cycletimes.py is a tool for trying to understand chrome productivity. Uses commit and review information to produce statistics on how long it takes changes between upload, review, test, commit, and release.

See cycletimes.py help for help.

Precache all the review/commit information into

/chromepath/cycletimes_cache

cycletimes.py /chrome/path update

Note 'update' can take a very long time.

Use -v to see verbose output, including when

requests have to hit the network.

Reads cached files from /chromepath/cycletimes_cache

and dumps per-repository stats to stdout.

cycletimes.py /chrome/path stats

Reads cached files from /chromepath/cycletimes_cache

and dumps js to stdout for use with graph.html

cycletimes.py /chrome/path graph

TODO:

• Verify that timezones are correct for all timestamps! (Timezones can mean hours, which is a lot of time!)
• Blink Rolls - list DEPS, grab webkit_revs and commit_date.
• Could build a list of date -> rev and filter out any rollouts?
• Understand Reverts
• Could show what % of time was due to tree-closures by overlapping lgtm -> commit time with tree closure data?
• Could show what % of time was due to CQ by overlapping CQ data with lgtm -> commit time.
• Want to show what % of a typical patch time is spent waiting for X.

Should be possile to do skia by grabbing the DEPS from the branch and using that to get Skia revisions.