From ctlaux at gmail.com  Sun Jun  8 13:47:51 2014
From: ctlaux at gmail.com (Christopher Laux)
Date: Sun, 8 Jun 2014 15:47:51 +0200
Subject: [Closure-devel] Bug in html parsing
Message-ID: <CAD6+YkYXDNPD2wMy1Vf-NtLrQ4Y1OsE1gBtHbo-GqyuR6Ohipw@mail.gmail.com>

Hi,

I think I've found a bug in closure. If I execute

(chtml:parse "<small><p>test1</p><p>test2</p></small>"
(chtml:make-lhtml-builder))

> (:HTML NIL (:HEAD NIL)
 (:BODY NIL (:SMALL NIL) (:P NIL (:SMALL NIL "test1")) (:P NIL "test2")))

I get that incorrect parse tree. The example is taken from a real website
and the same happens inside the entire page. That happens both with the
current git version and the quicklisp version (which might just be the same
one).

Any help?

Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/closure-devel/attachments/20140608/1ae4966d/attachment.html>

From bhyde at pobox.com  Sun Jun  8 15:16:46 2014
From: bhyde at pobox.com (Ben Hyde)
Date: Sun, 8 Jun 2014 11:16:46 -0400
Subject: [Closure-devel] Bug in html parsing
In-Reply-To: <CAD6+YkYXDNPD2wMy1Vf-NtLrQ4Y1OsE1gBtHbo-GqyuR6Ohipw@mail.gmail.com>
References: <CAD6+YkYXDNPD2wMy1Vf-NtLrQ4Y1OsE1gBtHbo-GqyuR6Ohipw@mail.gmail.com>
Message-ID: <C269C970-27FE-45E7-8913-0F07B45F2236@pobox.com>

Ah, the joys of HTML - see, for example: http://stackoverflow.com/questions/9852312/list-of-html5-elements-that-can-be-nested-inside-p-element

On Jun 8, 2014, at 9:47 AM, Christopher Laux <ctlaux at gmail.com> wrote:
> Hi,
> 
> I think I've found a bug in closure. If I execute
> 
> (chtml:parse "<small><p>test1</p><p>test2</p></small>" (chtml:make-lhtml-builder))
> 
> > (:HTML NIL (:HEAD NIL)
>  (:BODY NIL (:SMALL NIL) (:P NIL (:SMALL NIL "test1")) (:P NIL "test2")))
> 
> I get that incorrect parse tree. The example is taken from a real website and the same happens inside the entire page. That happens both with the current git version and the quicklisp version (which might just be the same one).
> 
> Any help?
> 
> Chris
> 
> _______________________________________________
> Closure-devel mailing list
> Closure-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/closure-devel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/closure-devel/attachments/20140608/a605c51f/attachment.html>

From ctlaux at gmail.com  Sun Jun  8 15:26:01 2014
From: ctlaux at gmail.com (Christopher Laux)
Date: Sun, 8 Jun 2014 17:26:01 +0200
Subject: [Closure-devel] Bug in html parsing
In-Reply-To: <C269C970-27FE-45E7-8913-0F07B45F2236@pobox.com>
References: <CAD6+YkYXDNPD2wMy1Vf-NtLrQ4Y1OsE1gBtHbo-GqyuR6Ohipw@mail.gmail.com>
 <C269C970-27FE-45E7-8913-0F07B45F2236@pobox.com>
Message-ID: <CAD6+YkaLa2VYGEhyavUz9vg1Pn3OLuBPWhgEeBqkdfcuDfjoPA@mail.gmail.com>

If that is not valid html, should it not raise an error instead of
returning an incorrect parse tree?

Is there any way of making it parse it the way it is written anyway?

Chris
 Am 08.06.2014 17:16 schrieb "Ben Hyde" <bhyde at pobox.com>:

> Ah, the joys of HTML - see, for example:
> http://stackoverflow.com/questions/9852312/list-of-html5-elements-that-can-be-nested-inside-p-element
>
> On Jun 8, 2014, at 9:47 AM, Christopher Laux <ctlaux at gmail.com> wrote:
>
> Hi,
>
> I think I've found a bug in closure. If I execute
>
> (chtml:parse "<small><p>test1</p><p>test2</p></small>"
> (chtml:make-lhtml-builder))
>
> > (:HTML NIL (:HEAD NIL)
>  (:BODY NIL (:SMALL NIL) (:P NIL (:SMALL NIL "test1")) (:P NIL "test2")))
>
> I get that incorrect parse tree. The example is taken from a real website
> and the same happens inside the entire page. That happens both with the
> current git version and the quicklisp version (which might just be the same
> one).
>
> Any help?
>
> Chris
>
> _______________________________________________
> Closure-devel mailing list
> Closure-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/closure-devel
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/closure-devel/attachments/20140608/9cd47a62/attachment.html>

