Hi Xiao
1. If you use only symbolic arguments, use "klee-replay ./a.out
klee-last/*.ktest;" (a.out is the program under test compiled normally)
If you use klee_int or other methods, use "KTEST_FILE=xxx; ./a.out;"
(a.out needs to be compiled with the libraries under /Release/lib/)
2. I don't check the info in run.stats carefully, but it is convenient
to use gcov to collect the coverage information. You can write a simple
parser for the .gcov files. For overall coverage of files in a
directory, you can use the zcov tool.
Best wishes,
Wujie
Xiao Qu wrote:
>
> Hi,
>
> I have some questions on replaying test cases and collecting coverage
> information.
>
> 1. In Tutorial One and Three, there are two different methods to
> replay test cases. One is build with "libkleeRuntest library" while
> the other one is use "klee-replay" directly.
>
> I guess klee-replay only applies when the symbolic input is the
> "direct" input parameters for the main function?
>
> Plus, I didn't find libkleeRuntest.dylib in Release/lib ?
>
>
> 2. In order to collect coverage information, do I have to use gcov?
> Can we just get the similar information from "*run.stats* " ?
>
> Thanks,
> Xiao
> ------------------------------------------------------------------------
>
> _______________________________________________
> klee-dev mailing list
> klee-dev at keeda.stanford.edu
> http://keeda.Stanford.EDU/mailman/listinfo/klee-dev
>
