Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
156 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

DepRes: A Tool for Resolving Fully Qualified Names and Their Dependencies (2108.01165v1)

Published 2 Aug 2021 in cs.SE

Abstract: Reusing code snippets shared by other programmers on Q&A forums (e.g., StackOverflow) is a common practice followed by software developers. However, lack of sufficient information about the fully qualified name (FQN) of identifiers in borrowed code snippets, results in serious compile errors. Programmers either have to manually search for the correct FQN of identifiers which is a tedious and error-prone process, or use tools developed to automatically identify correct FQNs. Despite the efforts made by researchers to automatically identify FQNs in code snippets, the current approaches suffer from low accuracy when it comes to practice. Moreover, while these tools focus on resolving the FQN for an identifier in a code snippet, they leave the challenge of finding the correct third-party library (i.e., dependency) implementing that FQN unresolved. Using an incorrect dependency or incorrect version of a dependency might lead to a semantic error which is not detectable by compilers. Therefore, it can result in serious damages in the run-time. In this paper, we introduce DepRes, a tool that leverages a sketch-based approach to resolve FQNs in java-based code snippets and recommend the correct dependency for each FQN. The source code, documentation, as well as a demo video of DepRes tool is available from its code repository at https://github.com/SoftwareDesignLab/DepRes-Tool.

Citations (2)

Summary

We haven't generated a summary for this paper yet.