Changes between Initial Version and Version 1 of ReleaseSchedule


Ignore:
Timestamp:
May 23, 2008, 2:41:48 PM (14 years ago)
Author:
Beman Dawes
Comment:

Initial commit

Legend:

Unmodified
Added
Removed
Modified
  • ReleaseSchedule

    v1 v1  
     1= Release Schedule =
     2
     3== Next Release ==
     4
     5== Future Releases ==
     6
     7Releases are scheduled for the last day of the first month of each quarter:
     8
     9 * January 31
     10 * April 30
     11 * July 31
     12 * October 31
     13
     14Release milestones
     15
     16 * One week after prior release ships: branches/release opens for all stable changes, including bug fixes, major library upgrades, and, with permission, new libraries. Breaking changes should be coordinated with libraries affected.
     17 * Six weeks before release: branches/release closes for new libraries and major upgrades or breaking changes to existing libraries. Still open for bug fixes and minor changes to all libraries.
     18 * Four weeks before release: branches/release closes for routine minor changes and fixes. Still open for serious problem fixes.
     19 * Three weeks before release: branches/release closes for all library changes except when specific permission given by release manager(s).
     20 * Two weeks before release (sooner if possible): Beta 1 target date. Further betas and/or release candidates as feedback dictates.
     21
     22== Schedule Rationale ==
     23
     24The quarterly schedule is based on mailing list discussions and the results of
     25a straw poll at the BoostCon '08 "Future of Boost" session:
     26
     27How often do we want to have Boost releases?
     28
     29 4 weeks: 0 votes, 6 weeks: 0 votes, 8 weeks: 10 votes, 12 weeks: 25 votes, 16 weeks or longer: 6 votes
     30
     31The last day of the month is chosen rather than the first day, because January 1st
     32comes after a holiday period in which it is hard to get anything done.
     33
     34The schedule for release milestones is based on practical issues like the length of time it takes regression tests to cycle, and is adjusted as experience dictates.