain files which are applied automatically after
loading such a file into a buffer. The possible filters are the same as
offered by the command `ecb-methods-filter' and they are applied in the
same manner - the only difference is they are applied automatically. The
files can be specified on a combined major-mode- and
filename-regexp-basis.

Usage-example: This can be used to display for outline-mode files (e.g.
NEWS) in the Methods-buffer only the level-1-headings by defining a
regexp-filter "^\* .*" (see the default-value of this new option).

*** New keybindings and commands for faster applying certain tag-filters.
Here are the new bindings:
[C-c . fr] --> `ecb-methods-filter-regexp'
[C-c . ft] --> `ecb-methods-filter-tagclass'
[C-c . fp] --> `ecb-methods-filter-protection'
[C-c . fn] --> `ecb-methods-filter-nofilter'
[C-c . fl] --> `ecb-methods-filter-delete-last'
[C-c . ff] --> `ecb-methods-filter-function'

*** The popup-menu contains now mode-dependend tag-filter entries.
This means for sources not supported by semantic no protection- or
tag-class filters will be offered. And for semantic-supported sources
exactly these tag-classes are offered the semantic-parser for the current
major-mode offers. Same for the command `ecb-methods-filter'. For example
texi-sources can only be filtered by the tag-classes "Definitions" and
"Sections" and java-sources can be filtered by "Methods", "Variables",
"Classes" etc. In general the semantic-variable
`semantic-symbol->name-assoc-list' is used to get the right tag-classes.

*** The option `ecb-show-tags' is now defined on a major-mode-basis.
This means you can have different settings for each major-mode. ECB autom.
upgrades your old setting to the new option-type.

*** Distinction between functions and function-prototypes in the Methods-buffer
This is for example useful for C++ and C because these languages distinct
between a method-prototype (rsp. function-prototype for C) and the method
(rsp. function for C) itself. The new default value of `ecb-show-tags'
displayes per default the methods as flattened and the method-prototype as
collapsed when opening a C- or C++-buffer. ECB autom. updates your old
setting of this option to the new default value.

** The command `ecb-toggle-layout' now has a prefix-argument:
   If this optional argument is not nil (e.g. if called with a prefix-arg)
   then always the last selected layout was choosen regardless of the setting
   in `ecb-toggle-layout-sequence'. The last selected layout is always that
   layout which was current direct before the most recent layout-switch. So
   now a user can switch to another layout via `ecb-change-layout' [C-c . l c]
   and always come back to his previous layout via [C-u C-c . l t].

** Better internal self-monitoring of ECB
   ECB now monitors itself when Emacs is idle, ie. it monitors if all
   necessary ecb-functions are still member of the hooks `post-command-hook'
   and `pre-command-hook'. This is because Emacs resets these hooks to nil if
   any of the contained hook-functions fails with an error-signal. The
   ECB-hook-functions are error-save but if any of the "other" hook-functions
   of these hooks fails then the ECB-hooks would also have been removed from
   these hooks. The new monitoring-mechanism checks this periodically and
   re-adds the ECB-hooks if necessary.

** New option group for the integrated speedbar and one new option:
   The new hook `ecb-speedbar-before-activate-hook' runs directly before ECB
   activates the integrated speedbar. For example this hook can be used to
   change the expansion-mode of the integrated speedbar via
   `speedbar-change-initial-expansion-list'. Example:
   (speedbar-change-initial-expansion-list "buffers").

** Added a new section

RE: Roadmap for Java 1.5 support ?

[Berndl, Klaus]

Hervé Bitteur wrote:
> Hello,
> 
> I've switched to Java 1.5 beta for more than one month now. I just
> made the bare minimum in terms of modifications in JDEE lisp files,
> since I'm not at all a Lisp expert.
> 
> I can keep on using JDEE as my favorite development environment, since
> browsing, text editing, compiling, running, ant builds, ... are OK.
> 
> But of course I lack full support for:
> - Syntax highlighting
> - Jalopy formatting
> - ECB

what is missed in ECB concerning supporting Java 1.5?

