Implementing the Symbolic Method of Verification in the C-Light Project

Kondratyev, Dmitry A.

doi:10.1007/978-3-319-74313-4_17

Cited by 14 publications

(40 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Based on these methods, we conducted some experiments on error localization. We can also mention successful verification [6,8] of function asum from the interface BLAS. The more complex data structures and the other functions from BLAS will be considered in future work.…”

Section: Resultsmentioning

confidence: 98%

“…These strategies are the new contribution relative to previous research [6][7][8]11]. Based on these methods, we conducted some experiments on error localization.…”

Section: Resultsmentioning

confidence: 99%

“…In contrast to the original idea, the labels in our VCs form hierarchy more suitable to explanation generation [6]. We perform the depth search in the label tree and for each label its common text is expanded by corresponding pattern filled by line numbers.…”

Section: Methods Used In C-light Projectmentioning

confidence: 99%

“…The loop execution is modeled symbolically by replacement function rep. Some results of such adaptation were discussed in [6,8].…”

Section: Introductionmentioning

confidence: 99%

“…A natural language explanation can be generated from such protocol. The MetaVCG approach allowed us to easily introduce semantic labels in our axiomatic systems [6,7,12]. Now, after this background overview let us address the current problematic task, error localization for programs with loops.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Towards Automated Error Localization in C Programs With Loops

Kondratyev¹,

Promsky²

2019

Self Cite

View full text Add to dashboard Cite

The most recent trends in the C-light verification system are MetaVCG, semantic labels appropriate for verification condition (VC) explanation and symbolic method of definite iterations. MetaVCG takes a C-light program together with some Hoare's logic and produces on-the-fly a VC generator (VCG), which in turn processes the input program. Hoare's logic for definite iterations is a good choice if we try to get rid of loop invariants. Finally, if a theorem prover was unable to validate some VCs we could follow two ways. Obviously, we could revise/enrich specifications or/and underlying proof theory to prove the truth of VCs. Or, perhaps, we could concentrate upon establishment of falsity, which meant there were errors in annotated program. This is where semantic labels play crucial role providing some natural language comments about wrong VC as well as a back-trace to the error location. The newly developed ACL2 heuristics to prove VC falsity is the main theme of this paper.

show abstract

Section: Resultsmentioning

confidence: 98%

“…These strategies are the new contribution relative to previous research [6][7][8]11]. Based on these methods, we conducted some experiments on error localization.…”

Section: Resultsmentioning

confidence: 99%

Section: Methods Used In C-light Projectmentioning

confidence: 99%

“…The loop execution is modeled symbolically by replacement function rep. Some results of such adaptation were discussed in [6,8].…”

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations