Reproduction Package for the ICSE 2023 Article `CoVeriTeam Service: Verification as a Service'

This reproduction package is a virtual machine (VM) to host CoVeriTeam Service described in article CoVeriTeam Service: Verification as a Service. Abstract. The research community has developed numerous tools for solving verification problems, but we are missing a common interface for executing them. This means users have to spend considerable effort on the installation and parameter setup, for each new tool (version) they want to execute. The situation could make a verification researcher wanting to experiment with a new verification tool turn away from it. We aim to make it easier for users to execute verification tools, as well as provide mechanism for tool developers to make their tools easily accessible. Our solution combines a web service and a common interface for verification tools. Contents of this archive |-- LICENSE.txt : specifies the license |-- README.txt : this file |-- coveriteam-service/ : folder containing CoVeriTeam Service. It has been created from the repository: https://gitlab.com/sosy-lab/software/coveriteam-service |-- coveriteam-service-vm.ova: A virtual machine with CoVeriTeam-Service installed and running CoVeriTeam Service Virtual Machine The provided Artifact coveriteam-service-vm.ova is a virtual machine with the service pre-installed. The virtual machine was created with vagrant; the username and password are vagrant. We recommend to run the VM in headless mode: You do not need to log into the VM, everything can be done from your host machines command line once the vm is running. The VM runs Ubuntu22.04 with 2 cores and 5 Gb of memory. CoVeriTeam Service is included as the systemd service coveriteam-service. To run the artifact, import it with VirtualBox and launch it. The VM is configured to forward CoVeriTeam Service to 127.0.0.1:5000 on the host machine. Trying CoVeriTeam Service You can open the overview page in your browser under http://127.0.0.1:5000. To invoke the Service you need to make a POST request to http://127.0.0.1:5000/execute. Running an example: Method 1) logging into the VM: Log into the VM - username and password are vagrant cd coveriteam-service/doc ./send_request.sh http://127.0.0.1:5000/execute sudo apt install unzip unzip cvt_remote_output.zip -d cvt-output cat cvt-ouput/LOG Method 2) from your own machine; The commands assumed to start at the root of the unpacked provided zip archive cd coveriteam-service/doc ./send_request.sh http://127.0.0.1:5000/execute unzip cvt_remote_output.zip -d cvt-output cat cvt-ouput/LOG Execution using cURL This section provides details about what is happening in the send_request.sh script for interested readers. You can use cURL or any other REST client to call the service. The service expects a multipart/form-data encoded POST request with the form field args containing a json string containing the inputs to CoVeriTeam and all the files with the names matching how they are referenced in args. All of the data for the following example can be found in coveriteam-service/doc. Example: example_request.json: { "coveriteam_inputs": { "verifier_path": "../actors/cpa-seq.yml", "program_path": "../../sv-benchmarks/c/ldv-regression/test02.c", "specification_path": "../../sv-benchmarks/c/properties/unreach-call.prp", "verifier_version": "default", "data_model": "ILP32" }, "cvt_program": "verifier.cvt", "working_directory": "coveriteam/examples", } curl -X POST -H "ContentType: multipart/form-data" -k \ --form args="$(cat example_request.json)" \ --form "../actors/cpa-seq.yml"=@data/cpa-seq.yml \ --form "../../sv-benchmarks/c/ldv-regression/test02.c"=@data/test02.c \ --form "../../sv-benchmarks/c/properties/unreach-call.prp"=@data/unreach-call.prp \ --form "verifier.cvt"=@data/verifier.cvt \ --form "../actors/verifier_resource.yml"=@data/verifier_resource.yml \ --output cvt_remote_output.zip \ 127.0.0.1:5000/execut