Re: xml-mode (from sgml-mode) tag skipping fix

Re: xml-mode (from sgml-mode) tag skipping fix

[Stefan Monnier]

> The included patch allows `xml-mode' from sgml-mode.el to understand
> how to skip around XML-style empty tags:
> 
> Example:
> 
> <?xml version="1.0"?>
> <foo>
>   <bar>
>     <foo/>
>     <bar/>
>   </bar>
> </foo>
> 
> `C-c C-f' before the first <bar> should go to the end of the </bar>.
> `C-c C-b' after the last </foo> should go to the beginning of the
> first <foo>.

Thanks.  I believe that SGML should handle that as well (although
I'm not quite sure what <foo/> means in SGML) so I think we
could/should use the patch below instead.

Does anybody know better ?


	Stefan


--- sgml-mode.el.~1.91.~	Wed Oct 30 20:55:54 2002
+++ sgml-mode.el	Thu Oct 31 15:47:20 2002
@@ -691,43 +691,46 @@
     (if (looking-at "</\\([^ \n\t>]+\\)")
 	;; end tag, skip any nested pairs
 	(let ((case-fold-search t)
-	      (re (concat "</?" (regexp-quote (match-string 1)))))
+	      (re (concat "</?" (regexp-quote (match-string 1))
+			  ;; Ignore empty tags like <foo/>.
+			  "\\([^>]*[^/]\\)?>")))
 	  (while (and (re-search-backward re nil t)
 		      (eq (char-after (1+ (point))) ?/))
 	    (forward-char 1)
 	    (sgml-skip-tag-backward 1))))
     (setq arg (1- arg))))
 
-(defun sgml-skip-tag-forward (arg &optional return)
+(defun sgml-skip-tag-forward (arg)
   "Skip to end of tag or matching closing tag if present.
 With prefix argument ARG, repeat this ARG times.
 Return t iff after a closing tag."
   (interactive "p")
-  (setq return t)
+  (let ((return t))
   (while (>= arg 1)
     (skip-chars-forward "^<>")
     (if (eq (following-char) ?>)
 	(up-list -1))
-    (if (looking-at "<\\([^/ \n\t>]+\\)")
+      (if (looking-at "<\\([^/ \n\t>]+\\)\\([^>]*[^/]\\)?>")
 	;; start tag, skip any nested same pairs _and_ closing tag
 	(let ((case-fold-search t)
-	      (re (concat "</?" (regexp-quote (match-string 1))))
+		(re (concat "</?" (regexp-quote (match-string 1))
+			    ;; Ignore empty tags like <foo/>.
+			    "\\([^>]*[^/]\\)?>"))
 	      point close)
 	  (forward-list 1)
 	  (setq point (point))
 	  (while (and (re-search-forward re nil t)
 		      (not (setq close
 				 (eq (char-after (1+ (match-beginning 0))) ?/)))
-		      (not (up-list -1))
+			(goto-char (match-beginning 0))
 		      (sgml-skip-tag-forward 1))
 	    (setq close nil))
-	  (if close
-	      (up-list 1)
+	    (unless close
 	    (goto-char point)
-	    (setq return)))
+	      (setq return nil)))
       (forward-list 1))
     (setq arg (1- arg)))
-  return)
+    return))
 
 (defun sgml-delete-tag (arg)
   "Delete tag on or after cursor, and matching closing or opening tag.

Re: xml-mode (from sgml-mode) tag skipping fix

[md5i]

"Stefan Monnier" <monnier+gnu/emacs/pretest@rum.cs.yale.edu> writes:

> Thanks.  I believe that SGML should handle that as well (although
> I'm not quite sure what <foo/> means in SGML) so I think we
> could/should use the patch below instead.
>
> Does anybody know better ?

