Import py-html5lib-0.11:

html5lib is a pure-python library for parsing HTML. The parser is
designed to handle all flavours of HTML and  parses invalid documents
using well-defined error handling rules compatible with the behaviour of
major desktop web browsers.

Output is to a tree structure; the current release supports output to
DOM, ElementTree, lxml and BeautifulSoup tree formats as well as a
simple custom format.

4 files changed, 159 insertions, 0 deletions

diff --git a/textproc/py-html5lib/DESCR b/textproc/py-html5lib/DESCR
new file mode 100644
index 00000000000..996dd87fcdf
--- /dev/null
+++ b/textproc/py-html5lib/DESCR
@@ -0,0 +1,8 @@
+html5lib is a pure-python library for parsing HTML. The parser is
+designed to handle all flavours of HTML and  parses invalid documents
+using well-defined error handling rules compatible with the behaviour of
+major desktop web browsers.
+
+Output is to a tree structure; the current release supports output to
+DOM, ElementTree, lxml and BeautifulSoup tree formats as well as a
+simple custom format.
diff --git a/textproc/py-html5lib/Makefile b/textproc/py-html5lib/Makefile
new file mode 100644
index 00000000000..980cd7a6afe
--- /dev/null
+++ b/textproc/py-html5lib/Makefile
@@ -0,0 +1,18 @@
+# $NetBSD: Makefile,v 1.1.1.1 2009/01/27 17:27:07 joerg Exp $
+#
+
+DISTNAME=	html5lib-0.11
+PKGNAME=	${PYPKGPREFIX}-${DISTNAME}
+CATEGORIES=	textproc www python
+MASTER_SITES=	http://html5lib.googlecode.com/files/
+
+EXTRACT_SUFX=	.zip
+
+MAINTAINER=	joerg@NetBSD.org
+HOMEPAGE=	http://code.google.com/5/html5lib/
+COMMENT=	HTML5 parser and tokenizer
+
+PKG_DESTDIR_SUPPORT=	user-destdir
+
+.include "../../lang/python/egg.mk"
+.include "../../mk/bsd.pkg.mk"
diff --git a/textproc/py-html5lib/PLIST b/textproc/py-html5lib/PLIST
new file mode 100644
index 00000000000..0faaf01fe94
--- /dev/null
+++ b/textproc/py-html5lib/PLIST
@@ -0,0 +1,128 @@
+@comment $NetBSD: PLIST,v 1.1.1.1 2009/01/27 17:27:07 joerg Exp $
+${PYSITELIB}/${EGG_INFODIR}/PKG-INFO
+${PYSITELIB}/${EGG_INFODIR}/SOURCES.txt
+${PYSITELIB}/${EGG_INFODIR}/dependency_links.txt
+${PYSITELIB}/${EGG_INFODIR}/top_level.txt
+${PYSITELIB}/html5lib/__init__.py
+${PYSITELIB}/html5lib/__init__.pyc
+${PYSITELIB}/html5lib/__init__.pyo
+${PYSITELIB}/html5lib/constants.py
+${PYSITELIB}/html5lib/constants.pyc
+${PYSITELIB}/html5lib/constants.pyo
+${PYSITELIB}/html5lib/filters/__init__.py
+${PYSITELIB}/html5lib/filters/__init__.pyc
+${PYSITELIB}/html5lib/filters/__init__.pyo
+${PYSITELIB}/html5lib/filters/_base.py
+${PYSITELIB}/html5lib/filters/_base.pyc
+${PYSITELIB}/html5lib/filters/_base.pyo
+${PYSITELIB}/html5lib/filters/formfiller.py
