Deep Just-In-Time Inconsistency Detection Between Comments and Source Code
نویسندگان
چکیده
Natural language comments convey key aspects of source code such as implementation, usage, and pre- post-conditions. Failure to update accordingly when the corresponding is modified introduces inconsistencies, which known lead confusion software bugs. In this paper, we aim detect whether a comment becomes inconsistent result changes body code, in order catch potential inconsistencies just-in-time, i.e., before they are committed base. To achieve this, develop deep-learning approach that learns correlate with changes. By evaluating on large corpus comment/code pairs spanning various types, show our model outperforms multiple baselines by significant margins. For extrinsic evaluation, usefulness combining it build more comprehensive automatic maintenance system can both resolve based
منابع مشابه
Comments on Inconstancy and Inconsistency
Overview In this very interesting paper, David Ripley brings together formal work on the semantics of vague predicates with experimental work on speakers’ reactions to statements involving such predicates. Working within a broadly Kaplanian framework—in which predicate expressions have characters, which together with contexts determine contents, which together with circumstances determine exten...
متن کاملFilms as Graphical Comments in the Source Code of Programs
We suggest to use animated pictures (films) as graphical comments in the source code of programs. Such pictures can be played forwards and backwards in steps under the control of the user. They can have multiple branches, which lead to different pictures. Animation effects can be applied to show how a picture changes over time. This can be useful for visualizing the dynamic behavior of programs...
متن کاملData Integration: Inconsistency Detection and Resolution Based on Source Properties
This paper addresses the problem of integration of multiple heterogeneous information sources. The sources may conflict with each other on the following three levels: their schema, data representation, or data themselves. Most of the approaches in this area of research resolve inconsistencies among different schemas and data representations, and ignore the possibility of data-level conflict alt...
متن کاملJust-In-Time Compilers and Dynamic Code Generation
• Partial Evaluation (Run-Time Specialization) Proposed as a way to speed up selected pieces of code. Some function arguments may be known to be (almost) constant at run-time. A specialized (partially evaluated) version of those functions can be generated dynamically, and will hopefully execute faster than the general version. Some of the work done in this area includes: Auslander’s extensions ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Proceedings of the ... AAAI Conference on Artificial Intelligence
سال: 2021
ISSN: ['2159-5399', '2374-3468']
DOI: https://doi.org/10.1609/aaai.v35i1.16119