Need help from Thinking Minds
I am doing some trend analysis on the Firefox bugs. I am able to generate Regression bugs trends like - How many bugs are fixed in a given month, what is the average time for the fixed bugs between confirmed and fixed, How many of them had more than one re-fix [ rework metric for the metrics geeks ] etc.,
Results are available here
http://people.mozilla.com/~mnandigama/timeplot2.html
But I am trying to find the bug 'soak time' : meaning, when was the code that was changed to fix the bug was introduced / last modified that could have resulted in the said bug report.
I have the following inputs to attack the problem.
I don't want to use hg bisect to build and run the tests to validate the bug presence. I just want to figure out heuristically when this bug could have been introduced.
Can any of you have done some thing like this already. If so, can you please share your expertise !!
Results are available here
http://people.mozilla.com/~mnandigama/timeplot2.html
But I am trying to find the bug 'soak time' : meaning, when was the code that was changed to fix the bug was introduced / last modified that could have resulted in the said bug report.
I have the following inputs to attack the problem.
- The code base for Mozilla Firefox is in HG.
- I have the actual change set used to fix the bug.
- The entire code change history of the source file(s) that are modified to fix the bug.
I don't want to use hg bisect to build and run the tests to validate the bug presence. I just want to figure out heuristically when this bug could have been introduced.
Can any of you have done some thing like this already. If so, can you please share your expertise !!
Comments
hg annotate -c -r $previous-changeset $file
will tell you the changesets each line _last_ changed in. From log you can get the time of that change.
Perhaps this isn't very practical, it certainly goes the long away round!
In the long run it would provide more accurate information.