Re: [gpsd-dev] Test coverage

[Christian Gagneraud]

Hi all,

Any comments on this?

I find it really useful to see what kind of code coverage you get with 
the test-suites.
For example,from [1], you can clearly see that the broadcast case of 
type 22, the long type 16 case, the "trim spaces on right end" code are 
all not tested. (and plenty of other things as well).

Chris

[1] 
http://www.techworks.ie/pub/chris/gpsd/coverage/aivdm/gpsd/gpsd.git/driver_ais.c.gcov.html

On 27/05/12 02:03, Christian Gagneraud wrote:
> Hi,
>
> Today, I gave a try to gcov and lcov, and the results are encouraging I
> think, this is what I've done:
> - 1 coverage report per testsuite run
> - 1 coverage report with all the runs
>
> The results when run on a x86 machine can be found here:
> http://www.techworks.ie/pub/chris/gpsd/coverage
>
> This is experimental stuff, I'm sure there's some problems here and there.
>
> For the "all" testsuite, I think it is possible to merge all the
> testcases belonging to each other testsuite into a corresponding single
> testcase.
> Attached is the diff to SConstruct. After banging my head against the
> wall, I stopped trying to find a nice way to add conditionally
> "--coverage", to LDFLAGS, CFLAGS and CCFLAGS, and went for a less subtle
> one...
> I had to change the calls to mktemp as it was complaining about not
> having enough X ...
> As well, the coveraging calls is unconditional, this is still
> experimental, thought.
>
> Any comments welcome!
>
> Chris


--
Christian Gagneraud,
Embedded systems engineer.
Techworks Marine
1 Harbour road
Dun Laoghaire
Co. Dublin
Ireland
Tel: + 353 (0) 1 236 5990
Web: http://www.techworks.ie/