+${PYSITELIB}/html5lib/filters/formfiller.pyc
+${PYSITELIB}/html5lib/filters/formfiller.pyo
+${PYSITELIB}/html5lib/filters/inject_meta_charset.py
+${PYSITELIB}/html5lib/filters/inject_meta_charset.pyc
+${PYSITELIB}/html5lib/filters/inject_meta_charset.pyo
+${PYSITELIB}/html5lib/filters/iso639codes.py
+${PYSITELIB}/html5lib/filters/iso639codes.pyc
+${PYSITELIB}/html5lib/filters/iso639codes.pyo
+${PYSITELIB}/html5lib/filters/lint.py
+${PYSITELIB}/html5lib/filters/lint.pyc
+${PYSITELIB}/html5lib/filters/lint.pyo
+${PYSITELIB}/html5lib/filters/optionaltags.py
+${PYSITELIB}/html5lib/filters/optionaltags.pyc
+${PYSITELIB}/html5lib/filters/optionaltags.pyo
+${PYSITELIB}/html5lib/filters/rfc2046.py
+${PYSITELIB}/html5lib/filters/rfc2046.pyc
+${PYSITELIB}/html5lib/filters/rfc2046.pyo
+${PYSITELIB}/html5lib/filters/rfc3987.py
+${PYSITELIB}/html5lib/filters/rfc3987.pyc
+${PYSITELIB}/html5lib/filters/rfc3987.pyo
+${PYSITELIB}/html5lib/filters/sanitizer.py
+${PYSITELIB}/html5lib/filters/sanitizer.pyc
+${PYSITELIB}/html5lib/filters/sanitizer.pyo
+${PYSITELIB}/html5lib/filters/validator.py
+${PYSITELIB}/html5lib/filters/validator.pyc
+${PYSITELIB}/html5lib/filters/validator.pyo
+${PYSITELIB}/html5lib/filters/whitespace.py
+${PYSITELIB}/html5lib/filters/whitespace.pyc
+${PYSITELIB}/html5lib/filters/whitespace.pyo
+${PYSITELIB}/html5lib/html5parser.py
+${PYSITELIB}/html5lib/html5parser.pyc
+${PYSITELIB}/html5lib/html5parser.pyo
+${PYSITELIB}/html5lib/inputstream.py
+${PYSITELIB}/html5lib/inputstream.pyc
+${PYSITELIB}/html5lib/inputstream.pyo
+${PYSITELIB}/html5lib/liberalxmlparser.py
+${PYSITELIB}/html5lib/liberalxmlparser.pyc
+${PYSITELIB}/html5lib/liberalxmlparser.pyo
+${PYSITELIB}/html5lib/sanitizer.py
+${PYSITELIB}/html5lib/sanitizer.pyc
+${PYSITELIB}/html5lib/sanitizer.pyo
+${PYSITELIB}/html5lib/serializer/__init__.py
+${PYSITELIB}/html5lib/serializer/__init__.pyc
+${PYSITELIB}/html5lib/serializer/__init__.pyo
+${PYSITELIB}/html5lib/serializer/htmlserializer.py
+${PYSITELIB}/html5lib/serializer/htmlserializer.pyc
+${PYSITELIB}/html5lib/serializer/htmlserializer.pyo
+${PYSITELIB}/html5lib/serializer/xhtmlserializer.py
+${PYSITELIB}/html5lib/serializer/xhtmlserializer.pyc
+${PYSITELIB}/html5lib/serializer/xhtmlserializer.pyo
+${PYSITELIB}/html5lib/tokenizer.py
+${PYSITELIB}/html5lib/tokenizer.pyc
+${PYSITELIB}/html5lib/tokenizer.pyo
+${PYSITELIB}/html5lib/treebuilders/__init__.py
+${PYSITELIB}/html5lib/treebuilders/__init__.pyc
+${PYSITELIB}/html5lib/treebuilders/__init__.pyo
+${PYSITELIB}/html5lib/treebuilders/_base.py
+${PYSITELIB}/html5lib/treebuilders/_base.pyc
+${PYSITELIB}/html5lib/treebuilders/_base.pyo
+${PYSITELIB}/html5lib/treebuilders/dom.py
