Compile-time Variability Research Articles

Effective Analysis of C Programs by Rewriting Variability

Context. Variability-intensive programs (program families) appear in many application areas and for many reasons today. Different family members, called variants, are derived by switching statically configurable options (features) on and off, while reuse of the common code is maximized. Inquiry. Verification of program families is challenging since the number of variants is exponential in the number of features. Existing single-program analysis and verification tools cannot be applied directly to program families, and designing and implementing the corresponding variability-aware versions is tedious and laborious. Approach. In this work, we propose a range of variability-related transformations for translating program families into single programs by replacing compile-time variability with run-time variability (non-determinism). The obtained transformed programs can be subsequently analyzed using the conventional off- the-shelf single-program analysis tools such as type checkers, symbolic executors, model checkers, and static analyzers. Knowledge. Our variability-related transformations are outcome-preserving, which means that the relation between the outcomes in the transformed single program and the union of outcomes of all variants derived from the original program family is equality. Grounding. We show our transformation rules and their correctness with respect to a minimal core imperative language IMP. Then, we discuss our experience of implementing and using the transformations for efficient and effective analysis and verification of real-world C program families. Importance. We report some interesting variability-related bugs that we discovered using various state-of-the-art single-program C verification tools, such as Frama-C, Clang, LLBMC.

Variability encoding: From compile-time to load-time variability

Many software systems today are configurable. Analyzing configurable systems is challenging, especially as (1) the number of system variants may grow exponentially with the number of configuration options, and (2) often existing analysis tools cannot be used for configurable systems. Recent work proposes to automatically transform compile-time variability into load-time variability—called variability encoding—with the goal of reusing existing analysis tools for analyzing configurable systems and improving analysis performance compared to analyzing all system variants in a brute-force manner. However, it is not clear whether one can automatically find an efficiently analyzable load-time configurable system for any given compile-time configurable system. Also, for many analyses, we need guarantees that the load-time configurable system precisely encodes the behavior of all system variants that can be statically derived. We address both issues (1) by developing a formal model of variability encoding based on Featherweight Java, (2) by proving that variability encoding preserves variant behavior with respect to a core set of language mechanisms, (3) by discussing how our work extends to more complex language mechanisms that elude our formal model, and (4) by sharing our experience with implementing and using variability encoding in real-world applications.

Family-based performance measurement

Most contemporary programs are customizable. They provide many features that give rise to millions of program variants. Determining which feature selection yields an optimal performance is challenging, because of the exponential number of variants. Predicting the performance of a variant based on previous measurements proved successful, but induces a trade-off between the measurement effort and prediction accuracy. We propose the alternative approach of family-based performance measurement , to reduce the number of measurements required for identifying feature interactions and for obtaining accurate predictions. The key idea is to create a variant simulator (by translating compile-time variability to run-time variability) that can simulate the behavior of all program variants. We use it to measure performance of individual methods, trace methods to features, and infer feature interactions based on the call graph. We evaluate our approach by means of five feature-oriented programs. On average, we achieve accuracy of 98%, with only a single measurement per customizable program. Observations show that our approach opens avenues of future research in different domains, such an feature-interaction detection and testing.

A variability-aware module system

Module systems enable a divide and conquer strategy to software development. To implement compile-time variability in software product lines, modules can be composed in different combinations. However, this way, variability dictates a dominant decomposition. As an alternative, we introduce a variability-aware module system that supports compile-time variability inside a module and its interface. So, each module can be considered a product line that can be type checked in isolation. Variability can crosscut multiple modules. The module system breaks with the antimodular tradition of a global variability model in product-line development and provides a path toward software ecosystems and product lines of product lines developed in an open fashion. We discuss the design and implementation of such a module system on a core calculus and provide an implementation for C as part of the TypeChef project. Our implementation supports variability inside modules from #ifdef preprocessor directives and variable linking at the composition level. With our implementation, we type check all configurations of all modules of the open source product line Busybox with 811~compile-time options, perform linker check of all configurations, and report found type and linker errors -- without resorting to a brute-force strategy.