Re: [lmi] [PATCH] wx_test_calculation_summary.cpp

[Greg Chicares]

On 2015-03-08 23:53, Greg Chicares wrote:
> On 2015-03-08 01:09, Vadim Zeitlin wrote:
[...patch applied yesterday...]
>> AFAICS it does follow it faithfully, but there are 2 remaining
>> questions/potential problems: first one is the previously mentioned issue
>> with getting the default summary columns and is discussed in more details
>> below.

Addressed in email sent a few moments ago.

>>  The second one is similar to the previously discussed issues due to
>> external files modification in the default update test. Namely, with this
>> test, the "distribution only" part of it can pass at most once because it
>> relies on the fixed values of the "calculation_summary_columns" and
>> "use_builtin_calculation_summary" keys in the configuration file which are
>> changed by this test itself. So after running successfully once, the test
>> is _guaranteed_ to fail when run the next time with the "--distribution"
>> option. As previously stated, this might not be a problem in your workflow,
>> but I wanted to at least mention this because it does seem rather
>> user/tester-unfriendly to me.
> 
> I'll have to confer with Kim.

We like this. It fits our workflow perfectly. It's all right that it's
guaranteed to fail when we fail to follow our workflow.