matrix/weechat-matrix
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Merge pull request #9 from dkasak/master

Browse source 
HTML ref parsing fix.
This commit is contained in:
poljar 2018-03-22 21:10:52 +01:00 committed by GitHub
commit 13bfb33b73
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 24 additions and 16 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

32
matrix/colors.py
View file

@ -328,14 +328,13 @@ class MatrixHtmlParser(HTMLParser):
except AttributeError: except AttributeError:
return HTMLParser.unescape(self, text) return HTMLParser.unescape(self, text)
def feed(self, text): def add_substring(self, text, attrs):
text = self.unescape(text) fmt_string = FormattedString(text, attrs)
return HTMLParser.feed(self, text) self.substrings.append(fmt_string)
def _toggle_attribute(self, attribute): def _toggle_attribute(self, attribute):
if self.text: if self.text:
self.substrings.append( self.add_substring(self.text, self.attributes.copy())
FormattedString(self.text, self.attributes.copy()))
self.text = "" self.text = ""
self.attributes[attribute] = not self.attributes[attribute] self.attributes[attribute] = not self.attributes[attribute]
@ -352,11 +351,9 @@ class MatrixHtmlParser(HTMLParser):
self._toggle_attribute("quote") self._toggle_attribute("quote")
elif tag == "br": elif tag == "br":
if self.text: if self.text:
self.substrings.append( self.add_substring(self.text, self.attributes.copy())
FormattedString(self.text, self.attributes.copy()))
self.text = "\n" self.text = "\n"
self.substrings.append( self.add_substring(self.text, DEFAULT_ATRIBUTES.copy())
FormattedString(self.text, DEFAULT_ATRIBUTES.copy()))
self.text = "" self.text = ""
elif tag == "font": elif tag == "font":
for key, value in attrs: for key, value in attrs:
@ -367,8 +364,7 @@ class MatrixHtmlParser(HTMLParser):
continue continue
if self.text: if self.text:
self.substrings.append( self.add_substring(self.text, self.attributes.copy())
FormattedString(self.text, self.attributes.copy()))
self.text = "" self.text = ""
self.attributes["fgcolor"] = color self.attributes["fgcolor"] = color
else: else:
@ -387,20 +383,24 @@ class MatrixHtmlParser(HTMLParser):
self._toggle_attribute("quote") self._toggle_attribute("quote")
elif tag == "font": elif tag == "font":
if self.text: if self.text:
self.substrings.append( self.add_substring(self.text, self.attributes.copy())
FormattedString(self.text, self.attributes.copy()))
self.text = "" self.text = ""
self.attributes["fgcolor"] = None self.attributes["fgcolor"] = None
else: else:
pass pass
def handle_data(self, data): def handle_data(self, data):
self.text = self.text + data self.text += data
def handle_entityref(self, name):
self.text += self.unescape("&{};".format(name))
def handle_charref(self, name):
self.text += self.unescape("&{};".format(name))
def get_substrings(self): def get_substrings(self):
if self.text: if self.text:
self.substrings.append( self.add_substring(self.text, self.attributes.copy())
FormattedString(self.text, self.attributes.copy()))
return self.substrings return self.substrings

8
tests/http_parser_test.py
View file

@ -31,3 +31,11 @@ def test_html_numeric_reference_parsing(entitydef):
num = entitydef[2] num = entitydef[2]
parser = MatrixHtmlParser() parser = MatrixHtmlParser()
assert parser.unescape('&#{};'.format(num)) == character assert parser.unescape('&#{};'.format(num)) == character
def test_parsing_of_escaped_brackets():
p = MatrixHtmlParser()
p.feed('<pre><code>&lt;faketag&gt;</code></pre>')
s = p.get_substrings()
print(s)
assert s[0].text == '<faketag>' and len(s) == 1