> - (perhaps debug, but I don't use it with JDEE anyway)
> 
> Switching a whole Java application to Java 1.5, taking full advantage
> of its new features such as generics, is not something you can easily
> undo. So now I'm on the 1.5 side and I stay there :-)
> 
> My question is thus about your roadmap for upgrading this nice
> environment to a full support of Java 1.5.

ECB and JDEE are independent from each other

Klaus

RE: Roadmap for Java 1.5 support ?

[Berndl, Klaus]

Hervé Bitteur wrote:
> Klaus,
> 
> On the 2 following methods, I noticed an inconsistent behavior:
> 
> // First method
> public void setVerticalLag (VerticalLag vLag)
> {
>  this.vLag = vLag;
> }
> 
> // Second method
> public VerticalLag getVerticalLag ()
> {
>  return vLag;
> }
> 
> The former is correctly flagged as a public method (according to the
> small icon in the hierarchical display), while the latter is flagged
> as a package private method.
> 
> I just tried this with using your latest ECB version (2.23).
> 
> Of course, I can survive this ...

This is not a matter of ECB but its a matter of the semantic-package. ECB is
just an ontop-displayor for the parsing-results of the semantic-package.
if semantic supports parsing a grammar then ECB does it autom. too if semantic
doesn't it ECB doesn't ist too.

AFAIK David (Ponce) has checked in a new wisent java-grammar which has full(!)
Java 1.5 support - i do not know if this is contained in the released beta2 -
but at least when you use the newest CVS-cedet you should get it. You can ask
for this in the cedet-mailing-list!

Klaus

RE: JDE 2.3.4beta3

[Berndl, Klaus]

Prof. Dr. Jobst Hoffmann wrote:
> I started JDE 2.3.4beta3 and the following error came out:
> 
> (1) (error/warning) Error in
> `post-command-hook' (setting hook to nil): (wrong-type-argument
> string-or-itimer-p nil)
> 
> Just another little glitch: when I'm starting a console application
> (C-c C-v C-r) the first time, the cursor is positioned at the right
> place (after the prompt for example). When I'm doing the same thing a
> second time, the cursor is positioned at the top of the window (line
> 1, column 0) in the text coming from the bean shell (cd ...)
> 
> I'm using XEmacs 21.4.15, jde 2.3.4beta3, and semantic 1.4.4
> 
> By the way, what is your preferred directory structure in the future:
> jde, semantic ... in the same level or jde and cedet in the same level
> and semantic ... below cedet?

The new cedet suite introduces a new directory-structure for all subcomponents
of cedet:

cedet
  - common
  - eieio
  - speedbar
  - semantic
  - ...

At least this true until cedet will become a XEmacs-packages and has to
be included in the given directory-structure of all XEmacs-packages.
But this is somehow future-music ;-)

Where all the other packages like JDEE, ECB or others stay is completely
independent from cedet - but a comman way would be to place jdee and cedet
in the same level...

Klaus

> 
> Jobst

RE: Question/Bug to 'jde-detect-java-buffer-activation

[Berndl, Klaus]

>> ;; to a buffer belonging to another.
>> (make-local-hook 'post-command-hook)

>Hmm, I don't think I need the above form because the add-hook form below
>makes the variable local anyway.

Indeed.

> > (add-hook 'post-command-hook
> >   'jde-detect-java-buffer-activation nil t)
> > 
> >  
> > This is the code from `jde-mode' (JDEE 2.3.4beta3).
>This should work even if the "from" buffer is not a jde-mode buffer. Are
>you saying that it doesn't?

