Improve VCS url splits #755

Signed-off-by: Philippe Ombredanne <[email protected]>

src/packagedcode/utils.py

@@ -1,5 +1,5 @@
 #
-# Copyright (c) 2015 nexB Inc. and others. All rights reserved.
+# Copyright (c) 2017 nexB Inc. and others. All rights reserved.
 # http://nexb.com and https:/nexB/scancode-toolkit/
 # The ScanCode software is licensed under the Apache License version 2.0.
 # Data generated with ScanCode require an acknowledgment.
@@ -75,8 +75,12 @@ def parse_repo_url(repo_url):
  return repo_url
 
  if repo_url.startswith('git@'):
- left, right = repo_url.split('@', 1)
- host, repo = right.split(':', 1)
+ left, _, right = repo_url.partition('@')
+ if ':' in repo_url:
+ host, _, repo = right.partition(':')
+ else:
+ # [email protected]/Filirom1/npm2aur.git
+ host, _, repo = right.partition('/')
  if any(h in host for h in ['github', 'bitbucket', 'gitlab']):
  return 'https://%(host)s/%(repo)s' % locals()
  else:
@@ -89,7 +93,7 @@ def parse_repo_url(repo_url):
  'gitlab': 'https://gitlab.com/%(repo)s',
  'gist': 'https://gist.github.com/%(repo)s',
  }
- hoster, repo = repo_url.split(':', 1)
+ hoster, _, repo = repo_url.partition(':')
  return hoster_urls[hoster] % locals()
  elif len(repo_url.split('/')) == 2:
  # implicit github

tests/packagedcode/test_package_utils.py

@@ -101,3 +101,8 @@ def test_parse_repo_url_13(self):
  test = '[email protected]:foo/private.git'
  expected = 'https://gitlab.com/foo/private.git'
  assert expected == parse_repo_url(test)
+
+ def test_parse_git_repo_url_without_slash_slash(self):
+ test = '[email protected]/Filirom1/npm2aur.git'
+ expected = 'https:/Filirom1/npm2aur.git'
+ assert expected == parse_repo_url(test)