arXivDaily arXiv每日学术速递 周一至周五更新
arXiv cs/0012007cs.SEcs.PL
收藏

Kima --并发逻辑程序的自动纠错系统

Kima - an Automated Error Correction System for Concurrent Logic Programs

Yasuhiro Ajiro, Kazunori Ueda

摘要:

我们已经实现了Kima,一个并发逻辑程序的自动纠错系统。Kima纠正了诸如在没有显式声明程序属性的情况下出现错误的变量等近似错误。强建模/类型和基于约束的分析在调试并发逻辑程序以及建立通信协议和数据类型的一致性方面起着重要作用。模态平板GHC的模态/型分析是一个具有许多简单模态/型约束的约束满足问题,可以有效求解。我们提出了一个简单而有效的技术,给出了一个非良好的模式/类型的程序,诊断的“原因”的不一致,通过找到最小的不一致的子集的模式/类型的约束。由于每个约束都跟踪程序中出现的符号,因此最小子集也可以告诉程序错误的可能来源。Kima通过系统地替换可能源周围的符号出现并重新计算重写程序的模式和类型来实现自动校正。只要bug是接近失败的,Kima就提出了相当少的替代方案,包括一个预期的程序。

英文摘要:

We have implemented Kima, an automated error correction system for concurrent logic programs. Kima corrects near-misses such as wrong variable occurrences in the absence of explicit declarations of program properties. Strong moding/typing and constraint-based analysis are turning to play fundamental roles in debugging concurrent logic programs as well as in establishing the consistency of communication protocols and data types. Mode/type analysis of Moded Flat GHC is a constraint satisfaction problem with many simple mode/type constraints, and can be solved efficiently. We proposed a simple and efficient technique which, given a non-well-moded/typed program, diagnoses the ``reasons'' of inconsistency by finding minimal inconsistent subsets of mode/type constraints. Since each constraint keeps track of the symbol occurrence in the program, a minimal subset also tells possible sources of program errors. Kima realizes automated correction by replacing symbol occurrences around the possible sources and recalculating modes and types of the rewritten programs systematically. As long as bugs are near-misses, Kima proposes a rather small number of alternatives that include an intended program.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