html2text

html2text

[Alfred M. Szmidt]

html2text is quite nice, but it doesn't strip all HTML files into
something that is readable.  The following patch makes it strip some
"newer" tags that have croped up.  Though, it still doesn't make
things as nice as they could be, tables and comments are still left
intact.

I guess that a better way to do this is to convert all known tags to
something nice, and then just strip all remaining tags that are left.

diff -ur html2text.el html2text.el.new
--- html2text.el	2004-10-31 19:23:06.000000000 +0100
+++ html2text.el.new	2004-10-31 19:23:46.000000000 +0100
@@ -75,8 +75,10 @@
 
 (defvar html2text-format-tag-list
   '(("b" 	  . html2text-clean-bold)
+    ("strong"     . html2text-clean-bold)
     ("u" 	  . html2text-clean-underline)
     ("i" 	  . html2text-clean-italic)
+    ("em"         . html2text-clean-italic)
     ("blockquote" . html2text-clean-blockquote)
     ("a"          . html2text-clean-anchor)
     ("ul"         . html2text-clean-ul)

Diff finished.  Sun Oct 31 19:23:56 2004

Re: html2text

[Reiner Steib]

On Sun, Oct 31 2004, Alfred M. Szmidt wrote:

> html2text is quite nice, but it doesn't strip all HTML files into
> something that is readable.  The following patch makes it strip some
> "newer" tags that have croped up.  Though, it still doesn't make
> things as nice as they could be, tables and comments are still left
> intact.
>
> I guess that a better way to do this is to convert all known tags to
> something nice, and then just strip all remaining tags that are left.

Would you like to work on this?

> diff -ur html2text.el html2text.el.new
> --- html2text.el	2004-10-31 19:23:06.000000000 +0100
> +++ html2text.el.new	2004-10-31 19:23:46.000000000 +0100
> @@ -75,8 +75,10 @@
>
>  (defvar html2text-format-tag-list
>    '(("b" 	  . html2text-clean-bold)
> +    ("strong"     . html2text-clean-bold)
>      ("u" 	  . html2text-clean-underline)
>      ("i" 	  . html2text-clean-italic)
> +    ("em"         . html2text-clean-italic)
>      ("blockquote" . html2text-clean-blockquote)
>      ("a"          . html2text-clean-anchor)
>      ("ul"         . html2text-clean-ul)

Committed in Gnus repository (will be synced to Emacs within a couple
of days).

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: html2text

[Alfred M. Szmidt]

   > I guess that a better way to do this is to convert all known tags
   > to something nice, and then just strip all remaining tags that
   > are left.

   Would you like to work on this?

Yes I would, but I have other priorities that are far more important
to me so it won't happen any time soon.

   Committed in Gnus repository (will be synced to Emacs within a
   couple of days).

Thanks.

Re: html2text

[Katsumi Yamaoka]

Hi,

What's the `tag' argument used for?

(defun html2text-get-attr (p1 p2 tag)

[...]

(defun html2text-clean-anchor (p1 p2 p3 p4)
  (let* ((attr-list (html2text-get-attr p1 p2 "a"))

In relation to this; is it necessary to bind the following
variable?

(defun html2text-format-tags ()
[...]
        (let ((p1)
              (p2 (point))
              (p3) (p4)
              (attr (match-string 0)))
               ^^^^

Re: html2text

[Reiner Steib]

On Tue, Nov 02 2004, Katsumi Yamaoka wrote:

> What's the `tag' argument used for?
> (defun html2text-get-attr (p1 p2 tag)

For nothing, AFAICS.

[ other odd code / unused variables ]

I only fixed _some_ odd looking places in the code.  I didn't go thru
the code systematically.  Please fix it if you find more.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: html2text

[Katsumi Yamaoka]

>>>>> In <v9lldk4ncs.fsf@marauder.physik.uni-ulm.de> Reiner Steib wrote:

>> What's the `tag' argument used for?
>> (defun html2text-get-attr (p1 p2 tag)

> For nothing, AFAICS.

