I believe Hoan gave the wrong URL, here is the correct one:
And just to build on what he provided, here is my version:
This job computes the top-10 committers. It is also coded to easily modify what file names you match against (the array 'searchNames' at the top).
The basic intuition is to iterate over each code repository (line 6), then over each revision (line 7), then over each file (line 8), and if a file matches (line 9) any of the provided regex's (line 4) then we increase the count for that particular user. This is done by emitting their name and a weight of '1' (line 10). The top aggregator (line 2) then groups by each user's name, sums all their weights, sorts by the total weight, and gives the top (in this case, 10) results.
- Robert