This website aims to show my work, so I will just leave it plain-text. To see the real me, visit here.
The built system could be regarded as a customized version of Mozilla rr.
In my point of view, the goal of this research is to build a patch-validation system which could help user see if their patch could fix the user reported severe bug (memory leak, null pointer and so on). The assumption is that the patch only contains limited lines of update rather than a total renew, which is necessary for recording.
The whole process for system's processing is divided into two phases: record and muplay.
For recording, rr tech report has given its definition and usage. This usually happens in user's environment and records information.
For muplay, details are unavailable now, and I will update this when the paper gets published.
This research is based on Deepbase published in SIGMOD 19. The goal is to distribute the Deepbase system.
Think of use cases and translate them into SQL-like query.
Created customized parsers based on moz-sql-paser and generate logical plan.
Created Sqlite3 database to collect metadata for uploading data.
Created Google Storage for model, dataset and hypothesis storage.
This is an open-source project and the code is available at here.
Created multiple test cases.
Enhanced NVD-database by pulling data from offical cve website.
Refactored code for parsing json file.
Implemented bash command like file and string by Python.