The six way of doing this is to replace all occurences of "unicode"
with "six.text_type". However, "unicode" is non-ambiguous and
(arguably) easier to read. Also, using it makes the patches smaller,
which should help with backporting.
Reviewed-By: Petr Viktorin <pviktori@redhat.com>
In Python 3, zip() returns an iterator. To get a list, it must
be explicitly converted.
In most cases, zip() result is iterated over so this is not
necessary.
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Jan Cholasta <jcholast@redhat.com>
In Python 3, range() behaves like the old xrange().
The difference between range() and xrange() is usually not significant,
especially if the whole result is iterated over.
Convert xrange() usage to range() for small ranges.
Use modern idioms in a few other uses of range().
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Jan Cholasta <jcholast@redhat.com>
Python 2 has keys()/values()/items(), which return lists,
iterkeys()/itervalues()/iteritems(), which return iterators,
and viewkeys()/viewvalues()/viewitems() which return views.
Python 3 has only keys()/values()/items(), which return views.
To get iterators, one can use iter() or a for loop/comprehension;
for lists there's the list() constructor.
When iterating through the entire dict, without modifying the dict,
the difference between Python 2's items() and iteritems() is
negligible, especially on small dicts (the main overhead is
extra memory, not CPU time). In the interest of simpler code,
this patch changes many instances of iteritems() to items(),
iterkeys() to keys() etc.
In other cases, helpers like six.itervalues are used.
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Jan Cholasta <jcholast@redhat.com>
The deprecated has_key method will be removed from dicts in Python 3.
For custom dict-like classes, has_key() is kept on Python 2,
but disabled for Python 3.
Reviewed-By: Tomas Babej <tbabej@redhat.com>
Make all DNs, RDNs and AVAs immutable.
Immutability makes reasoning about DN-handling code easier,
as value objects can't be changed once created.
Instead of mutable DNs, one can use a list (or even a generator)
of RDNs that's converted to a DN on output.
Reviewed-By: Jan Cholasta <jcholast@redhat.com>
DN code was optimized to be faster if DNs are created from string. This is
the major use case, since most DNs come from LDAP.
With this patch, DN creation is almost 8-10x faster (with 30K-100K DNs).
Second mojor use case - deepcopy in LDAPEntry is about 20x faster - done by
custom __deepcopy__ function.
The major change is that DN is no longer internally composed of RDNs and
AVAs but it rather keeps the data in open ldap format - the same as output
of str2dn function. Therefore, for immutable DNs, no other transformations
are required on instantiation.
The format is:
DN: [RDN, RDN,...]
RDN: [AVA, AVA,...]
AVA: ['utf-8 encoded str - attr', 'utf-8 encode str -value', FLAG]
FLAG: int
Further indexing of DN object constructs an RDN which is just an encapsulation
of the RDN part of open ldap representation. Indexing of RDN constructs AVA in
the same fashion.
Obtained EditableAVA, EditableRDN from EditableDN shares the respected lists
of the open ldap repr. so that the change of value or attr is reflected in
parent object.
Reviewed-By: Petr Viktorin <pviktori@redhat.com>
Drop support for pylint < 1.0
Enable ignoring unknown attributes on modules (both nose and pytest
use advanced techniques, support for which only made it to pylint
recently)
Fix some bugs revealed by pylint
Do minor refactoring or add pylint:disable directives where the
linter complains.
Reviewed-By: Tomas Babej <tbabej@redhat.com>
pytest's support for Nose-style test generators is not bulletproof;
use a real function to please it.
https://fedorahosted.org/freeipa/ticket/4610
Reviewed-By: Tomas Babej <tbabej@redhat.com>
The setUp/dearDown names are used in the unittest module, but there is no reason
to use them in non-`unittest` test cases.
Nose supports both styles (but mixing them can cause trouble when
calling super()'s methods).
Pytest only supports the new ones.
https://fedorahosted.org/freeipa/ticket/4610
Reviewed-By: Tomas Babej <tbabej@redhat.com>
Updating a CIDict with data like {'A': 1, 'a': 2} would lead to data
loss since only one of the items would get to the CIDict.
This can result in non-obvious bugs similar to this one in python-ldap:
https://bugzilla.redhat.com/show_bug.cgi?id=1007820
Raise an error in this case; any resolution must be done by the caller.
Make the CIDict interface match standard dict (except view* methods).
Add __contains__, __iter__, clear.
Add keyword and iterable support for __init__, update.
Also add values() and itervalues(). Previously the dict versions were
used; the new ones guarantee that the order matches keys().
Mark view* methods as not implemented.
CIDict.copy() now returns a CIDict.
Test the above additions, and fromkeys() which worked but wasn't tested.
Nose doesn't pick up directories that don't begin with 'test'.
Rename ipatests/test_ipaserver/install to test_install so that it's run.
Also, merge test_ipautil.py from ipapython/test into tests/test_ipapython,
so the whole test suite is in one place.