As far as I am concerned, you can change it this way.  I've looked at
the SGML standard, and it doesn't seem to allow tags of the type
<foo/> anyway.  It does allow tage of the form

<phone-number/555-5555/

but sgml-mode.el doesn't handle these, so I think we're fine.  Seems
like a real wart anyway.

-- 
Michael Welsh Duggan
(md5i@cs.cmu.edu)

Re: xml-mode (from sgml-mode) tag skipping fix

[Karl Eichwalder]

"Stefan Monnier" <monnier+gnu/emacs/pretest@rum.cs.yale.edu> writes:

> (although I'm not quite sure what <foo/> means in SGML)

   Linkname: Web SGML Adaptations Extension                                           
        URL: http://www.sgml.u-net.com/book/sgml-4.htm#Extensions                     

   Web SGML Adaptations Extension                                                 
   When the Web SGML adaptations provided by Annex K of SGML are available an         
   optional net-enabling start-tag close (NESTC) delimiter can be defined in additon  
   to the null-end tag delimiter. When this delimiter has been defined, the start-tag 
   of an element to be ended with a null end-tag must end with the NESTC delimiter.   
                                                                                      
   Note: If no NESTC delimiter is defined it is assumed to have the same definition   
   as the NET delimiter                                                               

The Reference Concrete Syntax used by DocBook, HTML, etc. does not
define a NESTC.  Check the SGML declaration (.dcl or .decl) coming with
DTD; xml.dcl features the following definition:

         DELIM
             GENERAL SGMLREF
             HCRO "&#38;#x" -- 38 is the number for ampersand --
             NESTC "/"
             NET ">"
             PIC "?>"
             SHORTREF NONE

The Reference Concrete Syntax just says:

         DELIM
             GENERAL SGMLREF
             SHORTREF SGMLREF


Note: Annex K "was approved by ISO's SGML technical committee on 5th
December 1997" (Martin Bryan 1997, http://www.sgml.u-net.com/book/ ).

Note 2: As it stands, XML is SGML.

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.gnu.franken.de/ke/                            |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)

Re: xml-mode (from sgml-mode) tag skipping fix

[md5i]

"Stefan Monnier" <monnier+gnu/emacs/pretest@rum.cs.yale.edu> writes:

> Thanks.  I believe that SGML should handle that as well (although
> I'm not quite sure what <foo/> means in SGML) so I think we
> could/should use the patch below instead.

Small fix.  The "\\([^>]*[^/]\\)?>" references should probably read:
"\\([^>]*[^/>]\\)?>" instead.

-- 
Michael Welsh Duggan
(md5i@cs.cmu.edu)

Re: xml-mode (from sgml-mode) tag skipping fix

[Stefan Monnier]

> "Stefan Monnier" <monnier+gnu/emacs/pretest@rum.cs.yale.edu> writes:
> 
> > (although I'm not quite sure what <foo/> means in SGML)
> 
>    Linkname: Web SGML Adaptations Extension                                           
>         URL: http://www.sgml.u-net.com/book/sgml-4.htm#Extensions                     
> 
>    Web SGML Adaptations Extension                                                 
>    When the Web SGML adaptations provided by Annex K of SGML are available an         
>    optional net-enabling start-tag close (NESTC) delimiter can be defined in additon  
>    to the null-end tag delimiter. When this delimiter has been defined, the start-tag 
>    of an element to be ended with a null end-tag must end with the NESTC delimiter.   
>                                                                                       
>    Note: If no NESTC delimiter is defined it is assumed to have the same definition   
>    as the NET delimiter                                                               
> 
> The Reference Concrete Syntax used by DocBook, HTML, etc. does not
> define a NESTC.  Check the SGML declaration (.dcl or .decl) coming with
> DTD; xml.dcl features the following definition:
> 
>          DELIM
>              GENERAL SGMLREF
>              HCRO "&#38;#x" -- 38 is the number for ampersand --
>              NESTC "/"
>              NET ">"
>              PIC "?>"
>              SHORTREF NONE
> 
> The Reference Concrete Syntax just says:
> 
>          DELIM
>              GENERAL SGMLREF
>              SHORTREF SGMLREF
> 
> 
> Note: Annex K "was approved by ISO's SGML technical committee on 5th
> December 1997" (Martin Bryan 1997, http://www.sgml.u-net.com/book/ ).
> 
> Note 2: As it stands, XML is SGML.

