Closed StudentJamesChen closed 4 weeks ago
GitHub Bot Integration: The first basic requirement of the project will be to create the infrastructure needed to support a GitHub bot. Developer(s) will need to handle the technical implementation of the bot itself. Maybe not so hard, but it's essential.
Adapt Existing Fault Localization Code: The next basic requirement will be to adapt the fault localization code to work on the server. Developer(s) will need to ensure that the code is function on a remote environment to be leveraged by the GitHub bot. Could be pretty hard, who knows lmao.
Processing GUI Information: Lastly, we will need to process inputted getevent traces into GUI information to be used by the existing fault localizaion code. Developer(s) may need to leverage tools such as TraceReplayer to retrieve XML data. XML data may need to get parsed and passed over to existing fault localization code.
Brainstorming and Strech:
How would we match screenshots to XML data? If going with this method of user attaching screenshot to issue and querying DB based on this?
Possibly random crawling to get screenshots of app and GUI data
Not sure if this would be suitable for MVP
associated with monkey? Robo Test?
Even further... backend app crawling.
Guidance from Dr. Moran: MVP components could be reduced further by having GUI data processing as an optional step. In that case, the user would input only a bug report and get back a baseline ranking of potentially buggy files. Here, the backend would only be responsible for the code developed in the research paper.
There are also scopes beyond the MVP that could get discussed/documented for review.