x/stackalytics
Code Issues Proposed changes

Merge "Strip unsafe chars from author email"

Browse Source
This commit is contained in:
Jenkins
2015-01-15 13:33:36 +00:00
committed by Gerrit Code Review
parent cc39307286 41633f49ab
commit f9b36da748
3 changed files with 13 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
stackalytics/processor/utils.py
View File

@@ -234,6 +234,10 @@ def safe_encode(s):
return six.moves.urllib.parse.quote(s.encode('utf-8')) return six.moves.urllib.parse.quote(s.encode('utf-8'))
def keep_safe_chars(s):
return re.sub(r'[^\x21-\x7e\x80-\xff]+', '', s)
def make_module_group(module_group_id, name=None, modules=None, tag='module'): def make_module_group(module_group_id, name=None, modules=None, tag='module'):
return {'id': module_group_id, return {'id': module_group_id,
'module_group_name': name or module_group_id, 'module_group_name': name or module_group_id,

3
stackalytics/processor/vcs.py
View File

@@ -211,6 +211,9 @@ class Git(Vcs):
# ignore commits with empty email (there are some < Essex) # ignore commits with empty email (there are some < Essex)
continue continue
commit['author_email'] = utils.keep_safe_chars(
commit['author_email'])
diff_stat_str = rec.group('diff_stat') diff_stat_str = rec.group('diff_stat')
diff_rec = re.search(DIFF_STAT_PATTERN, diff_stat_str) diff_rec = re.search(DIFF_STAT_PATTERN, diff_stat_str)

6
tests/unit/test_utils.py
View File

@@ -78,6 +78,12 @@ class TestUtils(testtools.TestCase):
self.assertEqual(expected, utils.add_index( self.assertEqual(expected, utils.add_index(
sequence, start=0, item_filter=lambda x: x['name'] != 'B')) sequence, start=0, item_filter=lambda x: x['name'] != 'B'))
def test_keep_safe_chars(self):
self.assertEqual('somemoretext',
utils.keep_safe_chars('some more text'))
self.assertEqual(u'(unicode)',
utils.keep_safe_chars(u'(unicode \u0423) '))
def test_normalize_company_name(self): def test_normalize_company_name(self):
company_names = ['EMC Corporation', 'Abc, corp..', 'Mirantis IT.', company_names = ['EMC Corporation', 'Abc, corp..', 'Mirantis IT.',
'Red Hat, Inc.', 'abc s.r.o. ABC', '2s.r.o. co', 'Red Hat, Inc.', 'abc s.r.o. ABC', '2s.r.o. co',