Mark Mitchell wrote: > >1. Add a field to bugzilla for the SVN revision at which a particular > >regression was introduced. Display that in bugzilla as a link to the > >online SVN history browser so that clicking on a link takes us from the > >PR straight to the checkin. This field value ought to be the most > >recent revision to the GCC trunk such that the bug did not occur in the > >previous revision, but does occur in all subsequent revisions.
On Wed, May 23, 2007 at 12:12:30AM -0700, Brooks Moses wrote: > In practice, it seems that in conversations here this is often reported > as a range -- it is known to work at SVN revision N, but fails at > revision N+5 or somesuch. Might it be useful to have some way of > recording this, rather than expecting a single number? The Bugzilla field is just a string, so it's possible to put a range there as well as a single number. The mathematical form for an open/closed range could be used: (working_rev,failing_rev] to indicate that the "bad" revision is > working_rev and <= failing_rev. A volunteer who builds the compiler once a day can easily fill in this kind of range when a failure shows up on a given day that wasn't present the previous day.