[Freeipa-devel] [PATCH 0042] Make lint work on Fedora 22.

[David Kupka]

On 04/08/2015 10:23 AM, Jan Cholasta wrote:
> Dne 8.4.2015 v 10:22 David Kupka napsal(a):
>> On 04/06/2015 02:48 PM, Simo Sorce wrote:
>>> On Mon, 2015-03-30 at 12:15 +0200, David Kupka wrote:
>>>> On 03/30/2015 07:12 AM, Jan Cholasta wrote:
>>>>> Dne 28.3.2015 v 00:05 Petr Vobornik napsal(a):
>>>>>> On 27.3.2015 14:58, David Kupka wrote:
>>>>>>> pylint changed slightly so we must react otherwise we'll be
>>>>>>> unable to
>>>>>>> build freeipa rpms on Fedora 22. This patch should go to master
>>>>>>> for sure
>>>>>>> but I don't know if we want it in 4.1.
>>>>>>>
>>>>>>
>>>>>> ACK
>>>>>
>>>>> Are all the new disables really just false positives?
>>>>
>>>> It seems to me as a false positives.
>>>>
>>>> 1. ipalib/plugins/otptoken.py:552: [E1101(no-member),
>>>> otptoken_sync.forward] Module 'ssl' has no 'PROTOCOL_TLSv1' member)
>>>>
>>>>   >>> import ssl
>>>>   >>> ssl.PROTOCOL_TLSv1
>>>> 3
>>>>
>>>> 2. ipaserver/install/ipa_otptoken_import.py:63: [E1101(no-member),
>>>> convertDate] Instance of 'tuple' has no 'tzinfo' member)
>>>> ipaserver/install/ipa_otptoken_import.py:64: [E1101(no-member),
>>>> convertDate] Instance of 'tuple' has no 'timetuple' member)
>>>>
>>>> dateutil.parser.parse() returns datetime.datetime object and it has
>>>> both tzinfo and timetuple methods
>>>> (https://docs.python.org/2/library/datetime.html#datetime-objects)
>>>>
>>>> 3. ipapython/dnssec/ldapkeydb.py:26: [E1127(invalid-slice-index),
>>>> uri_escape] Slice index is not an int, None, or instance with
>>>> __index__)
>>>>
>>>> This is the line lint is complaining about:
>>>> out += '%'.join(hexval[i:i+2] for i in range(0, len(hexval), 2))
>>>> I don't see a chance for 'i' or 'i+1' to be anything else than
>>>> integers.
>>>>
>>>>>
>>>>>>
>>>>>> tested on:
>>>>>> - F21: ipa-4-1, master branch
>>>>>> - F22: master branch.
>>>>>>
>>>>>> IMHO it could got to ipa-4-1 branch because of FreeIPA 4.1.4 in F22
>>>>>
>>>
>>> This patch doesn't seem to fix all my issues building on F22, so
>>> tentative NACK.
>>
>> I tested it this way:
>> 1. started with Fedora-22-x86_64-minimal system
>> 2. dnf install git
>> 3. clone freeipa
>> 4. make version-update # to get freeipa.spec
>> 5. dnf install `awk '/^BuildRequires/ {print $2}' freeipa.spec`
>> 6. ./make-lint
>>
>>>
>>> It seem the main offenders are "No value for argument 'second' in method
>>> call" (this one only in test_ipautul.py) and "No value for argument
>>> 'extClass' in method call" sprinkled around various test plugins.
>>> These cause E1120(no-value-for-parameter).
>>
>> Could you please paste the output of make-lint somewhere?
>>>
>>> On a different note, make-lint takes forever to run, do we really need
>>> to run it in make rpms ? Shouldn't we rather just run it at make dist
>>> time, or whatever we use to generate the release tarballs ?
>>
>> It really takes forever and it would be nice to move it out of
>> Makefile's rpms target but I don't know if we can do it without breaking
>> anything.
>
> Well, you can always "make rpms DEVELOPER_MODE=1" to skip make-lint.

Quick "git grep" shows that it only passes "--no-fail" to make-lint so 
the process always succeeds but it takes the same amount of time.

>
>>
>>>
>>> Simo.
>>>
>>
>
>


-- 
David Kupka