Yes, i'm saying this - and it can not work in other modes:
This add-hook is called when `jde-mode' is called which is only called when a buffer X 
is opened which is "linked" to jde-mode. And because you add the function LOCALLY to 
post-command-hook only the local value (local in buffer X) contains this function - 
the global value of post-command-hook doesn't. Therefore this works only in 
jde-buffers - which is suboptimal IMHO ;-)

>Then the hook function would run in buffers that are not jde-mode
>buffers? This seemed a waste
>of processing cycles to me when I wrote the code.

No.
1. It should run in other mode buffers
2. You are right, that then the check if the current-buffer is running 
   jde-mode would be processed after each command in each buffer but this 
   is really not a problem - A powerful Emacs-setup contains some functions 
   in post-command-hook and this is not a problem if these funtions are 
   well designed, i.e. do their task only if needed and check this with a 
   efficient condition - and (equal major-mode 'jde-mode) is very efficient 
   ;-)

Thoughts?

>I will try to make the doc clearer on these issues.

Would be better to fix the things above ;-)

>jde-entering-java-buffer-hook functions are supposed to run whenever the
>user activates a jde-mode buffer,
>regardless of the mode of the previously active buffer. As far as I
>know, it works correctly and that
>is easily verifiable. Are you saying that the hook functions are not
>being run when a jde-mode buffer
>is entered from a buffer that is operating in a different mode?

Yes, indeed! See above...

Klaus

RE: Customizing JDE

[Berndl, Klaus]

It can:

I have the following to enable a side-wide Emacs-setup also for jdee:

First 2 help-funktions:

(defun customize-option-get-value (option type)
  "Return the value of a customizable option OPTION with TYPE, where
TYPE can either be 'standard-value \(the default-value of the
defcustom) or 'saved-value \(the value stored durable by the user via
  customize)." (let ((val (car (get option type
(cond ((not (listp val)) val)
  ((equal 'quote (car val)) (car (cdr val)))
  (t (car val)

(defun customize-save-variable-save (option value &optional override)
  "Calls `customize-save-variable' with OPTION and VALUE if OPTION is a
custom-type and if OPTION has no saved-value until now.
If OVERRIDE is a function or lambda-form then it is called with two
arguments: 
- OLD-SAVED-VAL: The saved value of OPTION
- NEW-VALUE: see argument VALUE.
OVERRIDE is only called if OPTION has already a saved-value. If OVERIDE
returns not nil then `customize-save-variable' is called for OPTION
with VALUE even if OPTION has no saved-value until now."
  (and (get option 'custom-type)
   (or (not (get option 'saved-value))
   (and (functionp override)
(funcall override
 (customize-option-get-value option
 'saved-value) value)))
   (progn
 (message "Overriding saved value for option %s with %s" option
 value) (customize-save-variable option value


Then you can customize jdee-veriables programatically like follows:

;; Here we set some options of JDE for the whole site:
;; We set default values for the following options but this default
valuies ;; will only be set if the user has *NOT* already customized
these options for ;; himself.
(customize-save-variable-save 'jde-mode-line-format
default-mode-line-format) (customize-save-variable-save
  'jde-java-font-lock-api-file
"~/.jde-java-font-lock.api") (customize-save-variable-save
'jde-complete-insert-method-signature nil)
(customize-save-variable-save 'jde-auto-parse-buffer-interval 240)
(customize-save-variable-save 'jde-auto-parse-enable t)
(customize-save-variable-save 'jde-auto-parse-max-buffer-size 0)
(customize-save-variable-save 'jde-which-method-mode nil)
(customize-save-variable-save 'jde-imenu-enable nil)
(customize-save-variable-save 'jde-imenu-sort (quote asc)) ;; If a user
has set jde-check-version-flag to not nil then we override this ;; with
nil! (customize-save-variable-save 'jde-check-version-flag nil
  (function (lambda (old-val new-val)
  ;; we return not nil (i.e.
  override ;; the old-value
  with nil when the ;; current
  set value is not nil. ;;
  Otherwise there is no need to
  ;; change the setting. We
  could also ;; simply return t
  but this is more ;; charming
;-) old-val))) 

Does this help?

