A method to aid in debugging and enhancing user understanding of the code.
Publication Date: 2010-Oct-29
The IP.com Prior Art Database
In event of absence of stack trace and other relevant information, it becomes difficult to locate a starting point for the code investigation and locate a defect. Proposed art suggests a method to identify starting point for the code investigation and locate probable source of the given bug in code.
Page 01 of 2
A method to aid in debugging and enhancing user understanding of the code .
When working on a bug, there are two versions of a server one on which there is no problem and one version on which the bug manifests. In case of absence of a stack trace (bugs with wrong results, etc) it becomes difficult to locate a starting point for the code investigation.
Problem: In event of absence of stack trace and other relevant information, give a quick handle to the developer to start the investigation.
Code Flow Graph: We define Code flow Graphs as the code path taken by the execution in a control flow graph. A code flow graph represents all the functions called during the execution, along with their caller-callee relationships, number of loop iterations, parameter values, etc.
(This page contains 00 pictures or other non-text object)
Fig 1: Code Flow Graph
Problem stated in section 1 would be addressed as follows: Let the user define two code points (starting point and ending point) in execution flow of the bug, these code points need not be exact code points where the bug lies, but these points should define a segment of the code which encompasses the cause of the bug. A code flow graph is generated for both the binaries (with and without the bug) of the server for the code segment defined by the user. Both the code flow graphs are now compared to each other to find out the points where they differ. These difference areas on the code flow graph give the user a good starting...