[PATCH] libtest: Change expected fail state string

Chris Johns chrisj at rtems.org
Wed Dec 4 08:49:25 UTC 2019


On 4/12/19 5:43 pm, Sebastian Huber wrote:
> On 04/12/2019 05:14, Chris Johns wrote:
>> On 3/12/19 9:59 pm, Sebastian Huber wrote:
>>> On 02/12/2019 23:32, Chris Johns wrote:
>>>> On 2/12/19 6:24 pm, Sebastian Huber wrote:
>>>>> Use the value expected by the RTEMS Tester.  There are two separator
>>>>> characters used ('-' and '_') for the states.  This is a bit
>>>>> inconsistent.
>>>> Sorry about this. I had not noticed.
>>>>
>>>>> ---
>>>>>    cpukit/libtest/testbeginend.c | 2 +-
>>>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/cpukit/libtest/testbeginend.c b/cpukit/libtest/testbeginend.c
>>>>> index b739e47b2e..9bdc86780c 100644
>>>>> --- a/cpukit/libtest/testbeginend.c
>>>>> +++ b/cpukit/libtest/testbeginend.c
>>>>> @@ -29,7 +29,7 @@ rtems_printer rtems_test_printer = {
>>>>>    static const char* const test_state_strings[] =
>>>>>    {
>>>>>      "EXPECTED-PASS",
>>>> Should this change or should `USER_INPUT` be changed and we update ...
>>>>
>>>> https://git.rtems.org/rtems-tools/tree/tester/rt/report.py#n159
>>>> https://git.rtems.org/rtems-tools/tree/tester/rt/report.py#n161
>>>>
>>>> ?
>>>>
>>>> It would be a simpler change because this change means we move to having 2 with
>>>> `_` and a bunch with `-`. There are internal states in the tester such as
>>>> `WRONG-VERSION` and changing all of those is a bigger change in the tester.
>>>
>>> If we update the tester, then it should accept '-' and '_' versions so that it
>>> can be used for test runs before and after a change of
>>> cpukit/libtest/testbeginend.c.
>>
>> Yes this is a good point. Shame it has to be this way but I cannot see an
>> alternative.
> 
> I checked in the patch as is. Using only one separator character for all states
> is just a beautification.

That is not what I was meaning. I would like RTEMS master to use only '-' and
the tester has backward compatibility.

Chris


More information about the devel mailing list