[PATCH] testsuite: Use printk for all test output where possible.

[Chris Johns]

On 19/10/17 5:22 pm, Chris Johns wrote:
> On 19/10/17 4:59 pm, Sebastian Huber wrote:
>> On 19/10/17 07:45, Chris Johns wrote:
>>> - Control the test's single init for functions and global data with
>>>    TEST_INIT and not CONFIGURE_INIT. They are now separate.
>>
>> Could we move the implementation of this stuff to cpukit/libmisc/testsupport?
>> This would avoid the need for this TEST_INIT define.
>>
> 
> That will be round 2. I did not want to step to far in a patch. If this gives
> good results with wider testing I can move to doing this.
> 

Test results are show too many failures:

Passed:        535
Failed:         17
User Input:      4
Expected Fail:   0
Indeterminate:   0
Benchmark:       3
Timeout:         3
Invalid:         3
------------------
Total:         565

Looking at block06.exe I see there is a local printf function using the stdout
stream directly and fflush. A grep for fflush in the test suite shows approx. 32
more fflush calls that need to looked into. The fflush call cannot be defined
away because it might be used in a real test.

Test result details:

Failures:
 psxtmsem04.exe.img
 sperror01.exe.img
 sperror02.exe.img
 psxtmcond03.exe.img
 fsjffs2gc01.exe.img
 rhdeadlockbrk.exe.img
 psxtmmutex05.exe.img
 psxtmcond06.exe.img
 psxsignal04.exe.img
 sperror03.exe.img
 sp51.exe.img
 psxsignal03.exe.img
 block06.exe.img
 psxtmcond07.exe.img
 psxtmsem05.exe.img
 psxtmrwlock06.exe.img
 psxtmrwlock07.exe.img
User Input:
 monitor.exe.img
 termios.exe.img
 fileio.exe.img
 top.exe.img
Benchmark:
 linpack.exe.img
 dhrystone.exe.img
 whetstone.exe.img
Timeouts:
 spintrcritical15.exe.img
 spintrcritical07.exe.img
 dl02.exe.img
Invalid:
 sptimecounter01.exe.img
 spinternalerror01.exe.img
 spsysinit01.exe.img
Average test time: 0:00:10.244275
Testing time     : 1:36:28.015860

Chris