From nikodemus at random-state.net  Tue May 19 09:33:10 2009
From: nikodemus at random-state.net (Nikodemus Siivola)
Date: Tue, 19 May 2009 12:33:10 +0300
Subject: [climacs-devel] M-q issues: undo doesn't work after,
	DREI-BUFFER:OFFSET-AFTER-END 	during
Message-ID: <633d72b0905190233k3998b2f3qaa4d619bc5e40251@mail.gmail.com>

in foo.lisp:

(defun foo ()
  "Line one.

Line two."
  (bar))

hit M-q on the start of `Line two.' in the docstring. Result is

(defun foo ()
  "Line one.  Line Two."
  (bar))

C-_ move the `Line two.' part back and forth, but does not restore the
original before running out of undo.

Add following forms in the the same file after FOO, and try the
reindent again: Climacs lands in the debugger with OFFSET-AFTER-END
error.

(defmacro qx-macro (type n)
   `(let ((n ,n))
      (declare (,type n))
      (decf n)))

 (defun qx (a)
   (declare (optimize speed))
   (qx-macro fixnum a))

Cheers,

 -- Nikodemus



From strandh at labri.fr  Sun May 31 06:21:43 2009
From: strandh at labri.fr (Robert Strandh)
Date: Sun, 31 May 2009 08:21:43 +0200
Subject: [climacs-devel]  M-q issues: undo doesn't work after,
	DREI-BUFFER:OFFSET-AFTER-END 	during
In-Reply-To: <633d72b0905190233k3998b2f3qaa4d619bc5e40251@mail.gmail.com>
References: <633d72b0905190233k3998b2f3qaa4d619bc5e40251@mail.gmail.com>
Message-ID: <18978.8695.754509.722768@serveur5.labri.fr>

Hello,

Nikodemus Siivola writes:
 > in foo.lisp:
 > 
 > (defun foo ()
 >   "Line one.
 > 
 > Line two."
 >   (bar))
 > 
 > hit M-q on the start of `Line two.' in the docstring. Result is
 > 
 > (defun foo ()
 >   "Line one.  Line Two."
 >   (bar))
 > 
 > C-_ move the `Line two.' part back and forth, but does not restore the
 > original before running out of undo.

Thanks for reporting this problem.

This appears to be because (setf buffer-object) doesn't create an undo
record.  I suspect the thing to do would be to create a new kind of
output record, CHANGE-RECORD (in addition to INSERT-RECORD and
DELETE-RECORD), to add all the proper methods for this new type of
record, and to add a :before method on (SETF BUFFER-OBJECT) that
creates such an output record.

-- 
Robert Strandh

---------------------------------------------------------------------
Greenspun's Tenth Rule of Programming: any sufficiently complicated C
or Fortran program contains an ad hoc informally-specified bug-ridden
slow implementation of half of Common Lisp.
---------------------------------------------------------------------



From strandh at labri.fr  Sun May 31 07:50:27 2009
From: strandh at labri.fr (Robert Strandh)
Date: Sun, 31 May 2009 09:50:27 +0200
Subject: [climacs-devel]  M-q issues: undo doesn't work after,
	DREI-BUFFER:OFFSET-AFTER-END 	during
In-Reply-To: <633d72b0905190233k3998b2f3qaa4d619bc5e40251@mail.gmail.com>
References: <633d72b0905190233k3998b2f3qaa4d619bc5e40251@mail.gmail.com>
Message-ID: <18978.14019.72839.141821@serveur5.labri.fr>

Nikodemus Siivola writes:
 > Add following forms in the the same file after FOO, and try the
 > reindent again: Climacs lands in the debugger with OFFSET-AFTER-END
 > error.

Indeed.  This seems to be an unrelated problem.  I don't know (yet)
what causes it.

Thanks for reporting it. 
-- 
Robert Strandh

---------------------------------------------------------------------
Greenspun's Tenth Rule of Programming: any sufficiently complicated C
or Fortran program contains an ad hoc informally-specified bug-ridden
slow implementation of half of Common Lisp.
---------------------------------------------------------------------