Debugging smalltalk-mode.el

Hi All,

In emacs the smalltalk-mode have a bug with the new syntax when writing a
keyword method signature.

foo: a [
        ...
]

When writing the colon you end up in a parse error:

Debugger entered--Lisp error: (scan-error "Unbalanced parentheses" 2479 2446)
  apply(debug error (scan-error "Unbalanced parentheses" 2479 2446))
  edebug(error (scan-error "Unbalanced parentheses" 2479 2446))
  signal(scan-error ("Unbalanced parentheses" 2479 2446))
  edebug-signal(scan-error ("Unbalanced parentheses" 2479 2446))
  scan-sexps(2480 -1)
  forward-sexp(-1)
  backward-sexp(1)
  smalltalk-backward-sexp(1)
  smalltalk-find-statement-begin()
  smalltalk-indent-for-colon()
  (edebug-after (edebug-before 37) 38 (smalltalk-indent-for-colon))
  (and (edebug-after 0 36 needs-indent) (edebug-after (edebug-before 37) 38
(smalltalk-indent-for-colon)))
  (edebug-after (edebug-before 35) 39 (and (edebug-after 0 36 needs-indent)
(edebug-after ... 38 ...)))
  (progn (edebug-after (edebug-before 16) 34 (save-excursion ... ...))
(edebug-after (edebug-before 35) 39 (and ... ...)))
  (edebug-after (edebug-before 15) 40 (progn (edebug-after ... 34 ...)
(edebug-after ... 39 ...)))
  (if (edebug-after (edebug-before 10) 14 (null ...)) (edebug-after
(edebug-before 15) 40 (progn ... ...)))
  (edebug-after (edebug-before 9) 41 (if (edebug-after ... 14 ...) (edebug-after
... 40 ...)))
  (let (needs-indent state (parse-sexp-ignore-comments t)) (edebug-after
(edebug-before 1) 8 (setq state ...)) (edebug-after (edebug-before 9) 41 (if ...
...)) (edebug-after (edebug-before 42) 43 (self-insert-command 1)))
  (edebug-after (edebug-before 0) 44 (let (needs-indent state ...) (edebug-after
... 8 ...) (edebug-after ... 41 ...) (edebug-after ... 43 ...)))
  (lambda nil (edebug-after (edebug-before 0) 44 (let ... ... ... ...)))()
  edebug-enter(smalltalk-colon nil (lambda nil (edebug-after (edebug-before 0)
44 (let ... ... ... ...))))
  edebug-enter(smalltalk-colon nil (lambda nil (edebug-after (edebug-before 0)
44 (let ... ... ... ...))))
  smalltalk-colon()


I think the issue is in #'smalltalk-indent-for-colon who should check if you are
being writing a method pattern or just sending a message.

I am trying to fix it but since I am not an emacs expert tell me if you have an
idea of how to fix it.

Thanks

Mathieu





_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

In fact to reproduce the bug you need to have a method before.
For example :

----Test.st-----
Object subclass: A [
    foo: f [
    ]

      bar: e [
   ]
]
-----Test.st-----

Only the colon of the bar method will throw the error.
This is because the #'smalltalk-narrow-to-method will narrow to:

-----------------
]
     
    bar
-----------------

So we need to check before if we are in a message pattern for the new syntax.





_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

On 07/23/2010 02:24 PM, Mathieu Suen wrote:
> This is because the #'smalltalk-narrow-to-method will narrow to:
>
> -----------------
> ]
>
>      bar
> -----------------
>
> So we need to check before if we are in a message pattern for the new syntax.

Or maybe the narrowing is wrong.  I'll look at it in the next few days
if you don't beat me...

Paolo

_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

----- Message d'origine ----

:)

The narrowing function is using  smalltalk-scope-begin-of-defun.
What is the purpose of  smalltalk-scope-begin-of-defun?

Is it to move point at the beginning of the message patern ?

Thanks

--Mathieu





_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

Yes.

Paolo

_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

Hi Paolo

I manage to fix the bug.
See the attached file.

You can also take it from my github repository:
git://github.com/mathk/smalltalk.git

Let me know if that is ok.

Thanks




_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

On 07/29/2010 10:42 AM, Mathieu Suen wrote:
> Hi Paolo
>
> I manage to fix the bug.
> See the attached file.
>
> You can also take it from my github repository:
> git://github.com/mathk/smalltalk.git
>
> Let me know if that is ok.

Applied, thanks.

Paolo

_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk