table.el downcase HTML tags

table.el downcase HTML tags

[Tak Ota]

This has been requested for many times that HTML tags that table.el
generates to be downcased to be compatible with XHTML standard.  If no
one objects could someone check the following change into the
repository.  Thanks.

-Tak

2006-05-30  Takaaki Ota  <Takaaki.Ota@am.sony.com>

	* textmodes/table.el: Convert all HTML tags to lower case for
	XHTML compatibility.


bash-3.1$ diff ../../../../d/pub/emacs/pure/emacs/lisp/textmodes/table.el table.el
9c9
< ;; Revised: Sat Aug 06 2005 19:42:54 (CEST)
---
> ;; Revised: Tue May 30 2006 10:01:43 (PDT)
3107c3107
< 	      (format "<TABLE %s>\n" table-html-table-attribute)
---
> 	      (format "<table %s>\n" table-html-table-attribute)
3110c3110
< 		  (format "  <CAPTION>%s</CAPTION>\n" caption)
---
> 		  (format "  <caption>%s</caption>\n" caption)
3134c3134
<       (insert "</TABLE>\n"))
---
>       (insert "</table>\n"))
3155c3155
< 	(insert "  <TR>\n"))
---
> 	(insert "  <tr>\n"))
3163c3163
< 	(insert "  </TR>\n"))
---
> 	(insert "  </tr>\n"))
3210c3210
< 				   "TH" "TD"))))
---
> 				   "th" "td"))))
3269c3269
< 	    (insert "<BR />")
---
> 	    (insert "<br />")

Re: table.el downcase HTML tags

[Mathias Dahl]

> This has been requested for many times that HTML tags that table.el
> generates to be downcased to be compatible with XHTML standard.  If no
> one objects could someone check the following change into the
> repository.  Thanks.
>
> <             (format "<TABLE %s>\n" table-html-table-attribute)

Why not making it configurable? I don't care about the case in this
case (haha!) but I am sure some do.

Re: table.el downcase HTML tags

[Lennart Borgman]

Mathias Dahl wrote:
>> This has been requested for many times that HTML tags that table.el
>> generates to be downcased to be compatible with XHTML standard.  If no
>> one objects could someone check the following change into the
>> repository.  Thanks.
>>
>> <             (format "<TABLE %s>\n" table-html-table-attribute)
>
> Why not making it configurable? I don't care about the case in this
> case (haha!) but I am sure some do.
I do not know of any case where uppercase tags is useful. Are there 
really any such cases?

Re: table.el downcase HTML tags

[Ted Zlatanov]

On 30 May 2006, lennart.borgman.073@student.lu.se wrote:

> I do not know of any case where uppercase tags is useful. Are there
> really any such cases?

Old, badly written parsers may look for "TABLE" and break with
"table".  Generally, why break old behavior if it can be left in as an
option?

For table.el, why not define something like this:

(defcustom table-uppercase-html-tags nil
  "*All generated HTML tags are in uppercase when this is set."
  :tag "Uppercase HTML tags"
  :type 'boolean
  :group 'table)

(defun table--generate-source-html-tag (tag-symbol text &optional options)
  (let ((tag-symbol (if table-uppercase-html-tags
		       (upcase tag-symbol)
		     tag-symbol)))
    (format "<%s %s> %s </%s>" tag-symbol options text tag-symbol)))

This would also simplify a lot of the table.el code that generates HTML :)

Ted

Re: table.el downcase HTML tags

[Kevin Rodgers]

Lennart Borgman wrote:
> Mathias Dahl wrote:
>>> This has been requested for many times that HTML tags that table.el
>>> generates to be downcased to be compatible with XHTML standard.  If no
>>> one objects could someone check the following change into the
>>> repository.  Thanks.
>>>
>>> <             (format "<TABLE %s>\n" table-html-table-attribute)
>>
>> Why not making it configurable? I don't care about the case in this
>> case (haha!) but I am sure some do.
> I do not know of any case where uppercase tags is useful. Are there 
> really any such cases?

No, because HTML is a case-insensitive SGML application:

See http://www.w3.org/TR/html4/

	HTML 4 is an SGML application conforming to International
	Standard ISO 8879 -- Standard Generalized Markup Language
	[ISO8879].

See http://www.w3.org/TR/html4/intro/sgmltut.html#h-3.2.1

	Element names are always case-insensitive.

See http://www.w3.org/TR/html4/sgml/sgmldecl.html

	                  NAMECASE GENERAL YES
	                           ENTITY  NO

Note however, that clause 9.3 of ISO 8879 states that

	The upper-case form of each character in a name, name token,
	number, or number token, as specified by the "NAMECASE"
	parameter of the SGML declaration, is substituted for the
	character actually entered.

so the uppercase names are actually canonical in HTML, whereas the
lowercase names are mandatory in XHTML which is a (case-sensitive)
XML application.

-- 
Kevin

Re: table.el downcase HTML tags

[Stuart D. Herring]

[Sorry if this is a resend; mail client broke.]

