[Freeipa-devel] [PATCH 0070] Remove hardcoded values from idrange plugin tests

[Alexander Bokovoy]

On Wed, 19 Jun 2013, Tomas Babej wrote:
> On 06/11/2013 01:15 PM, Tomas Babej wrote:
>> On 06/11/2013 12:59 PM, Alexander Bokovoy wrote:
>>> On Tue, 11 Jun 2013, Tomas Babej wrote:
>>>> On 06/10/2013 01:13 PM, Tomas Babej wrote:
>>>>> Hi,
>>>>> 
>>>>> Hardcoded values for range parameters such as base RID or range
>>>>> size could be the reason the tests produced incorrect results,
>>>>> as the ranges could get in conflict with already existing ranges
>>>>> on the server.
>>>>> 
>>>>> Patch dynamically chooses ID and RID range space at the end of
>>>>> all ranges already present on the server.
>>>>> 
>>>>> https://fedorahosted.org/freeipa/ticket/3662
>>>>> 
>>>>> Tomas
>>>> 
>>>> Patch altered to incorporate minor fixes for recent idrange 
>>>> objectclass changes.
>>>> 
>>>> Tomas
>>> 
>>>> From b35b10f1356c9714776f16aadec7ffbe95e2f41e Mon Sep 17 00:00:00 2001
>>>> From: Tomas Babej <tbabej at redhat.com>
>>>> Date: Mon, 10 Jun 2013 13:08:50 +0200
>>>> Subject: [PATCH] Remove hardcoded values from idrange plugin tests
>>>> 
>>>> Hardcoded values for range parameters such as base RID or range
>>>> size could be the reason the tests produced incorrect results,
>>>> as the ranges could get in conflict with already existing ranges
>>>> on the server.
>>>> 
>>>> Patch dynamically chooses ID and RID range space at the end of
>>>> all ranges already present on the server.
>>>> 
>>>> https://fedorahosted.org/freeipa/ticket/3662
>>>> ---
>>>> ipalib/plugins/idrange.py              |  2 +-
>>>> tests/test_xmlrpc/test_range_plugin.py | 90 
>>>> ++++++++++++++++++++++------------
>>>> 2 files changed, 60 insertions(+), 32 deletions(-)
>>>> 
>>>> diff --git a/ipalib/plugins/idrange.py b/ipalib/plugins/idrange.py
>>>> index abca492978d04c71b78a89df8e5c2d1d51c06398..54b835e244fb60ee212a9c00223d4294ff8f4363 
>>>> 100644
>>>> --- a/ipalib/plugins/idrange.py
>>>> +++ b/ipalib/plugins/idrange.py
>>>> @@ -224,7 +224,7 @@ class idrange(LDAPObject):
>>>>         if not any((options.get('pkey_only', False),
>>>>                     options.get('raw', False))):
>>>>             range_type = entry_attrs['iparangetype'][0]
>>>> -            entry_attrs['iparangetype'] = 
>>>> self.range_types.get(range_type, None)
>>>> +            entry_attrs['iparangetype'] = 
>>>> [self.range_types.get(range_type, None)]
>>>> 
>>>>         # Remove the objectclass
>>>>         if not keep_objectclass:
>>> Could you please extract this change into an independent patch? I'm
>>> thinking purely from possible backporting perspective.
>>> 
>>> Otherwise looks good.
>> 
>> Sure. Patches 0070 and 0071 attached.
>> 
>> I'll link 0071 to the ticket for extending ID range types once 
>> it's pushed, for record's sake.
>> 
>> Tomas
>> 
> Patches needed rebase.
The tests now pass on a machine with existing trusts.

However, I'm getting following errors in dirsrv's error log:

[19/Jun/2013:15:34:58 +0300] find_sid_for_ldap_entry - [file ipa_sidgen_common.c, line 522]: Cannot convert Posix ID [1447902850]
into an unused SID.
[19/Jun/2013:15:34:58 +0300] ipa_sidgen_add_post_op - [file ipa_sidgen.c, line 149]: Cannot add SID to new entry.
[19/Jun/2013:15:34:59 +0300] find_sid_for_ldap_entry - [file ipa_sidgen_common.c, line 522]: Cannot convert Posix ID [1447902950]
into an unused SID.
[19/Jun/2013:15:34:59 +0300] ipa_sidgen_add_post_op - [file ipa_sidgen.c, line 149]: Cannot add SID to new entry.
[19/Jun/2013:15:35:01 +0300] ipa_range_check_pre_op - [file ipa_range_check.c, line 417]: New primary rid range overlaps with
existing primary rid range.
[19/Jun/2013:15:35:01 +0300] ipa_range_check_pre_op - [file ipa_range_check.c, line 417]: New secondary rid range overlaps with
existing secondary rid range.
[19/Jun/2013:15:35:01 +0300] ipa_range_check_pre_op - [file ipa_range_check.c, line 417]: New primary rid range overlaps with
existing secondary rid range.
[19/Jun/2013:15:35:01 +0300] ipa_range_check_pre_op - [file ipa_range_check.c, line 417]: New base range overlaps with existing base
range.

I think we still need to improve RID part of calculating the test range..

-- 
/ Alexander Bokovoy