CS 481 - Team 49 Meeting Minutes Weekly Team Meeting Friday, October 8th, 2007 10:30 - 11:30 AM JEB 324 Recorder: Shruti Upadhyaya Summary: Travis did the presentation for Release 1. As Travis went through each slide of PowerPoint presentation, different aspects of the project were discussed. Attendees Brandon Riggers, Shruti Upadhyaya, Peter Brown-Hayes, Travis Thoroman, Bruce Mayes (over the phone), Bill Junk Topics discussed: Bill said that we will probably need a different link in our webpage for releases. We need an architectural diagram in our presentation, where the total solutions of CTMT (Code Turmoil Measurement Tool) are laid out. The operational overview diagram needs to be changed slightly. Command line interface is supposed to face with subversion. In the current diagram of operational overview, we can put a few blocks inside CTMT to explain how it works. We didn’t have good understanding of how we were going to wor with Cruise Control before. It was not quite clear how Cruise Control fits in the environment. In the Operational Context diagram, Cruise Control needs to have two ways relationship with CTMT. Cruise Control will run both turmoil_between_revision and run_turmoil. We got the opportunity to ask Bruce various questions about Cruise Control. Cruise Control commits Subversion directly. It runs every 15 minutes and if there are any changes in different files, it provides us the changes. It only kicks off changes once the commit is complete. There are two kinds of commits in Subversion – pre-commit and post-commit. Cruise Control will most likely give us changes after post-commit. We will be getting more information on status of Cruise Control from Bruce. After talking about Cruise Control, we moved to our next slide – Testing. Bruce said that he will provide us different input files that he had problems with before for our testing purposes. We need to test identical files as well. The largest file that we have tested so far has 10,000 lines. Bruce said that he will provide us with large files. He also said that those files that he will send us need to be destroyed by the end of the semester. Bill suggested sending us two versions of same file. The test files can be as large as 190,000 lines of codes. Bruce mentioned that we need to look at improving the flexibility of the code so we won’t be depending on specific Operating System. We need to compare the result of current tool with the old tool to see if we are getting same results. We need a summary of testing, where we have comparisons of different diff utilities like Cygwin diff, Linux diff, etc. In our testing output, we need two more tags for new and deleted files. Then we moved on to next slide – defects. Defects"“weirdnes" ” needs to be captured in Test Specification. We discussed what needs to be in our 2nd release. We will have our 2nd release 3 weeks from now, which will give us about 2 weeks of development time and about a week of testing. Adding on to more comments, Bruce again mentioned that we need to have a Quantitative Analysis about differences of different diff utilities such as Linux diff and Cygwin diff. We need to include the sum of amount of changes in Linux and Cygwin as well as percentage of different between different diff utilities. We have asked Bruce to run test harness in his environment and see how his end compares to ours. We have found that we have identical results in Linux and Unix distribution. We have tested some files with Cygwin, but not the whole test suite. Bruce will be taking a look at our tarball file this afternoon. We also need our CVS Stat. Our next meeting is this Friday, October 12th. After Travis completed the presentation, Bruce handed out peer evaluation form and concluded the meeting.