address@hidden: tcl.el tcl-end-of-defun failure]

[Richard Stallman]

I don't know TCL so I have no idea if this has a hidden pitfall or not.
Could someone who does know TCL report on whether it is the right fix?

------- Start of forwarded message -------
Envelope-to: address@hidden
Delivery-date: Tue, 19 Nov 2002 12:19:14 -0500
X-Authentication-Warning: kuiper.rlent.pnet: roland set sender to 
address@hidden using -f
Sender: address@hidden
To: address@hidden
Subject: tcl.el tcl-end-of-defun failure
From: Roland Roberts <address@hidden>
Date: 19 Nov 2002 12:19:04 -0500
X-Spam-Status: No, hits=-0.9 required=5.0
        tests=PATCH_UNIFIED_DIFF,SPAM_PHRASE_00_01,X_AUTH_WARNING
        version=2.41
X-Spam-Level: 

- --=-=-=

In 21.2.92, tcl.el uses a value for defun-prompt-regexp which fails to
correctly skip of Tcl proc argument lists which have default values,
e.g.,

proc foo {arg1 {arg2 "arg2_default"}} {
   puts "hello, world"
}


This patch fixes that problem and seems to still work fine for
non-proc cases e.g.,

mm arg {
   puts "hello, world"
}

Note that in all cases, tcl-end-of-defun does not work correctly if
the proc does not start in column 0.  E.g.,

namespace eval foo {

    proc bar {args} {
        puts $args
    }

}

NB: This patch includes the patches sent earlier to fix
comment-start-skip.  The only required change for this patch is to
tcl-omit-ws-regexp.

roland
- -- 
                       PGP Key ID: 66 BC 3B CD
Roland B. Roberts, PhD                             RL Enterprises
address@hidden                            6818 Madeline Court
address@hidden                           Brooklyn, NY 11220


- --=-=-=
Content-Type: text/x-patch
Content-Disposition: attachment; filename=tcl.el.diff
Content-Description: tcl.el tcl-end-of-defun fix

- --- /usr/local/share/emacs/21.2.92/lisp/progmodes/tcl.el      Fri Apr 12 
13:14:49 2002
+++ elisp/tcl.el        Tue Nov 19 11:33:07 2002
@@ -559,7 +559,7 @@
 ;; end-of-defun doesn't really use the brace matching the one that
 ;; trails defun-prompt-regexp.
 ;; ?? Is there a bug now ??
- -(defconst tcl-omit-ws-regexp "^[^ \t\n#}][^\n}]+}*[ \t]+")
+(defconst tcl-omit-ws-regexp "^[^ \t\n#][^\n]+}*[ \t]+")
 
 (defun tcl-internal-beginning-of-defun (&optional arg)
   "Move backward to next beginning of defun.
@@ -764,7 +764,7 @@
   (make-local-variable 'comment-start)
   (setq comment-start "# ")
   (make-local-variable 'comment-start-skip)
- -  (setq comment-start-skip "#+ *")
+  (setq comment-start-skip "\\(^\\s +#+\\s *\\|;#+\\s *\\)")
   (make-local-variable 'comment-column)
   (setq comment-column 40)
   (make-local-variable 'comment-end)
@@ -1874,7 +1874,7 @@
                (insert
                 (if (tcl-real-command-p) "" ";")
                 "# ")
- -             (backward-char))))))
+               (backward-char 2))))))
   ;; Point is just after the "#" starting a comment.  Move it as
   ;; appropriate.
   (let* ((indent (or (funcall comment-indent-function) comment-column))
@@ -1883,7 +1883,7 @@
                   (point))))
     (if (/= begpos indent)
        (progn
- -       (skip-chars-backward " \t" (save-excursion
+         (skip-chars-backward "; \t" (save-excursion
                                       (beginning-of-line)
                                       (point)))
          (delete-region (point) begpos)

- --=-=-=--
------- End of forwarded message -------