> [ other odd code / unused variables ]

> I only fixed _some_ odd looking places in the code.  I didn't go thru
> the code systematically.

Sorry, I misunderstood as you've renewed that function because I
couldn't find any logs except for the following:

	(html2text-get-attr, html2text-fix-paragraph): Simplify code.

> Please fix it if you find more.

I've done.  Thanks.

Re: html2text

[Reiner Steib]

On Tue, Nov 02 2004, Katsumi Yamaoka wrote:

>>>>>> In <v9lldk4ncs.fsf@marauder.physik.uni-ulm.de> Reiner Steib wrote:
[...]
>> I only fixed _some_ odd looking places in the code.  I didn't go thru
>> the code systematically.
>
> Sorry, I misunderstood as you've renewed that function because I
> couldn't find any logs except for the following:
>
> 	(html2text-get-attr, html2text-fix-paragraph): Simplify code.

Ah, you probably looked at the version in Gnus trunk.  It seems
something went wrong when Miles merged the changed from v5-10 (where I
installed the changes), see ...

cvs diff -r7.7 -r7.8 lisp/html2text.el

or ...

http://quimby.gnus.org/cgi-bin/cvsweb.cgi/gnus/lisp/html2text.el.diff?r1=7.7&r2=7.8

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: html2text

[Reiner Steib]

On Sat, Nov 06 2004, Jari Aalto+mail.emacs wrote:

> This is your copy. Article has been posted to the newsgroup(s).

I didn't see your message on emacs-devel, see
<URL:http://thread.gmane.org/1099247139.071920.12084.nullmailer@Update.UU.SE>.

> * Sun 2004-10-31 Alfred Szmidt <ams AT kemisten.nu> gmane.emacs.devel
> * Message-Id: 1099247139.071920.12084.nullmailer AT Update.UU.SE
> | html2text is quite nice, but it doesn't strip all HTML files into
> | something that is readable.  The following patch makes it strip some
> | "newer" tags that have croped up.
>
> There is more entities. This patch is against the Gnus CVS, but I
> assume it will work for Emacs as well. The entities are in
> alphabetical order.
>
> 2004-11-06 Sat  Jari Aalto  <jari dot aalto A T cante dot net>
>
>         * text2html (html2text-replace-list). Added more HTML 4.0
>         entities.

It seems you have signed papers for Emacs as you are listed in the
AUTHORS file.  But I can't check it myself.  Could you please confirm?

[ The suggested patch from Jari's original message was: ]

--8<---------------cut here---------------start------------->8---
--- html2text.el.7.10	2004-11-06 17:20:46.000000000 +0200
+++ html2text.el	2004-11-06 17:41:12.000000000 +0200
@@ -42,8 +42,42 @@
 (defvar html2text-format-single-element-list '(("hr" . html2text-clean-hr)))

 (defvar html2text-replace-list
-  '(("&nbsp;" . " ") ("&gt;" . ">") ("&lt;" . "<") ("&quot;" . "\"")
-    ("&amp;" . "&") ("&apos;" . "'"))
+  '(("&acute;" . "`")
+    ("&amp;" . "&")
+    ("&apos;" . "'")
+    ("&brvbar;" . "|")
+    ("&cent;" . "c")
+    ("&circ;" . "^")
+    ("&copy;" . "(C)")
+    ("&curren;" . "¤")
+    ("&deg;" . "degree")
+    ("&divide;" . "/")
+    ("&euro;" . "e")
+    ("&frac12;" . "½")
+    ("&gt;" . ">")
+    ("&iquest;" . "?")
+    ("&laquo;" . "<<")
+    ("&ldquo" . "\"")
+    ("&lsaquo;" . "(")
+    ("&lsquo;" . "`")
+    ("&lt;" . "<")
+    ("&mdash;" . "--")
+    ("&nbsp;" . " ")
+    ("&ndash;" . "-")
+    ("&permil;" . "%%")
+    ("&plusmn;" . "+-")
+    ("&pound;" . "£")
+    ("&quot;" . "\"")
+    ("&raquo;" . ">>")
+    ("&rdquo" . "\"")
+    ("&reg;" . "(R)")
+    ("&rsaquo;" . ")")
+    ("&rsquo;" . "'")
+    ("&sect;" . "§")
+    ("&sup1;" . "^1")
+    ("&sup2;" . "^2")
+    ("&sup3;" . "^3")
+    ("&tilde;" . "~"))
   "The map of entity to text.
--8<---------------cut here---------------end--------------->8---

