Three more papers were accepted for presentation at CBSoft 2026, this time at SBCARS, SAST, and VEM!

How Rust Projects Use Macros: A Large-Scale Empirical Study, co-authored with Nilo Drumond (CIn-UFPE), was accepted at SBCARS. In this work, we investigate how macros are used in practice across 99 popular open-source Rust repositories, comprising more than 4,000 crates and 2.2 million macro invocations. Among our findings, we show that macro usage differs substantially across macro categories, that derive and attribute macros are strongly concentrated around a small number of ecosystem crates, and that macro density also varies considerably across application domains.

Evaluating Shaker for Flaky Test Detection in Python Projects, co-authored with Gabriela Leal (CIn-UFPE) and Denini Silva (UFRPE), was accepted at SAST. We present the first empirical evaluation of Shaker for Python projects, comparing its resource-contention strategy against plain test re-execution over 137 known flaky tests. Contrary to previous results for Java and Android, Shaker did not provide a significant detection advantage for Python. The study also highlights a broader methodological issue: flaky tests identified in one execution environment may no longer reproduce in another, potentially distorting evaluations based on previously established ground-truth datasets.

Finally, Django-Specific Code Smells: Catalog, Prevalence, and Automated Detection, co-authored with Eduardo Silva (CIn-UFPE), was accepted at VEM 2026. In this paper, we investigate code smells specific to the Django web framework, deriving a catalog from Django-focused grey literature and implementing a static analysis tool for detecting five smells across models, views, and templates. An analysis of 33 open-source Django projects found more than 4,000 occurrences, while manual validation indicates that the proposed detectors are reasonably accurate, although their reliability varies across smells.

See you at CBSoft 2026!