Hello Nick, thanks for the valid comments I hope I addressed them all.
I used the dictionary hint, but I didn't use it everywhere - mostly just where the
same dict would be reused in the same function.
On 11/03/2015 04:40 PM, Nikolai Kondrashov wrote:
> +
> + # We can create override with name root.
> + ent.assert_passwd_by_name(
> + 'user2',
> + dict(name='root', passwd='*', uid=10020, gid=20020,
> + gecos='Overriden User 2',
> + dir='/home/ov/user2',
> + shell='/bin/ov_user2_shell'))
> +
> + ent.assert_passwd_by_uid(0, dict(name="root"))
Could you please add a comment to this line explaining what it tests?
Well, I can't :-). I just hit this case and I don't think this is a security bug,
but I decided to highlight it and at least canonize this behavior - whenever
implementation would change test would need to be changed too.
I can remove this part of the test - no trouble at all.
Could you make each set item start on a new line, so it's easier to read?
Something like this:
assert set(out.splitlines()) == \
set(['user1@LDAP:ov_user1:10010:20010:Overriden User 1:' \
'/home/ov/user1:/bin/ov_user1_shell',
'user2@LDAP:ov_user2:10020:20020:Overriden User 2:' \
'/home/ov/user2:/bin/ov_user2_shell'])
Sorry I could not use this format as pep8 didn't like it. I had to remove the
backslash and then '2nd' part of the string could not be intended (pep8 didn't
like too).