wassname/FanFicFare
1
0
Fork 0
mirror of https://github.com/wassname/FanFicFare.git synced 2026-09-10 11:40:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix for replace_br_with_p(htmlheuristics) when author includes <>.

Browse Source
This commit is contained in:
Jim Miller
2016-02-02 09:58:53 -06:00
parent 4bf3399e35
commit 1be99aa95c
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fanficfare/htmlheuristics.py
+5 -1
View File
@@ -34,7 +34,7 @@ def replace_br_with_p(body):
return body
# logger.debug(u'---')
# logger.debug(u'BODY start.: ' + body[:250])
# logger.debug(u'BODY start.: ' + body[:4000])
# logger.debug(u'--')
# logger.debug(u'BODY end...: ' + body[-250:])
# logger.debug(u'BODY.......: ' + body)
@@ -49,6 +49,9 @@ def replace_br_with_p(body):
if is_valid_block(body) and body.find('<div') == 0:
body = body[body.index('>')+1:body.rindex('<')].strip()
# BS is doing some BS on entities, meaning &lt; and &gt; are turned into < and >... a **very** bad idea in html.
body = re.sub(r'&(.+?);', r'XAMP;\1;', body)
body = soup_up_div(u'<div>' + body + u'</div>')
body = body[body.index('>')+1:body.rindex('<')]
@@ -238,6 +241,7 @@ def replace_br_with_p(body):
body = re.sub(r'\s*<(\S+)[^>]*>\s*</\1>', r'', body)
body = body.replace(u'{br /}', u'<br />')
body = re.sub(r'XAMP;(.+?);', r'&\1;', body)
body = body.strip()
# re-wrap in div tag.