+${PYSITELIB}/html5lib/treebuilders/dom.pyc
+${PYSITELIB}/html5lib/treebuilders/dom.pyo
+${PYSITELIB}/html5lib/treebuilders/etree.py
+${PYSITELIB}/html5lib/treebuilders/etree.pyc
+${PYSITELIB}/html5lib/treebuilders/etree.pyo
+${PYSITELIB}/html5lib/treebuilders/etree_lxml.py
+${PYSITELIB}/html5lib/treebuilders/etree_lxml.pyc
+${PYSITELIB}/html5lib/treebuilders/etree_lxml.pyo
+${PYSITELIB}/html5lib/treebuilders/simpletree.py
+${PYSITELIB}/html5lib/treebuilders/simpletree.pyc
+${PYSITELIB}/html5lib/treebuilders/simpletree.pyo
+${PYSITELIB}/html5lib/treebuilders/soup.py
+${PYSITELIB}/html5lib/treebuilders/soup.pyc
+${PYSITELIB}/html5lib/treebuilders/soup.pyo
+${PYSITELIB}/html5lib/treewalkers/__init__.py
+${PYSITELIB}/html5lib/treewalkers/__init__.pyc
+${PYSITELIB}/html5lib/treewalkers/__init__.pyo
+${PYSITELIB}/html5lib/treewalkers/_base.py
+${PYSITELIB}/html5lib/treewalkers/_base.pyc
+${PYSITELIB}/html5lib/treewalkers/_base.pyo
+${PYSITELIB}/html5lib/treewalkers/dom.py
+${PYSITELIB}/html5lib/treewalkers/dom.pyc
+${PYSITELIB}/html5lib/treewalkers/dom.pyo
+${PYSITELIB}/html5lib/treewalkers/etree.py
+${PYSITELIB}/html5lib/treewalkers/etree.pyc
+${PYSITELIB}/html5lib/treewalkers/etree.pyo
+${PYSITELIB}/html5lib/treewalkers/genshistream.py
+${PYSITELIB}/html5lib/treewalkers/genshistream.pyc
+${PYSITELIB}/html5lib/treewalkers/genshistream.pyo
+${PYSITELIB}/html5lib/treewalkers/lxmletree.py
+${PYSITELIB}/html5lib/treewalkers/lxmletree.pyc
+${PYSITELIB}/html5lib/treewalkers/lxmletree.pyo
+${PYSITELIB}/html5lib/treewalkers/pulldom.py
+${PYSITELIB}/html5lib/treewalkers/pulldom.pyc
+${PYSITELIB}/html5lib/treewalkers/pulldom.pyo
+${PYSITELIB}/html5lib/treewalkers/simpletree.py
+${PYSITELIB}/html5lib/treewalkers/simpletree.pyc
+${PYSITELIB}/html5lib/treewalkers/simpletree.pyo
+${PYSITELIB}/html5lib/treewalkers/soup.py
+${PYSITELIB}/html5lib/treewalkers/soup.pyc
+${PYSITELIB}/html5lib/treewalkers/soup.pyo
+${PYSITELIB}/html5lib/utils.py
+${PYSITELIB}/html5lib/utils.pyc
+${PYSITELIB}/html5lib/utils.pyo
+@dirrm ${PYSITELIB}/html5lib/treewalkers
+@dirrm ${PYSITELIB}/html5lib/treebuilders
+@dirrm ${PYSITELIB}/html5lib/serializer
+@dirrm ${PYSITELIB}/html5lib/filters
+@dirrm ${PYSITELIB}/${EGG_INFODIR}
+@dirrm ${PYSITELIB}/html5lib
diff --git a/textproc/py-html5lib/distinfo b/textproc/py-html5lib/distinfo
new file mode 100644
index 00000000000..4ea5b06ab58
--- /dev/null
+++ b/textproc/py-html5lib/distinfo
@@ -0,0 +1,5 @@
+$NetBSD: distinfo,v 1.1.1.1 2009/01/27 17:27:07 joerg Exp $
+
+SHA1 (html5lib-0.11.zip) = cfacf8feed09bd0d53bc713965d70c8e9a416e92
+RMD160 (html5lib-0.11.zip) = fd6e377fa4af43d008147ee3fa50a18533eaa19d
+Size (html5lib-0.11.zip) = 187505 bytes