Reusing Static Analysis across Different Domain-Specific Languages using Reference Attribute Grammars (Artefact)
This repository contains the source code for the publication: Johannes Mey, Thomas Kühn, René Schöne, and Uwe Aßmann. “Reusing Static Analysis across Different Domain-Specific Languages Using Reference Attribute Grammars.” Programming 4, no. 3 (February 17, 2020).
DOI of the publication: 10.22152/programming-journal.org/2020/4/15
A git repository containing this artefact is available at https://git-st.inf.tu-dresden.de/jastadd/reusable-analysis/
Running the Evaluation
Preparation for the Qualitas Corpus
Note, that for all parts involving Java, the Qualitas Corpus is required to be downloaded first. To do that:
- Visit http://qualitascorpus.com/docs/faq.html#download and follow the instructions there to download both parts of the archive.
- Unpack the archives.
- Create a symlink to
QualitasCorpus-$VERSION/Systems/namedqualitas(alternatively, or if docker does not follow symbolic links for security reasons, create a directory namedqualitasand copy all systems into it) - Create a directory
docker-resultswhich serves as a shared directory with the container.
Preparation for the Docker Image
Load the docker image using docker load --input reusable-analysis.tar
Running the Docker Container
Run the container using docker run --rm -it -v "$PWD"/docker-results:/reusable-analysis/benchmark:Z -v "$PWD"/qualitas:/reusable-analysis/qualitas:Z reusable-analysis
Inside the Container
For convenience, there are several scripts to execute parts of the evaluation found in the paper named after the respective part and start with run_.
The main evaluation is performed with run_scc_java.
Those scripts internally call the correct Gradle tasks.
Measured Data
The measurements were performed on an Intel i7-8700 workstation with 64GB of memory using Fedora Linux 29 running on kernel 4.18, OpenJDK version 1.8 and JastAdd version 2.3.
The data obtained using the provided artefact are contained in the file measured_results.csv.
This file is comma-separated file with the following columns:
| Title | Explanation | |
|---|---|---|
| 1 | Domain | language on which the analysis is performed |
| 2 | Analysis | kind of analysis (type or package) |
| 3 | Internal |
true if direct, false if reusable |
| 4 | JavaFiles | number of files analyzed |
| 5 | Nodes | number of nodes in the dependency graph |
| 6 | Edges | number of edges in the dependency graph |
| 7 | NodesAndEdges | number of elements in the dependency graph |
| 8 | SCCs | number of computed SCCs |
| 9 | FullTime | total time of the run (parse+generation+analysis) |
| 10 | ParseTime | parse time |
| 11 | GenerationTime | generation time of the problem-specific structure |
| 12 | AnalysisTime | analysis time |
| 13 | GenAnaTime | sum of generation and analysis time |
| 14 | Run | number of the run (0-100) |
| 15 | Scenario | name of the analyzed program |
| 16 | TimeTime | wall-clock time as taken by the time command |
| 17 | Exit | return value of the benchmark run (always 0) |