Fetching latest commit…
Cannot retrieve the latest commit at this time.
|Failed to load latest commit information.|
This contains ALL of the illumos repositories and related projects. It is essentially read-only and not intended for building and/or compilation. It is to be used to analysis of commit activity and so forth. This is a framework for downloading illumos-related repositories from git-hub, and doing data analysis on the commit-logs of those repositories. There is a surprising amount of information in those logs. To use for first time: cd tools ./get_repos.ksh # This generates elaborate git-logs in JSON format in the ../data directory # You'll figure it out. ./commit_jlogs.ksh To update the repos: cd tools ./pull_repos.ksh ./commit_jlogs.ksh To get some stats: cd tools ./repo_stats.ksh You'll then want to go get a cup of coffee, and maybe play some video-games -- it will take 15 minutes or so get the stats over all of the repositories. If you think that's too long, keep in mind that it used to take 51 minutes -- for _one_ repository. This will distill the JSON commit logs into independent commit logs for each author, to facilitate easier (and more parallelizable) per-author analysis. Also creates an author summary. All output is in JSON, so it can be consumed by additional custom scripts. The output is stored in: data/commit_logs/<project type>/<repo name>_authors/ The repo_stats.ksh script and supporting code are still in the works, and will be getting support for more advanced statistics. Also, ideas welcome! If something doesn't work, please open an issue on github, or email me at email@example.com If you are the owner of an Illumos-related repo, please send me an email or open an issue to have your project's repo included in the data. Illumos-related repos are anything. Even userland code that depends on or leverages Illumos-specific facilities (for example the DTraceToolkit or even Joyent's Manta). The more data the better. Future work may include fetching the mailing list data for every illumos-related mailing-list and doing some kind of (as of yet unknown) analysis on this data. For example, I want to see if there is a correlation between commit-activity and email-activity. Also, I suspect, but would like to verify/falsify that the most political and opinionated mailing-list posts are more likely to be written by people that have zero or close to zero commits in the kernel repos (and other related repos perhaps). Special thanks to Mike Dimitropoulos (sdimitro on github) for suggesting this side-project which was a breath of fresh air, from the more sophisticated stuff I work on.