Can anybody translate it into English for me ?


	Stefan

Re: xml-mode (from sgml-mode) tag skipping fix

[md5i]

"Stefan Monnier" <monnier+gnu/emacs/pretest@rum.cs.yale.edu> writes:

>> Note 2: As it stands, XML is SGML.
>
> Can anybody translate it into English for me ?

Reading through it, and perusing the included web site, it seems to
say that SGML can be set up to handle self-closing tags like <foo/>.

-- 
Michael Welsh Duggan
(md5i@cs.cmu.edu)

Re: xml-mode (from sgml-mode) tag skipping fix

[Karl Eichwalder]

"Stefan Monnier" <monnier+gnu/emacs/pretest@rum.cs.yale.edu> writes:

>> Note: Annex K "was approved by ISO's SGML technical committee on 5th
>> December 1997" (Martin Bryan 1997, http://www.sgml.u-net.com/book/ ).
>> 
>> Note 2: As it stands, XML is SGML.
>
> Can anybody translate it into English for me ?

Last time I tried to explain the mysteries about SGML declarations and
you asked for a web reference.  This time I posted such a reference:
http://www.sgml.u-net.com/book/ -- Martin Bryan's book is worth reading.

SGML documents can look very different, depending on the SGML
declaration in use.

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.gnu.franken.de/ke/                            |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)

Re: xml-mode (from sgml-mode) tag skipping fix

[Stefan Monnier]

> >> Note 2: As it stands, XML is SGML.
> >
> > Can anybody translate it into English for me ?
> 
> Reading through it, and perusing the included web site, it seems to
> say that SGML can be set up to handle self-closing tags like <foo/>.

Thanks.  I thus installed the patch I suggested with the little
fixup of [^/] -> [^/>] you just sent.


	Stefan

Re: xml-mode (from sgml-mode) tag skipping fix

[Stefan Monnier]

> >> Note: Annex K "was approved by ISO's SGML technical committee on 5th
> >> December 1997" (Martin Bryan 1997, http://www.sgml.u-net.com/book/ ).
> >> 
> >> Note 2: As it stands, XML is SGML.
> >
> > Can anybody translate it into English for me ?
> 
> Last time I tried to explain the mysteries about SGML declarations and
> you asked for a web reference.  This time I posted such a reference:
> http://www.sgml.u-net.com/book/ -- Martin Bryan's book is worth reading.
> 
> SGML documents can look very different, depending on the SGML
> declaration in use.

Sure.  I just wanted to know whether it's safe for sgml-mode
to treat <foo/> in the same way as xml-mode does.
I know that tags can use something else than < and >, but sgml-mode
doesn't try to handle those cases.
I.e. my question didn't have much to do with the SGML definition
but more with the way SGML is usually used.


	Stefan

Re: xml-mode (from sgml-mode) tag skipping fix

[Karl Eichwalder]

"Stefan Monnier" <monnier+gnu/emacs/pretest@rum.cs.yale.edu> writes:

> I.e. my question didn't have much to do with the SGML definition
> but more with the way SGML is usually used.

By definition, <foo/> is neither allowed for HTML documents
(cf. html.d[e]?cl) nor for DocBook documents (docbook.dcl).  Thus it's
usually not used.

-- 
ke@suse.de (work) / keichwa@gmx.net (home):              |
http://www.gnu.franken.de/ke/                            |      ,__o
Free Translation Project:                                |    _-\_<,
http://www.iro.umontreal.ca/contrib/po/HTML/             |   (*)/'(*)