[IGSTK-Developers] Re: ITK Logger Bug

[Luis Ibanez]

Hi Patrick,

Thanks for pointing this out.

Andinet is in vacations this week.

We should be able to fix the submissions
from his machine (Tigris) early next week.


   Thanks


      Luis


----------------------
Patrick Cheng wrote:
> Hi Luis,
> 
> I am trying to update the ITK on the windows build machine here, so we 
> can see the results on dashboard.
> 
> I noticed that most of the tests on "tigris.­kitware" failed in the past 
> few days. The version of the ITK on that machine is May 11, 2007, I 
> think it does not take the fix of the missing "break", so it's calling 
> the itksys::SystemTools::GetCurrentDateTime("%Y %b %d %R %S"), which 
> causes most of the tests to fail.
> 
> We should:
> 1. update the ITK code on all build machines
> 2. add a test case in IGSTK,
>    SetTimeStampFormat( itk::LoggerBase::HUMANREADABLE );
>    So we can see whether this new feature works or not.
> 
> Could you help take a look at this?
> 
> Thank you,
> 
> Patrick
> 
> 
> 
> Luis Ibanez wrote:
> 
>>
>> Hi Patrick,
>>
>> Thanks for pointing this out.
>>
>> You are right, "break"s were missing
>> from the switch statement.
>>
>> A fix has been committed now.
>>
>> A test case was added to itkLoggerTest.cxx.
>>
>> I couldn't reproduce the problem you reported
>> with the Exception. Maybe we should experiment
>> with other combinations of the format string.
>>
>>
>>
>>    Luis
>>
>>
>>
>> -----------------------
>> Patrick Cheng wrote:
>>
>>> Hi Luis,
>>>
>>> Thank you for the fix. However I found two problems:
>>>
>>> 1) in itkLoggerBase.cxx line 73, there should be a "break"
>>>
>>> 2) My system doesn't seem to support this call:
>>>    itksys::SystemTools::GetCurrentDateTime("%Y %b %d %R %S");
>>>    It throws an exception as in attached image.
>>>
>>> Patrick
>>>
>>> Luis Ibanez wrote:
>>>
>>>>
>>>> Hi Patrick,
>>>>
>>>>
>>>> 1)  Thanks for pointing out this bug in ITK.
>>>>     A fix has now been committed for it.
>>>>
>>>>
>>>> 2) The functionality of writing out the time
>>>>    in a human-readable format has also been
>>>>    added to the itkLoggerBase class. You will
>>>>    have to update your ITK checkout in order
>>>>    to get this functionality.
>>>>
>>>>
>>>>    You should call in your logger the method:
>>>>
>>>>       logger->SetTimeStampFormat( itk::LoggerBase::HUMANREADABLE );
>>>>
>>>>    to get the time stamp in the format
>>>>    year, month, day, hours, minutes, seconds.
>>>>
>>>>    or you could call
>>>>
>>>>       logger->SetTimeStampFormat( itk::LoggerBase::REALVALUE );
>>>>
>>>>    in order to get the standard behavior, printing the
>>>>    time stamp in milliseconds.
>>>>
>>>>
>>>>
>>>> If you have a chance,
>>>> please give it a try and let us know if you find any problem.
>>>>
>>>>
>>>>
>>>>     Thanks
>>>>
>>>>
>>>>        Luis
>>>>
>>>>
>>>>
>>>> --------------------
>>>> Patrick Cheng wrote:
>>>>
>>>>> Hi Luis and Andinet,
>>>>>
>>>>> There is a minor bug in itkLoggerBase.cxx line 62
>>>>> I believe the third string should be "CRITICAL" instead of "ERROR" 
>>>>> according to the enum.
>>>>>
>>>>> Could you guys also add the feature of logging formated time early 
>>>>> next week? I really appreciate it.
>>>>>
>>>>> 4989 igstkLogMacro() & TimeStamp should output time in brokendown 
>>>>> date/time, instdead of milliseconds. Have both options, 
>>>>> milliseconds for easy computation, and broken down time for 
>>>>> readability
>>>>>
>>>>> Thank you.
>>>>>
>>>>> Patrick
>>>>>
>>>>
>>>>
>>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>
>>
>>
>