This release consists of binaries (Debian/LinuxMint), sources, and resulting data
from evaluation of the approach "Anonymous On-line Communication Between
Program Analyses" on SV-COMP 2015 benchmark suite, revision 571.
Installation is very simple. Download the ZIP package, unzip it into some
directory, and the tool is ready for an execution. The tool is located in
STATOR-tool sub-directory, the result from our evaluation in
SVCOMP-repo.ORIG.EVAL sub-directory. Note that SV-COMP benchmarks are
not included in the package. They have to be downloaded separately. List of
all used benchmarks can be found in the file BENCHLIST_SELECTION.txt.
Use Python script STATOR-tool/start.py to run the tool on a single C
file you have to execute . Use the option --help to list detailed info
of usage. In order to run the evaluation for several benchmarks you have to
execute Python script STATOR-tool/start_evaluation.py. Use the option
--help to list detailed info of usage. In all cases input C file(s)
have to always be preprocessed.
In case you want to recompute results of our evaluation we prepared a dedicated
shell script STATOR_SVCOMP_evaluate.sh located in the root directory
which calls STATOR-tool/start_evaluation.py with our settings. Note
the shell script assumes that SV-COMP repository was cloned into sub-directory
SVCOMP-repo (which is empty in our package).