coverage : error when running --coverage with rtems-test for whole testsuites

[Vijay Kumar Banerjee]

On 21 July 2018 at 03:49, Vijay Kumar Banerjee <vijaykumar9597 at gmail.com>
wrote:

> On Fri, Jul 20, 2018, 10:08 PM Joel Sherrill <joel at rtems.org> wrote:
>
>>
>>
>> On Fri, Jul 20, 2018 at 9:14 AM, Gedare Bloom <gedare at rtems.org> wrote:
>>
>>> On Thu, Jul 19, 2018 at 6:29 PM, Vijay Kumar Banerjee
>>> <vijaykumar9597 at gmail.com> wrote:
>>> > Hello,
>>> >
>>> > I used the following command
>>> >
>>> > ====================
>>> > $HOME/development/rtems/test/rtems-tools/tester/rtems-test \
>>> > --rtems-tools=$HOME/development/rtems/5 --log=coverage_analysis.log \
>>> > --no-clean --coverage=score --rtems-bsp=leon3-qemu-cov \
>>> > /home/lunatic/development/rtems/kernel/leon3/sparc-
>>> rtems5/c/leon3/testsuites
>>> > --debug-trace=cov
>>> > ====================
>>> >
>>> > and I'm getting the following error
>>> >
>>> > ===================
>>> > error: coverage: covoar failure:: -9
>>> > ===================
>>> >
>>>
>>> What does error code -9 mean from covoar?
>>>
>>> Does it make any progress at all?
>>>
>>
>> Can you capture the command line used to invoke covoar and
>> run it in gdb? That's an odd error message. covoar is usually
>> good at printing something useful. It was written to be paranoid.
>>
> yes I'm trying to run it in gdb, the laptop
> freezes for a long time and I have to
> restart it, I'll give it more time tommorow
> and let it take as much time as it wants
>
> It runs fine for samples/ so we're searching
> for a small test case where it trips, so
> I'll run it separately for each set of tests,
> like benchmarks/ , fstests/ ...
>
>
>>
>>>
>>> > This, however, runs fine for samples/
>>> >
>>>
>>> does this command work for you without using the cov options?
>>>
>>> > I think this will probably be on hold as Chris seems to be
>>> > on a break, meanwhile, I want to do a wrapup work on
>>> > the non-gcov coverage reports, I seek suggestions/advice
>>> > for the same.
>>> >
>>> > The current state is that the coverage reports can be generated
>>> > for one symbol-set only, There's a ticket for the support of
>>> > generating separate reports of multiple sets from covoar.
>>> >
>>> > https://devel.rtems.org/ticket/3441
>>
>>
>> I thought originally that the Python would invoke covoar multiple
>> times. It was slower but that was how it was originally designed.
>> Is there no support in the Python for doing this?
>>
> I was building it in a way that the script invokes
> covoar multiple times, but Chris' modifications
> to covoar added support for reading multiple symbol-sets
> after which I changed the coverage script.
>
> If we're planning to have this support in
> the python script untill covoar is updated,
> then I can add it.
>
I managed to get this work, please see the attached image.
If we're looking to get this support in the script, I'll send the patch. :)

>
>
>>
>>>
>>> >
>>> > Please let me know of any suggestions including suggestions
>>> > regarding documentation as Coverage needs more
>>> > documentation.
>>> >
>>>
>>> What is the existing documentation for coverage?
>>>
>>
>> That's an important part of reproducability.
>>
>>
>>>
>>> > Thanks
>>> > -- vijay
>>> >
>>> > _______________________________________________
>>> > devel mailing list
>>> > devel at rtems.org
>>> > http://lists.rtems.org/mailman/listinfo/devel
>>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20180723/b29789c7/attachment-0002.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screenshot from 2018-07-23 03-12-01.png
Type: image/png
Size: 85442 bytes
Desc: not available
URL: <http://lists.rtems.org/pipermail/devel/attachments/20180723/b29789c7/attachment-0002.png>