Day 58 - Thursday August 15th
A new listener made up much of today's work, and the idea is observation only. The listener is not meant to interfere with choice generation at all, but rather to run the SUT as if the JPF VM was any old Java VM, and have it run it's course. Then it looks to see at given times (or at the end of the run) is a given property is true or false (e.g. through a new method similar to others added this week: EvaluateHere(String propertyName), and similarly for many properties). Then it records this information, and then runs it all again, and again, up to a given bound, to see how many times each property was true or false, and reports this. Today was the start of that listener (which for the most part is a re-purpose of VarRecorder, with some changes), but unfortunately not the end of it.
Day 59 - Friday August 16th
Friday was a continuation and completion of Thursday's work. It looks like I've got the listener running the way it should, and I spent a bit of the day testing it, and comparing results with that of results from earlier runs. I also toyed around with the notion of letting it run the SUT for as many runs as it can before X minutes pass, but I wasn't really successful there.
I was also updating data for all the examples and I still had some larger examples to try that I'll just leave running at various times over the weekend so that I can get some really interesting (read: larger size) scalability results, and possible track memory usage in addition to time), so that I can pass off that report/spreadsheet I mentioned earlier. Expect that Monday (though probably not on this blog!).
No comments:
Post a Comment