The actual script is done by the commiters. You can read about the process in the spark commiter guidelines:
https://cwiki.apache.org/confluence/display/SPARK/Wiki+Homepage So the steps would be 1) Adopt the script for processing pull requests 2) Add it to our bigtop repo 3) Create a pull request and test the script for comitting it. Some might dispute one issue here : The patch system record of is a little diluted. Is that a big deal? Im not sure . I think its a tough decision - its no simple answer. If losing track of patches builds a significantly more vibrant ecosystem, i think its an okay tradeoff. Any other opinions on this? On Jun 30, 2014, at 6:38 PM, Andrew Bayer <andrew.ba...@gmail.com> wrote: > builds