URL and repo regex should be non-greedy (maybe)

[GreatBahram]

Hello, I think there is a problem in this regex or maybe all of them:

import re

pattern = r'https://(?P<domain>.+)/(?P<owner>.+)/(?P<repo>.+?)(?:\.git)?$'
url = 'https://github.com/nephila/giturlparse/blob/master/giturlparse/platforms/github.py'

match = re.match(pattern, url).groupdict()

I suppose, this is an invalid URL, for simple reason git clone would not accept it; however, when we check the match object, things are a bit strange,:

{'domain': 'github.com/nephila/giturlparse/blob/master/giturlparse',
 'owner': 'platforms',
 'repo': 'github.py'}

If we make them non-greedy, it would a little better:

pattern = r'https://(?P<domain>.+?)/(?P<owner>.+?)/(?P<repo>.+?)(?:\.git)?$'
match = re.match(pattern, url).groupdict()

{'domain': 'github.com',
 'owner': 'nephila',
 'repo': 'giturlparse/blob/master/giturlparse/platforms/github.py'}

• I don't know whether I'm right or not, but I think it's not normal to see a / in repo field, is it? If it is not, so you guys should return False for valid attribute, am I right?
• Is there any harm to make all the patterns non-greedy? Is there any circumstances that there would be '/' in the repo field?

@yakky

[GreatBahram]

I think the issue #10 is somehow related to this problem.

[yakky]

@GreatBahram thanks for reporting. In PR #19 I implemented a slightly different logic, which should be more consistent across url schemas