error while running rtems-test with coverage

[Chris Johns]

On 14/4/18 8:02 am, Vijay Kumar Banerjee wrote:
> Getting this error now seems like something is missing.
> 
> ........
> 
> RTEMS Testing - Tester, 5 (b2ba08b0da42 modified)
> Coverage environment prepared
> Traceback (most recent call last):
>    File 
> "/home/lunatic/development/rtems/test/rtems-tools/tester/rtems-test", 
> line 40, in <module>
>      rt.test.run()
>    File 
> "/home/lunatic/development/rtems/test/rtems-tools/tester/rt/test.py", 
> line 353, in run
>      tst.reraise()
>    File 
> "/home/lunatic/development/rtems/test/rtems-tools/tester/rt/test.py", 
> line 160, in reraise
>      _test_reraise(*self.result)
> NameError: global name '_test_reraise' is not defined
> 

The reraise logic was moved to the rtemstoolkit last year:

https://git.rtems.org/rtems-tools/commit/?id=fa81491764861e63785a3faa241c488469087209

It had some issues that have been fixed and it is being reused else 
where in the tester.

> 
> -- vijay
> 
> On 13 April 2018 at 23:57, Vijay Kumar Banerjee 
> <vijaykumar9597 at gmail.com <mailto:vijaykumar9597 at gmail.com>> wrote:
> 
> 
> 
>     On 13 April 2018 at 23:51, Cillian O'Donnell <cpodonnell8 at gmail.com
>     <mailto:cpodonnell8 at gmail.com>> wrote:
> 
>         Hmmm it's not there. Even so I can almost guarantee you, the
>         reason it's not running is that coverage_supported variable is
>         not being found as true in qemu.cfg and so the the command to
>         generated the trace files isn't run and there's no .cov files in
>         the end. So that's what you'll have to dig into.
> 
>     Okay , I'll set it to global then.
> 

Changes need to be on the latest version of master before I will review 
them. The failure above indicate the code you have is not up to date.

The change you have made is generating an exception in the tester 
thread. If you want to continue down this path you will need to fix what 
ever you have added.

The tests are run as threads separate from the main thread so the 
exception is caught and re-raised in the main thread letting you know. 
This is broken in the old code base you are using.

Why not investigate the reason as indicated and sort if out what the new 
format configuration file needs?

Chris