Scholarlink

This paper explores optimizing static code defect detection across multiple standards (e.g., GJB8114, CWE, MISRA2012) and AST-based template repair strategies to improve software quality and efficiency. It introduces a universal mechanism for detecting code defects and an AST-based manual template method for precise automated repair. Through experiments on real-world code and benchmarks, the proposed methods show superior accuracy and repair precision over popular neural network-based approaches, highlighting their practical applicability.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cross-Standard Code Defect Identification and AST-Based Template-Driven Automated Repair Methods

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Tianyi Guan,
  • Yongfeng Yin,
  • Qingran Su,
  • Ruinan Qiu

摘要

This paper explores optimizing static code defect detection across multiple standards (e.g., GJB8114, CWE, MISRA2012) and AST-based template repair strategies to improve software quality and efficiency. It introduces a universal mechanism for detecting code defects and an AST-based manual template method for precise automated repair. Through experiments on real-world code and benchmarks, the proposed methods show superior accuracy and repair precision over popular neural network-based approaches, highlighting their practical applicability.