Week2ProgressReport.md

This week I had the chance to look at: mv, nm and strip

• Testing mv: The command that was being replayed was in the form of: mv file1 dir. This command was tested for filetype anomalies for both file1 and dir. Changing the type for dir causes rr to diverge, which is the expected behavior. This signifies that the application is doing something differently to handle the anomaly as opposed to running with it. However, changing the type for file1 doesn't cause rr to diverge. One possible cause could be that the lstat system call is being used obtain information about the file1. Maybe since it's being treated as a symbolic link, changing the filetype doesn't quite cause the test to diverge, since the actual type of the file doesn't change.

• Testing nm: nm is a built-in Linux command to list the symbolic links in an object file. Ran into a major problem with rr on this one. For some reason replaying the test even without any change to trace_snip causes rr to hang. Still investigating this problem.

• Opened Issue #60: After testing nm, I tried to test the command strip. However, I discovered that even though I'm creating new tests, the strace in the new test directory was actually the strace from the last successful rrtest for a different command. Turned out it was an issue with error handling in rr. I was missing a pre-configuration for rr that I needed to configure everytime the system boots. Since rr wasn't throwing an exception and quitting out of the test creation when that configuration is missing, I wasn't able to test anything new. This one took us until Tuesday to figure out. Thus ends the week 2.