> (defun table--generate-source-html-tag (tag-symbol text &optional options)
>   (let ((tag-symbol (if table-uppercase-html-tags
> 		       (upcase tag-symbol)
> 		     tag-symbol)))
>     (format "<%s %s> %s </%s>" tag-symbol options text tag-symbol)))
>
> This would also simplify a lot of the table.el code that generates HTML :)

The resulting HTML would be ugly and somewhat incorrect:

(table--generate-source-html-tag "a" "")
 => "<a nil>  </a>"

I'd use

(defun table--generate-source-html-tag (tag &optional body attrs)
  (when table-uppercase-html-tags (setq tag (upcase tag)))
  (concat "<" tag (if attrs " ") attrs
          (if body (concat ">" body "</" tag) "/") ">"))

giving

(table--generate-source-html-tag "a" "b" "c=d")
 => "<a c=d>b</a>"
(table--generate-source-html-tag "a")
 => "<a/>"

I removed the spaces around the body because they could break things and
are trivial to include in the argument.  I'd really be tempted to make
attrs an alist, but that might be overengineering the problem.  The
XML-style empty tags might be undesirable, but they're optional:

(table--generate-source-html-tag "a" "")
 => "<a></a>"

Hope it helps,
Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.

Re: table.el downcase HTML tags

[Lennart Borgman]

Stuart D. Herring wrote:
> I'd use
>
> (defun table--generate-source-html-tag (tag &optional body attrs)
>   (when table-uppercase-html-tags (setq tag (upcase tag)))
>   (concat "<" tag (if attrs " ") attrs
>           (if body (concat ">" body "</" tag) "/") ">"))
>
> giving
>
> (table--generate-source-html-tag "a" "b" "c=d")
>  => "<a c=d>b</a>"
> (table--generate-source-html-tag "a")
>  => "<a/>"
>
> I removed the spaces around the body because they could break things and
> are trivial to include in the argument.  I'd really be tempted to make
> attrs an alist, but that might be overengineering the problem.  The
> XML-style empty tags might be undesirable, but they're optional:
>   

Option values should be surrounded by "". An alist should be better to 
avoid problems with this.

Re: table.el downcase HTML tags

[Kim F. Storm]

Ted Zlatanov <tzz@lifelogs.com> writes:

> On 30 May 2006, lennart.borgman.073@student.lu.se wrote:
>
> Old, badly written parsers may look for "TABLE" and break with
> "table".  

This is not a reason for us to output badly written HTML...

>           Generally, why break old behavior if it can be left in as an
> option?

table.el is new in emacs 22.x, so we still are free to change it,
without considering what those changes might break.

If Tak thinks this change is good, he's the one to judge it.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: table.el downcase HTML tags

[Kim F. Storm]

Tak Ota <Takaaki.Ota@am.sony.com> writes:

> This has been requested for many times that HTML tags that table.el
> generates to be downcased to be compatible with XHTML standard.  If no
> one objects could someone check the following change into the
> repository.  Thanks.

Done.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: table.el downcase HTML tags

[Carsten Dominik]

No,  this should not be configurable.  What a waste of resources.

- Carsten

On May 30, 2006, at 21:23, Mathias Dahl wrote:

>> This has been requested for many times that HTML tags that table.el
>> generates to be downcased to be compatible with XHTML standard.  If no
>> one objects could someone check the following change into the
>> repository.  Thanks.
>>
>> <             (format "<TABLE %s>\n" table-html-table-attribute)
>
> Why not making it configurable? I don't care about the case in this
> case (haha!) but I am sure some do.
>
>

--
Carsten Dominik
Sterrenkundig Instituut "Anton Pannekoek"
Universiteit van Amsterdam
Kruislaan 403
NL-1098SJ Amsterdam
phone: +31 20 525 7477

Re: table.el downcase HTML tags

[Ted Zlatanov]

On 31 May 2006, storm@cua.dk wrote:

> Ted Zlatanov <tzz@lifelogs.com> writes:
>
>> On 30 May 2006, lennart.borgman.073@student.lu.se wrote:
>>
>> Old, badly written parsers may look for "TABLE" and break with
>> "table".  
>
> This is not a reason for us to output badly written HTML...

I'm sure it's not badly written, in fact (as pointed out) it's "more"
valid to output uppercase tags with HTML 4.0 and earlier.

I simply offered an option to preserve the old behavior, which also
(IMHO) would simplify the code.

>> Generally, why break old behavior if it can be left in as an
>> option?
>
> table.el is new in emacs 22.x, so we still are free to change it,
> without considering what those changes might break.
>
> If Tak thinks this change is good, he's the one to judge it.

Sure.

Ted

Re: table.el downcase HTML tags

[Ted Zlatanov]

On 30 May 2006, herring@lanl.gov wrote:

> I'd use
>
> (defun table--generate-source-html-tag (tag &optional body attrs)
> (when table-uppercase-html-tags (setq tag (upcase tag)))
> (concat "<" tag (if attrs " ") attrs
> (if body (concat ">" body "</" tag) "/") ">"))
>
> giving
>
> (table--generate-source-html-tag "a" "b" "c=d")
> => "<a c=d>b</a>"
> (table--generate-source-html-tag "a")
> => "<a/>"

Sure, that works :)

> I removed the spaces around the body because they could break things and
> are trivial to include in the argument.  I'd really be tempted to make
> attrs an alist, but that might be overengineering the problem.  The
> XML-style empty tags might be undesirable, but they're optional:
>
> (table--generate-source-html-tag "a" "")
> => "<a></a>"

Actually all this and more is covered by sgml-mode's sgml-tag (and
friends).  Perhaps they should be used here instead of reinventing the
wheel.

Ted

Re: table.el downcase HTML tags

[Kim F. Storm]

Ted Zlatanov <tzz@lifelogs.com> writes:

> I simply offered an option to preserve the old behavior, which also
> (IMHO) would simplify the code.

But table.el already used a mixture of lower and uppercase tags, so
making all of them lowercase does not make things worse.

A more thorough cleanup (if needed), should wait until after the release.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk