|
| 1 | +These tasks were generated by Martin Spiessl |
| 2 | +from information in the tool repository of zilu, |
| 3 | +a tool for invariant generation throug selective sampling, |
| 4 | +which is described in the following publication: |
| 5 | + |
| 6 | +Li, J., Sun, J., Li, L., Le, Q. L., & Lin, S. W. (2017, October). |
| 7 | +Automatic loop-invariant generation anc refinement through selective sampling. |
| 8 | +In 2017 32nd IEEE/ACM International Conference on Automated Software Engineering (ASE) (pp. 782-792). IEEE. |
| 9 | + |
| 10 | +These tasks represent the 53 benchmarks |
| 11 | +that are listed in table 1 in this publication. |
| 12 | + |
| 13 | +The original repository for zilu was at (according to the paper): |
| 14 | +https://github.com/lijiaying/zilu |
| 15 | + |
| 16 | +It seems this repository was deleted or set to private. |
| 17 | +The original repository was published unter GPL-2.0, |
| 18 | +a copy I made can be found at: |
| 19 | +https://github.com/MartinSpiessl/zilu |
| 20 | + |
| 21 | +So I supply these tasks unter GPL-2.0 as well. |
| 22 | + |
| 23 | +The paper mentions that the tasks were collected |
| 24 | +from various papers/sources, one of them being SV-COMP. |
| 25 | +It is not easy to identify which tasks were taken from which source, |
| 26 | +and the authors state many of the SV-COMP tasks had to be excluded anyway |
| 27 | +because they were too complicated/ out of scope. |
| 28 | +Also the tasks are written down in a simplified format |
| 29 | +in the zilu repository in special files with the extension "cfg". |
| 30 | + |
| 31 | +I did the translation into SV-COMP compliant tasks |
| 32 | +myself with a small python script. |
| 33 | +Due to these 2 transformations, |
| 34 | +I do not believe that the tasks contain |
| 35 | +any copyrightable material anymore |
| 36 | +from the publications they were originally taken from. |
| 37 | +Also the tasks are very simple, |
| 38 | +so they will probably fall below the |
| 39 | +threshold of originality anyway. |
0 commit comments