Closed andymeneely closed 11 years ago
I was thinking that this should be collected for each vulnerability introduction commit. So there should be an analysis table at the end of RebuildHistory that collects "ProjectChurn" and "RecentFileChurn", each of them with a default parameter of 30 days.
Ok, next thing is to port this query over to a Java class and input the date as a timestamp. Also, modify the query to do project churn as well. You could do a separate analysis table, index the table, then do a view on both of those tables. Or something else.
Push what you've got, and then I'll optimize the speed.
Commit have been pushed.
On Thu, Jan 17, 2013 at 2:09 PM, Andy Meneely notifications@github.comwrote:
Push what you've got, and then I'll optimize the speed.
— Reply to this email directly or view it on GitHubhttps://github.com/apmeneel/httpd-history/issues/5#issuecomment-12383758.
Integrated into my branch as 08ea650b8312eb3c772b60ca9236fec9b9249550
The field should be RecentChurn in GitLogFiles. For each file on each commit, provide the sum of the TotalChurn (LinesInserted + LinesDeleted) for the last X days (e.g. 30 days). Include "this commit" in that count.
Parameters should be in the props file.
For example, suppose we have these changes:
January 1st: 5 churn --> Recent Churn = 5 January 2nd: 3 churn --> Recent Churn = 8 May 1st: 10 churn --> Recent Churn = 10 May 4th: 10 churn --> Recent Churn = 20