But im not sure if and how this work together with the
project-management of JDEE?! But if not maybe it can be a good starting
point for developing an enhanced version... ;-) 

Klaus

Nascif Abousalh-Neto wrote:
> I think the question is, can it be done programatically?
> -Original Message-
> From: Berndl, Klaus [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 09, 2004 2:52 PM
> To: [EMAIL PROTECTED]
> Subject: FW: Customizing JDE
> 
> 
> M-x customize-option RET jde-make-program RET
> 
> This opens a customize-buffer where you can set an appropriate value
> for an option. 
> 
> Klaus
> 
> -Original Message-
> From: Nallathambi, Balaguru [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 09, 2004 8:01 PM
> To: Berndl, Klaus; '[EMAIL PROTECTED]'
> Subject: RE: Customizing JDE
> 
> 
> How do I customize via customize?
> What is AFAIK?
> -Original Message-
> From: Berndl, Klaus [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 09, 2004 4:58 AM
> To: Nallathambi, Balaguru; '[EMAIL PROTECTED]'
> Subject: RE: Customizing JDE
> 
> 
> AFAIK you can not customize JDEE-options via setq - JDEE accepts only
> customizations done via customize! 
> 
> Klaus
> 
> -Original Message-
> From: Nallathambi, Balaguru [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, June 08, 2004 8:44 PM
> To: '[EMAIL PROTECTED]'
> Subject: Customizing JDE
> 
> 
> We have our custom build environment, which generates shell wrapper
> scripts for java programs. 
> I managed to configure 

ECB 2.25 released!

[Berndl, Klaus]

  ECB 2.25 is released!




What's new in the new release:
--

* Changes for ECB version 2.25

** More flexible sorting of the Sources- and the History-buffer

*** `ecb-sort-history-items' has been renamed to `ecb-history-sort-method'
and offers now the same choices as the already existing option
`ecb-sources-sort-method'.

*** Two new options `ecb-sources-sort-ignore-case' and
`ecb-history-sort-ignore-case' which allow to ignore case when sorting the
Sources- and/or the History-buffer. Thanks for suggestion to Markus
Gritsch <[EMAIL PROTECTED]>. Per default case is now ignored.

** New icons for parent-display in the Methods-buffer.
   Thanks to Markus Gritsch <[EMAIL PROTECTED]> for contributing the
   icon-images.

** Fixed Bugs

*** Fixed an annoying bug which results sometimes in an error "stack-overflow
error in equal" when using `senator-try-expand-semantic' which is called
when you use hippie-expand for example. Maybe the bug occured also in
other situation but now this bug has been extirpated!

*** Fixed a bug in the mechanism which prescanes directories for emptyness.
Now a directory is checked if it is accessible before it is prescanned -
otherwise the prescan could fail with an error.

*** Fixed a bug in the autom. option-upgrading-mechanism.
Thanks to Javier Oviedo <[EMAIL PROTECTED]> for pointing to the
underlying problem.



How to get it:
--

Sorry, it's too big to post the sources here...

If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if
you are online. ECB will then download autom. latest and newest ECB and
install it for you.

Or go to the homepage at http://ecb.sourceforge.net and download it from
there.



General description of ECB:
---

ECB stands for "Emacs Code Browser".  While Emacs already has good
*editing* support for many modes, its *browsing* support is somewhat
lacking. That's where ECB comes in: it displays a number of informational
windows that allow for easy source code navigation and overview.

The informational windows can contain:

  - A directory tree,
  - a list of source files in the current directory,
  - a list of functions/classes/methods/... in the current file, (ECB uses the
Semantic Bovinator, or Imenu, or etags, for getting this list so all
languages supported by any of these tools are automatically supported by
ECB too)
  - a history of recently visited files, 
  - the Speedbar and
  - output from compilation (the "*compilation*" window) and other modes like
help, grep etc. or whatever a user defines to be displayed in this window.