From michal at lisp.pl  Wed Jun  4 10:34:46 2014
From: michal at lisp.pl (=?UTF-8?Q?Micha=C5=82_Psota?=)
Date: Wed, 04 Jun 2014 10:34:46 -0000
Subject: [Closure-devel] Bug in closure-html,
	when use with *print-case* set to :downcase
Message-ID: <CAAe-Mv8Uq-+H9vQ0nqtCBMHkwhvZ9-XFvLJDKhmoiCpJkmr0dA@mail.gmail.com>

Hello,

I've noticed that in 2 places strings are not correctly interned when
you load closure-html with *print-case* set to :downcase. To make it
work, I've changed
(intern ?) to (intern (string-upcase ?)).

Please find patch attached.

Best regards,
Micha? Psota
-------------- next part --------------
diff -rupN /tmp/closure-html-2010-09-20/src/parse/sgml-parse.lisp /tmp/closure-html/src/parse/sgml-parse.lisp
--- /tmp/closure-html-2010-09-20/src/parse/sgml-parse.lisp	2010-09-20 00:07:37.000000000 +0200
+++ /tmp/closure-html/src/parse/sgml-parse.lisp	2014-05-20 12:50:01.817148769 +0200
@@ -860,7 +860,7 @@
 
 (let ((kw-pkg (find-package :keyword)))
   (defun kintern (x)
-    (intern x kw-pkg)))
+    (intern (string-upcase x) kw-pkg)))
 
 (defun canon-value (input dtd tag slot value)
   (let* ((attlist (find-element-attlist dtd tag))
diff -rupN /tmp/closure-html-2010-09-20/src/util/clex.lisp /tmp/closure-html/src/util/clex.lisp
--- /tmp/closure-html-2010-09-20/src/util/clex.lisp	2010-09-20 00:07:37.000000000 +0200
+++ /tmp/closure-html/src/util/clex.lisp	2014-05-20 12:53:08.437157801 +0200
@@ -365,7 +365,7 @@
      (mapcar #'(lambda (x y) (setf (cdr x) y))
              starts (ndfsa->dfsa (mapcar #'cdr starts))))
     ;;    (print (number-states starts))
-    `(DEFUN ,(intern (format nil "MAKE-~A-LEXER" name)) (INPUT)
+    `(DEFUN ,(intern (format nil "~:@(MAKE-~A-LEXER~)" name)) (INPUT)
         (LET* ((STARTS ,(loadable-states-form starts))
                (SUB-STATE 'INITIAL)
                (STATE NIL)
@@ -538,7 +538,7 @@
                                (mungle-transitions (state-transitions state))))
                        (mapcar #'cdr starts))
       (format T "~&~D states." n))
-    `(DEFUN ,(intern (format nil "MAKE-~A-LEXER" name)) (INPUT)
+    `(DEFUN ,(intern (format nil "~:@(MAKE-~A-LEXER~)" name)) (INPUT)
         (LET* ((STARTS ,(loadable-states-form starts))
                (SUB-STATE 'INITIAL)
                (STATE NIL)

From look.wangluke at gmail.com  Mon Jun 30 07:34:42 2014
From: look.wangluke at gmail.com (lookwong)
Date: Mon, 30 Jun 2014 07:34:42 -0000
Subject: [Closure-devel] a bug in read-pcdata (source file: sgml-parse.lisp)
Message-ID: <CAJuEQWDtQCzbMWbtKRE2=v3=VD3G_24gEcewcK=Lb7W5J3StWw@mail.gmail.com>

In function read-pcdata:

*(t *
*      (setf (aref scratch sp) ch)    ;recode character read*
*      (setf sp (the fixnum (+ sp 1)))*
*
<http://repo.or.cz/w/closure-html.git/blob/HEAD:/src/parse/sgml-parse.lisp#l371>
(cond ((= sp se)                         ;end of scratch pad reached?*
*            (enlarge-scratch-pad input)*
*            (setf scratch (a-stream-scratch input)*
*                   se (length scratch))))))))*



Should first enlarge the input buffer, or will cause out of index error.


*(t**
(cond ((= sp se)                         ;end of scratch pad reached?*
*       (enlarge-scratch-pad input)*
*       (setf scratch (a-stream-scratch input)*
*               se (length scratch))))*

*      (setf (aref scratch sp) ch)    ;recode character read*
*      (setf sp (the fixnum (+ sp 1)))**))))*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/closure-devel/attachments/20140630/ab3b9dd4/attachment.html>