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: html2text

[David Kastrup]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Reiner Steib <reinersteib+gmane@imap.cc> writes:

> On Sat, Nov 06 2004, Jari Aalto+mail.emacs wrote:
>
> It seems you have signed papers for Emacs as you are listed in the
> AUTHORS file.  But I can't check it myself.  Could you please
> confirm?

copyright.list (privately accessible to GNU maintainers) lists Jari
with a general assignment of past and future changes to Emacs.

- -- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>

iD8DBQFBj7S2Bo350SLJfmgRAq3jAJ9b01k5renyp+1lwFuxasAGSMuW+ACeLJq+
YD4G1OUD17RECupVKH3SkGI=
=SyiB
-----END PGP SIGNATURE-----

Re: html2text

[Reiner Steib]

On Mon, Nov 08 2004, Reiner Steib wrote:

> [ The suggested patch from Jari's original message was: ]
>
> --8<---------------cut here---------------start------------->8---
> --- html2text.el.7.10	2004-11-06 17:20:46.000000000 +0200
> +++ html2text.el	2004-11-06 17:41:12.000000000 +0200
> @@ -42,8 +42,42 @@
>  (defvar html2text-format-single-element-list '(("hr" . html2text-clean-hr)))
>
>  (defvar html2text-replace-list
> -  '(("&nbsp;" . " ") ("&gt;" . ">") ("&lt;" . "<") ("&quot;" . "\"")
> -    ("&amp;" . "&") ("&apos;" . "'"))
> +  '(("&acute;" . "`")

This should be "´".

> +    ("&amp;" . "&")
> +    ("&apos;" . "'")
> +    ("&brvbar;" . "|")
> +    ("&cent;" . "c")
> +    ("&circ;" . "^")
> +    ("&copy;" . "(C)")
> +    ("&curren;" . "¤")
> +    ("&deg;" . "degree")
> +    ("&divide;" . "/")
> +    ("&euro;" . "e")
> +    ("&frac12;" . "½")
[...]

It seems strange to use Latin-1 characters for some entities, but not
for all encodable by Latin-1.

On a second thought, it looks like there are already more or less
complete lists[1] e.g. in `mm-url-html-entities' (from Gnus),
`sgml-char-names', `sgml-char-names-table', `iso-iso2sgml-trans-tab'
(Emacs) or `w3m-entity-alist' (emacs-w3m).

Probably one of these could be used.  Hm, maybe the function
`iso-sgml2iso' could be used in `html2text.el'?

Bye, Reiner.

[1] Might be checked with
    http://www.w3.org/TR/REC-html40/sgml/entities.html or other
    tables.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: html2text

[Jari Aalto]

Reiner Steib <reinersteib+gmane@imap.cc> writes:
| On Sat, Nov 06 2004, Jari Aalto+mail.emacs wrote:
| 
| > This is your copy. Article has been posted to the newsgroup(s).
| 
| I didn't see your message on emacs-devel, see
| <URL:http://thread.gmane.org/1099247139.071920.12084.nullmailer@Update.UU=
| .SE>.
| >
| > 2004-11-06 Sat  Jari Aalto  <jari dot aalto A T cante dot net>
| >
| >         * text2html (html2text-replace-list). Added more HTML 4.0
| >         entities.
| 
| It seems you have signed papers for Emacs as you are listed in the
| AUTHORS file.  But I can't check it myself.  Could you please confirm?

Confirmed. I have signed the papers.

Jari