jenkins-bot has submitted this change. ( https://gerrit.wikimedia.org/r/c/pywikibot/core/+/634834 )

Change subject: [IMPR] Reduce code complexity of HttpRequest.header_encoding() ......................................................................

[IMPR] Reduce code complexity of HttpRequest.header_encoding()

Change-Id: Ie00cc6bba6b6f1edf02cb22a41387014206bb033 --- M pywikibot/comms/threadedhttp.py 1 file changed, 20 insertions(+), 19 deletions(-)

Approvals: Xqt: Looks good to me, approved jenkins-bot: Verified

diff --git a/pywikibot/comms/threadedhttp.py b/pywikibot/comms/threadedhttp.py index 8b4338e..208dc8e 100644 --- a/pywikibot/comms/threadedhttp.py +++ b/pywikibot/comms/threadedhttp.py @@ -108,26 +108,27 @@ @property def header_encoding(self): """Return charset given by the response header.""" - if not hasattr(self, '_header_encoding'): - content_type = self.response_headers.get('content-type', '') - pos = content_type.find('charset=') - if pos >= 0: - pos += len('charset=') - encoding = self.response_headers['content-type'][pos:] - self._header_encoding = encoding - elif 'json' in content_type: - # application/json | application/sparql-results+json - self._header_encoding = 'utf-8' - elif 'xml' in content_type: - header = self.raw[:100].splitlines()[0] # bytes - m = re.search( - br'encoding=(["'])(?P<encoding>.+?)\1', header) - if m: - self._header_encoding = m.group('encoding').decode('utf-8') - else: - self._header_encoding = 'utf-8' + if hasattr(self, '_header_encoding'): + return self._header_encoding + + content_type = self.response_headers.get('content-type', '') + m = re.search('charset=(?P<charset>.*?$)', content_type) + if m: + self._header_encoding = m.group('charset') + elif 'json' in content_type: + # application/json | application/sparql-results+json + self._header_encoding = 'utf-8' + elif 'xml' in content_type: + header = self.raw[:100].splitlines()[0] # bytes + m = re.search( + br'encoding=(["'])(?P<encoding>.+?)\1', header) + if m: + self._header_encoding = m.group('encoding').decode('utf-8') else: - self._header_encoding = None + self._header_encoding = 'utf-8' + else: + self._header_encoding = None + return self._header_encoding

@property