From df2de9c1759e43588ab07a7ffc491e66707e8d30 Mon Sep 17 00:00:00 2001
From: Tom Gross <itconsense@gmail.com>
Date: Fri, 29 May 2015 13:41:22 +0000
Subject: [PATCH] remove CMFDefault dependency

---
 .../PortalTransforms/transforms/html_body.py  |  2 +-
 .../transforms/pdf_to_html.py                 |  2 +-
 .../transforms/rtf_to_html.py                 |  2 +-
 .../PortalTransforms/transforms/safe_html.py  | 76 ++++++++++++++++---
 4 files changed, 68 insertions(+), 14 deletions(-)

diff --git a/Products/PortalTransforms/transforms/html_body.py b/Products/PortalTransforms/transforms/html_body.py
index e03fdaa..78181ab 100644
--- a/Products/PortalTransforms/transforms/html_body.py
+++ b/Products/PortalTransforms/transforms/html_body.py
@@ -1,6 +1,6 @@
 from Products.PortalTransforms.interfaces import ITransform
 from zope.interface import implements
-from Products.CMFDefault.utils import bodyfinder
+from Products.PortalTransforms.libtransforms.utils import bodyfinder
 
 
 class HTMLBody:
diff --git a/Products/PortalTransforms/transforms/pdf_to_html.py b/Products/PortalTransforms/transforms/pdf_to_html.py
index 8c6159d..27f9f24 100644
--- a/Products/PortalTransforms/transforms/pdf_to_html.py
+++ b/Products/PortalTransforms/transforms/pdf_to_html.py
@@ -7,7 +7,7 @@
 from Products.PortalTransforms.libtransforms.utils import sansext
 from Products.PortalTransforms.libtransforms.commandtransform import (
     commandtransform, popentransform)
-from Products.CMFDefault.utils import bodyfinder
+from Products.PortalTransforms.libtransforms.utils import bodyfinder
 import os
 
 
diff --git a/Products/PortalTransforms/transforms/rtf_to_html.py b/Products/PortalTransforms/transforms/rtf_to_html.py
index cb50d20..7df9d68 100644
--- a/Products/PortalTransforms/transforms/rtf_to_html.py
+++ b/Products/PortalTransforms/transforms/rtf_to_html.py
@@ -5,9 +5,9 @@
 from Products.PortalTransforms.interfaces import ITransform
 from zope.interface import implements
 from Products.PortalTransforms.libtransforms.utils import sansext
+from Products.PortalTransforms.libtransforms.utils import bodyfinder
 from Products.PortalTransforms.libtransforms.commandtransform import \
     commandtransform
-from Products.CMFDefault.utils import bodyfinder
 import os
 
 
diff --git a/Products/PortalTransforms/transforms/safe_html.py b/Products/PortalTransforms/transforms/safe_html.py
index 255a878..e5e243c 100644
--- a/Products/PortalTransforms/transforms/safe_html.py
+++ b/Products/PortalTransforms/transforms/safe_html.py
@@ -6,17 +6,74 @@
 from Products.PortalTransforms.interfaces import ITransform
 from zope.interface import implements
 from Products.PortalTransforms.utils import log
-from Products.CMFDefault.utils import bodyfinder
-from Products.CMFDefault.utils import IllegalHTML
-from Products.CMFDefault.utils import VALID_TAGS
-from Products.CMFDefault.utils import NASTY_TAGS
+from Products.PortalTransforms.libtransforms.utils import bodyfinder
 from Products.PortalTransforms.utils import safeToInt
 
-# tag mapping: tag -> short or long tag
-VALID_TAGS = VALID_TAGS.copy()
-NASTY_TAGS = NASTY_TAGS.copy()
+class IllegalHTML(ValueError):
+    """ Illegal HTML error.
+    """
+
+
+# These are the HTML tags that we will leave intact
+VALID_TAGS = { 'a'          : 1
+             , 'b'          : 1
+             , 'base'       : 0
+             , 'big'        : 1
+             , 'blockquote' : 1
+             , 'body'       : 1
+             , 'br'         : 0
+             , 'caption'    : 1
+             , 'cite'       : 1
+             , 'code'       : 1
+             , 'dd'         : 1
+             , 'div'        : 1
+             , 'dl'         : 1
+             , 'dt'         : 1
+             , 'em'         : 1
+             , 'h1'         : 1
+             , 'h2'         : 1
+             , 'h3'         : 1
+             , 'h4'         : 1
+             , 'h5'         : 1
+             , 'h6'         : 1
+             , 'head'       : 1
+             , 'hr'         : 0
+             , 'html'       : 1
+             , 'i'          : 1
+             , 'img'        : 0
+             , 'kbd'        : 1
+             , 'li'         : 1
+           # , 'link'       : 1 type="script" hoses us
+             , 'meta'       : 0
+             , 'ol'         : 1
+             , 'p'          : 1
+             , 'pre'        : 1
+             , 'small'      : 1
+             , 'span'       : 1
+             , 'strong'     : 1
+             , 'sub'        : 1
+             , 'sup'        : 1
+             , 'table'      : 1
+             , 'tbody'      : 1
+             , 'td'         : 1
+             , 'th'         : 1
+             , 'title'      : 1
+             , 'tr'         : 1
+             , 'tt'         : 1
+             , 'u'          : 1
+             , 'ul'         : 1
+             }
+
+NASTY_TAGS = { 'script'     : 1
+             , 'object'     : 1
+             , 'embed'      : 1
+             , 'applet'     : 1
+             , 'style'      : 1  # this helps improve Word HTML cleanup.
+             , 'meta'       : 1  # allowed by parsers, but can cause unexpected behavior
+             }
+
 
-# add some tags to allowed types. These should be backported to CMFDefault.
+# add some tags to allowed types.
 VALID_TAGS['ins'] = 1
 VALID_TAGS['del'] = 1
 VALID_TAGS['q'] = 1
@@ -61,9 +118,6 @@
 VALID_TAGS['time'] = 1
 VALID_TAGS['video'] = 1
 
-# add some tags to nasty.
-NASTY_TAGS['style'] = 1  # this helps improve Word HTML cleanup.
-NASTY_TAGS['meta'] = 1  # allowed by parsers, but can cause unexpected behavior
 
 
 msg_pat = """