From 576082bd6205da8bbf6d40e4eb3f6b6f0e60fcc0 Mon Sep 17 00:00:00 2001
From: Ross Jones <rossdjones@gmail.com>
Date: Fri, 25 May 2012 11:08:23 +0100
Subject: [PATCH] 2414 Remove LXML as a dependency from core

Removes LXML from the requirements by replacing the html stripping
when generating a markdown extract of the description.

Also removes LXML from the initialisation code that checks the
solr schema version
---
 ckan/lib/helpers.py         | 8 +++-----
 ckan/lib/search/__init__.py | 7 +++----
 doc/install-from-source.rst | 5 +----
 requires/lucid_present.txt  | 3 +--
 4 files changed, 8 insertions(+), 15 deletions(-)

diff --git a/ckan/lib/helpers.py b/ckan/lib/helpers.py
index 538e152b331..d91403424d5 100644
--- a/ckan/lib/helpers.py
+++ b/ckan/lib/helpers.py
@@ -25,7 +25,6 @@
 from routes import redirect_to as _redirect_to
 from routes import url_for as _routes_default_url_for
 from alphabet_paginate import AlphaPage
-from lxml.html import fromstring
 import i18n
 import ckan.exceptions
 from pylons import request
@@ -467,8 +466,7 @@ def group_name_to_title(name):
 def markdown_extract(text, extract_length=190):
     if (text is None) or (text.strip() == ''):
         return ''
-    html = fromstring(markdown(text))
-    plain = html.xpath("string()")
+    plain = re.sub(r'<.*?>', '', markdown(text))
     return unicode(truncate(plain, length=extract_length, indicator='...', whole_word=True))
 
 def icon_url(name):
@@ -607,7 +605,7 @@ def parse_rfc_2822_date(date_str, assume_utc=True):
 
     RFC 2822 is the date format used in HTTP headers.  It should contain timezone
     information, but that cannot be relied upon.
-    
+
     If date_str doesn't contain timezone information, then the 'assume_utc' flag
     determines whether we assume this string is local (with respect to the
     server running this code), or UTC.  In practice, what this means is that if
@@ -616,7 +614,7 @@ def parse_rfc_2822_date(date_str, assume_utc=True):
 
     If timezone information is available in date_str, then the returned datetime
     is 'aware', ie - it has an associated tz_info object.
-    
+
     Returns None if the string cannot be parsed as a valid datetime.
     """
     time_tuple = email.utils.parsedate_tz(date_str)
diff --git a/ckan/lib/search/__init__.py b/ckan/lib/search/__init__.py
index 0731d0e54b9..abc78b798fd 100644
--- a/ckan/lib/search/__init__.py
+++ b/ckan/lib/search/__init__.py
@@ -261,13 +261,12 @@ def check_solr_schema_version(schema_file=None):
         url = 'file://%s' % schema_file
         res = urllib2.urlopen(url)
 
-    from lxml import etree
-    tree = etree.fromstring(res.read())
+    import xml.dom.minidom
+    tree = xml.dom.minidom.parseString(res.read())
 
-    version = tree.xpath('//schema/@version')
+    version = tree.documentElement.getAttribute('version')
     if not len(version):
         raise SearchError('Could not extract version info from the SOLR schema, using file: \n%s' % url)
-    version = version[0]
 
     if not version in SUPPORTED_SCHEMA_VERSIONS:
         raise SearchError('SOLR schema version not supported: %s. Supported versions are [%s]'
diff --git a/doc/install-from-source.rst b/doc/install-from-source.rst
index c6dc209c2ca..bb324c32d25 100644
--- a/doc/install-from-source.rst
+++ b/doc/install-from-source.rst
@@ -125,7 +125,7 @@ WebOb has to be installed explicitly afterwards because by installing pylons wit
 
 Now to install the remaining dependencies in requires/lucid_present.txt and you are using Ubuntu Lucid 10.04 you can install the system versions::
 
-    sudo apt-get install python-pybabel python-psycopg2 python-lxml 
+    sudo apt-get install python-pybabel python-psycopg2
     sudo apt-get install python-pylons python-repoze.who 
     sudo apt-get install python-repoze.who-plugins python-tempita python-zope.interface
     
@@ -135,9 +135,6 @@ Alternatively, if you are not using Ubuntu Lucid 10.04 you'll need to install th
 
     pip install --ignore-installed -r pyenv/src/ckan/requires/lucid_present.txt
 
-This will take a **long** time. Particularly the install of the ``lxml``
-package.
-
 At this point you will need to deactivate and then re-activate your
 virtual environment to ensure that all the scripts point to the correct
 locations:
diff --git a/requires/lucid_present.txt b/requires/lucid_present.txt
index aabdb0cedad..a4cbfb03f12 100644
--- a/requires/lucid_present.txt
+++ b/requires/lucid_present.txt
@@ -2,13 +2,12 @@
 # apt-get if you are on that platform. If you are using a different platform
 # you can install these dependencies via pip instead.
 #
-# sudo apt-get install python-pybabel python-psycopg2 python-lxml
+# sudo apt-get install python-pybabel python-psycopg2
 # sudo apt-get install python-pylons python-repoze.who
 # sudo apt-get install python-repoze.who-plugins python-tempita python-zope.interface
 
 babel==0.9.4
 psycopg2==2.0.13
-lxml==2.2.4
 # Specifying particular version of WebOb because later version has incompatibility
 # with pylons 0.9.7 (change to imports of Multidict)
 webob==1.0.8