Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a 16-minute conference presentation from OOPSLA 2025 that introduces a novel program repair technique for fixing incomplete input generators in property-based testing. Learn how researchers from Purdue University and IIT Hyderabad developed an approach to automatically patch generators that fail to produce all valid inputs meeting specified preconditions. Discover the innovative enumerative synthesis algorithm that leverages coverage types to identify missing test values and evaluate potential repairs. Examine the implementation of Cobb, a repair tool for OCaml generators, and review its performance on benchmarks from the property-based testing literature. Understand how this type-guided repair method addresses the common problem of handwritten generators being unable to generate complete sets of valid inputs, ultimately improving the effectiveness of automated testing for programs with semantically rich preconditions.
