Using trace in tests

[James McKenzie]

James McKenzie wrote:
> Charles Davis wrote:
>   
>> James McKenzie wrote:
>>   
>>     
>>> Charles Davis wrote:
>>>     
>>>       
>>>> James McKenzie wrote:
>>>>   
>>>>       
>>>>         
>>>>> All:
>>>>>
>>>>> I'm asking a question again, don't 'shoot' me, but I've noticed the use
>>>>> of the trace() function in tests and I would like to use this to
>>>>> troubleshoot what is incorrect in the EM_FORMATRANGE patch as it is not
>>>>> passing on MacOSX builds.  The test appears to pass on WindowsXP sp2 as
>>>>> the failures did not appear in the winetest_latest for Wine-1.1.33.
>>>>>
>>>>> So, how do I enable the trace's I've added for testing so that I can fix
>>>>> the patch?
>>>>>     
>>>>>         
>>>>>           
>>>> They should be automatically enabled when you run 'make test' in
>>>> dlls/riched20/tests. Conversely, they aren't enabled when you run
>>>> winetest.exe from Wine. You can forcibly enable them by setting
>>>> WINETEST_DEBUG=1 in the environment.
>>>>
>>>>   
>>>>       
>>>>         
>>> Chip:
>>>
>>> Neither of these worked.  I'm not getting the outputs I expected to see.
>>>     
>>>       
>> That's odd. According to the definition of the winetest_trace() function
>> (which is used to implement the trace() macro), when the debug level is
>> greater than 0 (which it is by default, and it should be if you set
>> WINETEST_DEBUG to anything other than 0), it prints a message. Are you
>> sure you used trace() and not TRACE()? There is a difference.
>>
>>   
>>     
> Chip:
>
> I'm certain I used trace() as TRACE() caused the build to fail with an
> error that __TRACE was not defined.
>
> I'll try it again with WINETEST_DEBUG=1 make test  and see what happens.
>
>   
Ran through the include/wine/test.h code and found that the variable has
to be LOWER case. 

Ran the test and I got the results that were reported.  Now to do some
code work.

James McKenzie