As an added bonus, ECB makes sure to keep these informational windows visible,
even when you use C-x 1 and similar commands.

It goes without saying that you can configure the layout, ie which
informational windows should be displayed where. ECB comes with a number of
ready-made window layouts to choose from.

Here is an ascii-screenshot of what ECB offers you:

--
|  | |
| Directories  | |
|  | |
|--| |
|  | |
| Sources  | |
|  | |
|--|   Edit-area |
|  |(can be splitted in several edit-windows)|
| Methods/Vars | |
|  | |
|--| |
|  | |
| History  | |
|  | |
--
||
| Compilation-window (optional)  |
||
--

For more details go to http://ecb.sourceforge.net/


Enjoy,
Klaus


Klaus Berndl mailto: [EMAIL PROTECTED]
sd&m AG http://www.sdm.de
software design & management
Carl-Wery-Str. 42, 81739 Muenchen, Germany
Tel +49 89 63812-392, Fax -220

RE: Reporting problems concerning JDEE and ECB (Was: Trouble with JDE Completion)

[Berndl, Klaus]

Ok, thanks for the information... But in general it would be good to mention this in 
the first problem-posting becuase otherwise phrases like "I'm using ECB 2.2.5 and JDEE 
2.3.4" could cause confusion - at least in MY brain ;-))

Klaus

-Original Message-
From: exits funnel
To: Berndl, Klaus; '[EMAIL PROTECTED] '
Cc: '[EMAIL PROTECTED]'
Sent: 24.06.2004 14:45
Subject: Re: Reporting problems concerning JDEE and ECB (Was: Trouble with JDE  
Completion)

Klaus,

This problem has nothing to do with ECB which is why I
posted it to the JDE list.  I just triple checked this
by running JDE with ECB not even loaded and I see the
same behavior.  Thanks for taking the time to reply.

-exits

--- "Berndl, Klaus" <[EMAIL PROTECTED]> wrote:
> Not an answer to the problem below but a
> recommendation which would make it easier for me or
> Paul to reproduce problems.
> 
> It seems that there are some problem-reports either
> to the JDEE- or the ECB-mailing-list where the user
> writes he/she uses ECB and JDEE in combination. It
> would always good to test in which situations a
> problem occurs:
> 
> - ECB and JDEE active
> - only JDEE active but ECB deactivated
> - only ECB with deactivated JDEE (the latter one
> means, java-sources are 
>   running default java-mode and not jde-mode)
> 
> Then we can faster decide who has the trouble with
> the problem-fix ;-)
> 
> Klaus
> 
> -Original Message-
> From: exits funnel
> To: [EMAIL PROTECTED]
> Sent: 24.06.2004 03:38
> Subject: Trouble with JDE Completion
> 
> --text follows this line--
> 
> Please enter the details of your bug report here
> 
> Hello.  I'm having trouble using completion.  If I
> open a java buffer and enter the following:
> 
> String s; s.
> 
> and then, with point sitting just after the dot I do
> an M-x jde-complete I get an error which begins:
> 
> jde-parse-import-list: Wrong number of arguments:
> 
> I know very little lisp but I grepped through the
> jde
> code and it seems that jde-parse-import-list takes
> no
> arguments and is never called with any.  For this
> simple case my jde-global-classpath is nil.  I don't
> need to add anything for completion to work against
> the java.lang classes do I?  I should also mention
> that though I've just recently upgraded to
> JDE234Beta5
> and ECB2.2.5, I'm almost certain that this problem
> existed when I was running Beta4 against 2.2.4.  If
> I
> can provide any further information, let me know. 
> Thanks in advance!  Also, thanks for taking the time
> to work on the JDEE.  I'm just starting to come up
> to
> speed but so far it seems really promising.
> 
> -exits
> 
> Emacs  : GNU Emacs 21.3.1 (i386-redhat-linux-gnu, X
> toolkit, Xaw3d scroll bars)
>  of 2003-05-22 on porky.devel.redhat.com

