Lucas,

Fair enough - make sense not to apply this.

I am not sure though with tighter db schema if add_noncompliant=True still 
makes sense; perhaps to ponder later

Julius

-----Original Message-----
From: Lucas Meneghel Rodrigues [mailto:[email protected]] 
Sent: Monday, March 11, 2013 1:42 PM
To: Gawlas, Julius
Cc: [email protected]
Subject: Re: [Autotest] [PATCH] Fixes test importer to properly create new 
tests with test_type set

On Wed, Mar 6, 2013 at 10:18 PM, Lucas Meneghel Rodrigues <[email protected]> 
wrote:
> On 03/06/2013 08:21 PM, Julius Gawlas wrote:
>>
>> Since test_type is not nullable when creating new object we need to 
>> specify it; in the previous version we just set the path. Also move 
>> up the validation for test type.
>
>
> Julius, I've made some fixes to the install script pushed them to 
> next, and synced next and master.
>
> I've tested the install on a Fedora 17 VM. The install went flawless 
> and even the test importer worked perfectly, so as far as I understand 
> the patch is unnecessary.
>
> Now, I need to go over your patch and validate the changes. I'll do 
> this tomorrow.
>
> Cheers,
>
> Lucas
>
>
>> Signed-off-by: Julius Gawlas <[email protected]>
>> ---
>>   utils/test_importer.py |   37 ++++++++++++++++++-------------------
>>   1 files changed, 18 insertions(+), 19 deletions(-)
>>
>> diff --git a/utils/test_importer.py b/utils/test_importer.py index 
>> c1d8d8c..b1b4027 100755
>> --- a/utils/test_importer.py
>> +++ b/utils/test_importer.py
>> @@ -49,6 +49,9 @@ class
>> TestImporterLoggingConfig(logging_config.LoggingConfig):
>>   DRY_RUN = False
>>   DEPENDENCIES_NOT_FOUND = set()
>>
>> +test_type = {'client': 1,
>> +             'server': 2, }
>> +
>>
>>   def update_all(autotest_dir, add_noncompliant, add_experimental):
>>       """
>> @@ -228,12 +231,20 @@ def update_tests_in_db(tests, dry_run=False, 
>> add_experimental=False,
>>                               '%s not in %s, did you forget to use -z 
>> option?' %
>>                               (test, autotest_dir))
>>
>> +        logging.info("Processing %s", test)
>> +
>> +        # create test object, first validate test type
>> +        data = tests[test]
>> +        if not hasattr(data, 'test_type'):
>> +            raise Exception('Test type is required')
>> +        if not data.test_type.lower() in test_type:
>> +            raise Exception('Incorrect value %s for test_type' %
>> +                            data.test_type)

Although the changes are good, doing this so early here breaks the 
add_noncompliant=True case. As these are not strictly necessary, given that the 
script works on latest next/master, I won't apply this one.

Sorry about that, and thanks for your time dedicated on this issue.

Lucas

_______________________________________________
Autotest-kernel mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/autotest-kernel

Reply via email to