Description Florian Best 2021-09-15 22:51:10 CEST

The syntax class dnsName_umlauts is still broken / or doesn't match its error message:

 >>> univention.admin.syntax.dnsName_umlauts.parse('''foo-bar''')                                                                                                                                                                              
'foo-bar'
>>> univention.admin.syntax.dnsName_umlauts.parse('''foo_bar''')                                                                                                                                                                              
'foo_bar'
>>> univention.admin.syntax.dnsName_umlauts.parse('''foo.bar''')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python3/dist-packages/univention/admin/syntax.py", line 227, in parse
    raise univention.admin.uexceptions.valueError(self.error_message)
univention.admin.uexceptions.valueError: A host name or FQDN must start and end with a letter or number. In between additionally dashes, dots and underscores are allowed.


+++ This bug was initially created as a clone of Bug #35041 +++

The UDM syntax univention.admin.syntax.dnsName_umlauts is (amongst others) used in the UDM module computers/windows.py for checking hostnames seems to be broken.

class dnsName_umlauts(simple):
	regex = re.compile('(?u)(^\w[\w -.]*\w$)|\w*$')

>>> regex = re.compile('(?u)^\w[\w -.]*\w$')
>>> regex.match('''foo ( - " , ' _ ) bar''')
<_sre.SRE_Match object at 0x29176b0>
>>> regex.match(u'''foo ( - " ä , ü ' _ ) bar''')
<_sre.SRE_Match object at 0x2917850>

The part " -." specifies a character range which contains all characters between space and dot (e.g. ()"'-, ). So this does not look ok to me.

Side note: the second alternative "\w*$" of the regex should use ^ otherwise it would match to any string if re.search() is used instead of re.match().