Problem positioning cursor

Problem positioning cursor

[Victor Kirk]

Hi,

I've wrote a function to insert a skeleton for a java try-catch
block, but I've run into a problem leaving the cursor in the
correct place.

If there is no active region I want the following code to be
inserted (note X marks the spot I want to leave the cursor at.

    try {
        X
    } catch(Exception e) {
        e.printStackTrace();
    } 

This works fine.

If there is code selected I want:

    try {
        someCode(toBeWrapped);
    } catch(Exception e) {
        e.printStackTrace();
        X
    } 

unfortunatly the cursor is left at the `n' in Exception.

If anyone has the time to point me in the right direction
it would be appreciated.


(defun vics-java-catch-insert(&optional has-finally)
  "Generate a skeleton for a java try-catch block. If the optional
has-finally is true then a finally block is also inserted."
  (interactive)
  (let ((ex-type (read-from-minibuffer "Exception Type: " "Exception" nil
nil nil nil nil))
		(try-start)    ;; start of try catch block
		(try-end)      ;; end of 
		(try-body nil) ;; code to insert within try/catch
		(edit-point))  ;; point to leave user at
	;; If a region is selected save this to try-body
	(if (c-region-is-active-p)
		(setq try-body (delete-and-extract-region (mark) (point))))
	(setq try-start (point))
	(insert "try {\n")
	(if (not try-body)
		(setq edit-point (point))
	  (insert try-body))
	(insert (format "\n} catch (%s e) {\n" ex-type))
	(insert "e.printStackTrace();\n")
	;; if we had some text leave save location at the end of the catch
	;; block
	(if (eq nil (not try-body))
		(setq edit-point (point)))
	(insert "}")	;
	(if (eq nil (not has-finally))
		(insert " finally {\n\n}"))
	(setq try-end (point))
	(indent-region try-start try-end nil)
	(goto-char edit-point)
	(c-indent-command)))



Vic
--
Victor Kirk
Analyst
Serco Integrated Transport
--

This message, including attachments, is intended only for the use by the
person(s) to whom it is addressed. It may contain information which is
privileged and confidential. Copying or use by anybody else is not
authorised. If you are not the intended recipient, please contact the sender
as soon as possible. The views expressed in this communication may not
necessarily be the views held by Serco Integrated Transport.

Re: Problem positioning cursor

[Kai Großjohann]

Victor Kirk <Victor.Kirk@serco.com> writes:

> I've wrote a function to insert a skeleton for a java try-catch
> block, but I've run into a problem leaving the cursor in the
> correct place.

First of all, maybe you want to use skeletons rather than hand-coding
everything.  It will be easier to understand.

I think that skeletons can't do what you need, though.  There is a
change that adds the capability you need which has been committed to
the development release of Emacs in the last couple of days, or which
will be committed soon.

However, maybe you want to read up on skeletons anyway -- they might
come in useful for other things you are doing.

Failing the use of skeletons, a technique I suggest to use for the
hand-written functions is save-excursion.  For example, here is
something that inserts foo() and leaves point inside the parentheses:

    (insert "foo(")
    (save-excursion
      (insert ")")

I think you will find this much easier to understand that your
hand-coded remembering of positions, and thus it will be easier to
find the bugs.
-- 
file-error; Data: (Opening input file no such file or directory ~/.signature)

Re: Problem positioning cursor

[Oliver Scholz]

kai.grossjohann@gmx.net (Kai Großjohann) writes:
[...]
> I think that skeletons can't do what you need, though.  There is a
> change that adds the capability you need which has been committed to
> the development release of Emacs in the last couple of days, or which
> will be committed soon.

It should be possible even in Emacs 21.2. If you use something like
("lirum" _ "larum" _ "schubi"), then point is put at the first "_",
unless that is consumed (so to say) for the region.

Hmmm ...

(define-skeleton vics-java-catch-insert
  "Generate a skeleton for a java try-catch block."
  nil
  "try {" \n
  > _  \n
  > "} catch(Exception e) {" \n
  > "e.printStackTrace();" \n
  > _ 
  > "}" \n)


    Oliver
-- 
7 Floréal an 211 de la Révolution
Liberté, Egalité, Fraternité!

Re: Problem positioning cursor

[Kai Großjohann]

Oliver Scholz <alkibiades@gmx.de> writes:

> It should be possible even in Emacs 21.2. If you use something like
> ("lirum" _ "larum" _ "schubi"), then point is put at the first "_",
> unless that is consumed (so to say) for the region.

But the OP wants point to end up at the _second_ spot.

Hm.

Oh, are you saying that, if the region is active, then it gets put at
the first spot, and point at the second?  That'd work, then.
-- 
file-error; Data: (Opening input file no such file or directory ~/.signature)

Re: Problem positioning cursor

[Oliver Scholz]

kai.grossjohann@gmx.net (Kai Großjohann) writes:
[...]
> Oh, are you saying that, if the region is active, then it gets put at
> the first spot, and point at the second?  That'd work, then.
[...]

Yes.

    Oliver
-- 
8 Floréal an 211 de la Révolution
Liberté, Egalité, Fraternité!