Revision as of 13:35, 9 September 2007

The worst thing in release proces is a thing called showstopper. It is any issue which causes fixing, building, packing and deployment of one more RC (release canditate) instead of releasing the version.

Often there raise discussions on different mailinglists if a found issue ist or is not a showstopper (writing this we have 2.3rc2 in releasetest and an issue with PDF export as well as a crash when pasting into database tables as Showstopper candidates under discussion).

Although there is a comprehensive guideline for priorizing issues, there is no such thing for deciding if an issue is an showstopper. Some even think that there never can be a rule for that. But even if there is no strict rule, one can have some criteria made explicit. This page is about to collext such criteria to decide whether a found issue is a showstopper or not.

List of criteria

Reasons why an issue can possibly be regarded as a showstopper

• Installation fails

Reasons why an issue should not be regarded as a showstopper

• Although issue crashes it occurs only when using an "unusual way" to get something done.

Other reasons that may influence the decision

• There may be a known security isssue which needs releasing a new version before it's too late. So a potential showstopper may be not regarded as one because of this external need to have a new release as soon as possible.

• Marketing may wish to have a release in time for OOoCon

• Sun release engineering has other plans ;-)

This page is just a start. Everyone is invited to post argumenst or decision rules important for the "Showstopper-decision". Maybe it is useful you sign your statement so a discussion can be held in IRC or on an appropriate mailing list.