o2w/talon
4
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #86 from dougkeen/master

Browse Source 
Fix #85 (exception when stripping gmail quotes)
This commit is contained in:
Sergey Obukhov
2016-04-07 17:47:38 -07:00
parent f3c0942c49 333beb94af
commit d3de9e6893
2 changed files with 12 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
talon/html_quotations.py
View File

@@ -78,7 +78,7 @@ def delete_quotation_tags(html_note, counter, quotation_checkpoints):
def cut_gmail_quote(html_message): def cut_gmail_quote(html_message):
''' Cuts the outermost block element with class gmail_quote. ''' ''' Cuts the outermost block element with class gmail_quote. '''
gmail_quote = html_message.cssselect('div.gmail_quote') gmail_quote = html_message.cssselect('div.gmail_quote')
if gmail_quote and not RE_FWD.match(gmail_quote[0].text): if gmail_quote and (gmail_quote[0].text is None or not RE_FWD.match(gmail_quote[0].text)):
gmail_quote[0].getparent().remove(gmail_quote[0]) gmail_quote[0].getparent().remove(gmail_quote[0])
return True return True

11
tests/html_quotations_test.py
View File

@@ -131,6 +131,17 @@ def test_gmail_quote():
RE_WHITESPACE.sub('', quotations.extract_from_html(msg_body))) RE_WHITESPACE.sub('', quotations.extract_from_html(msg_body)))
def test_gmail_quote_compact():
msg_body = 'Reply' \
'<div class="gmail_quote">' \
'<div class="gmail_quote">On 11-Apr-2011, at 6:54 PM, Bob &lt;bob@example.com&gt; wrote:' \
'<div>Test</div>' \
'</div>' \
'</div>'
eq_("<html><body><p>Reply</p></body></html>",
RE_WHITESPACE.sub('', quotations.extract_from_html(msg_body)))
def test_gmail_quote_blockquote(): def test_gmail_quote_blockquote():
msg_body = """Message msg_body = """Message
<blockquote class="gmail_quote"> <blockquote class="gmail_quote">