> Package: JDE version 2.3.4beta5
> Required packages: cedet-1.0beta2
> 
> 
> current state:
> ==
> (setq
>  jde-gen-session-bean-template
> '("(jde-import-insert-imports-into-buffer " "  (list
> \"javax.ejb.*\""
>"
> \"java.rmi.RemoteException\"))" "'>"
> "(jde-wiz-update-implements-clause \"SessionBean\")"
> "'>"
>"(jde-gen-method-signature" "
> \"public\"" " 
> \"void\"" "  \"ejbActivate\"" "  nil"
>"  \"RemoteException\"" " )" "'>" "(if
> jde-gen-k&r " " ()" " 'n)" "\"{\"'>'n" "\"}\"'>'n
> 'n"
>"(jde-gen-method-signature" "
> \"public\"" " 
> \"void\"" "  \"ejbPassivate\"" "  nil"
>"  \"RemoteException\"" " )" "'>" "(if
> jde-gen-k&r " " ()" " 'n)" "\"{\"'>'n" "\"}\"'>'n
> 'n"
>"(jde-gen-method-signature" "
> \"public\"" " 
> \"void\"" "  \"ejbRemove\"" "  nil" " 
> \"RemoteException\""
>" )" "'>" "(if jde-gen-k&r " " ()" "
> 'n)"
> "\"{\"'&

ECB 2.26 is released!

[Berndl, Klaus]

  ECB 2.26 is released!




What's new in the new release:
--

* Changes for ECB version 2.26

** Improved the performance of the directories-buffer-display
   Reduced the need of completely rebuilding the whole directories-tree-buffer
   when just switching between buffers and the related directories are already
   expanded. This can dramatically increase the speed of the tree-buffer
   display when displaying the sources-files in the directories-buffer (see
   `ecb-show-sources-in-directories-buffer'). Thanks to Rob Walker
   <[EMAIL PROTECTED]> for tracking down the performance bottleneck and
   supplying a first patch.

** Per default all mouse-actions in the special ECB-buffers are now triggered
   first after releasing the mouse-button and not when pressed as with
   previous ECB-versions. This is the standard behavior of Emacs and of most
   applications. But with the new option `ecb-tree-mouse-action-trigger' a
   user can switch back to the old behavior.

   As a side effect now the ECB-tree-buffers can be used during an active
   isearch - at least with GNU Emacs. Thanks to Markus Gritsch
   <[EMAIL PROTECTED]> for pointing to this problem.

** ECB sets autom. temp-buffer-resize-mode rsp. temp-buffer-shrink-to-fit when
   needed. This modes are needed when a permanent compile-window is used
   because otherwise the correct sizing of the compile-window according to the
   settings in ecb-compile-window-temporally-enlarge and
   ecb-enlarged-compilation-window-max-height is not possible. After disabling
   the permanent compile-window or when deactivating ECB the original settings
   of these modes are restored!
  
** Some regexp-options has been changed to regexp-list-options:
   `ecb-excluded-directories-regexp' --> `ecb-excluded-directories-regexps'
   `ecb-source-file-regexps' --> `ecb-source-file-regexps'
   `ecb-exclude-parents-regexp'  --> `ecb-exclude-parents-regexps'

   All these options now accept (and require) a list of regexps instead of
   exactly one big regexp. This allows easier regexp-definition and follows
   now the way Emacs goes with other regexp-options like `same-window-regexps'
   or `special-display-regexps' which are also lists of regexps.

   ECB autom. upgrades your old settings to the new option-types rsp. also
   -names (if the option has been renamed).

** New option `ecb-history-exclude-file-regexps' for excluding certain files
   from being displayed in the history-buffer of ECB. Currently the filenames
   TAGS and semantic.cache are excluded per default. Thanks to Javier Oviedo
   <[EMAIL PROTECTED]> for the suggestion.

