Moving to mobile-l.
On Friday, June 12, 2015, Sam Smith samsmith@wikimedia.org wrote:
Hey web slingers,
If there is a regression introduced by a patch, then please revert that patch as soon as you've identified it and let the team know via Phabricator, email, or both. Reverting the commit will often be cheaper to do than fixing the regression in a follow-on patch, but there'll undoubtedly be exceptions, which we'll deal with (and learn from) as a team.
Fixing the regression in a follow-on patch means that:
- *master won't be deployable* until the patch has been reviewed,
tested, and merged, which should be communicated to the Release Engineering team
- reviewers might have to drop what they're working on in order to get
it reviewed - what if the original patch was lower priority? - we should be cognisant of the cost of context switching
- the commit history will be dirty
*Master should always be depoyable.*
–Sam