** ECB now displays read-only sourcefiles in a different face.
   For this a new option and a new face are offered both named
   `ecb-source-read-only-face' (default: italic) which is used to display
   source-files in the sources-tree-buffer (or in the directories-tree-buffer
   if `ecb-show-sources-in-directories-buffer' is not nil) if the file is
   read-only.
   
** Fixed bugs

*** Fixed merging faces with XEmacs.
Now options like `ecb-type-tag-display' (merges own ECB-faces to already
by semantic faced tags) work also correct with XEmacs. ECB-versions prior
to 2.26 has not merged the ECB-faces to the already applied faces of a
text but instead replaced the applied faces with the ECB ones. Now all
faces are merged.

*** Now the command `ecb-expand-methods-nodes' and all the expand-* menu-entries
in the popup-menu of the methods-buffer work for non-semantic-buffers too.

*** Fixed a bug concerning `ecb-auto-expand-tag-tree-collapse-other'.
Thanks to Javier Oviedo <[EMAIL PROTECTED]> for pointing to the problem.

*** Handles not existing source-paths in `ecb-source-path' correct.



How to get it:
--

Sorry, it's too big to post the sources here...

If you are using ECB >= 1.80 then you can just call "M-x ecb-download-ecb" if
you are online. ECB will then download autom. latest and newest ECB and
install it for you.

Or go to the homepage at http://ecb.sourceforge.net and download it from
there.



General description of ECB:
---

ECB stands for "Emacs Code Browser".  While Emacs already has good
*editing* support for many modes, its *browsing* support is somewhat
lacking. That's where ECB comes in: it displays a number of informational
windows that allow for easy source code navigation and overview.

The informational windows can contain:

  - A directory tree,
  - a list of source files in the current directory,
  - a list of functions/classes/methods/... in the current file, (ECB uses the
Semantic Bovinator, or Imenu, or etags, for getting this list so all
languages supported by any of these tools are automatically supported by
ECB too)
  - a history of recently visited files, 
  - the Speedbar and
  - output from compilation (the "*compilation*" window) and other modes like
help, grep

AW: ecb and jde 2.3.4

[Berndl, Klaus]

Title: AW: ecb and jde 2.3.4






From the ECB-maintainer:

Its very important, to ensure that the version of semantic/cedet which are used for bytecompiling ECB and which are loaded into Emasc are identical

If you now run JDE with cedet instead of the old semantic 1.X versions and you have ECB byte-compiled then you have to re-byte-compile ECB! Either via makefile (search the variable CEDET in the makefile) or remove all *.elc files of ECB, start Emacs and ECB unbyte-compiled and then run `ecb-bytecompile-ecb' - this will then automatically bytecompile ECB with that version of cedet/semantic/eieio/speedbar which is currently lloaded into Emacs!

In general ECB runs already a long time very well without any problems with all versions of the cedet-suite!

Ciao
Klaus


-Ursprüngliche Nachricht-
Von: Michael Schupp [mailto:[EMAIL PROTECTED]]
Gesendet: Mi 17.11.2004 13:35
An: [EMAIL PROTECTED]
Betreff: Re: ecb and jde 2.3.4


>
> What's the best version of ecb to work with jde 2.3.4 ?
>
> I'm upgrading jde from version 2.3.4beta3, so I've had to install
> cedet1.0beta2.  With that version of cedet, ECB fails when
> building the methods buffer:
>
> tree-node-set-expanded: Wrong type argument: arrayp, nil
>
> Is there some version of ecb that works fine with that versions of jde
> and cedet?
>
> Thanks in advance
>

this may be a case of the speedbar.elc that gets installed with emacs conflicting with
the one in CEDET [?] you might try just moving the one under emacs to .elc_OLD
and restarting..


--

michael schupp
+44.[0].20.754.50621




--

This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.