Re: pending-pullups

[David Holland <dholland%netbsd.org@localhost>]

On Mon, Nov 30, 2015 at 01:39:51PM +0100, Martin Husemann wrote:
 > > >      In FreeBSD, commits are often marked:
 > > > 
 > > > MFC: <some time period>
 > > 
 > > In FreeBSD, the individual committer is responsible for the merge.
 > > In our world, IMO it just adds noise. You can just as well add it to the
 > > pullup queue on commit with a notet to stale the ticket for now.
 > 
 > IMHO this just leads to "fire and forget" culture and broken pullup
 > requests, i.e. requests that do not have the full set of fixes that
 > were necessary to cleanup after the original commit.

concur

FreeBSD's pullup/merge system definitely doesn't work as well as ours.

In any case, pullups should be associated with issues, not commits, so
the state belongs in gnats and not in the cvs logs, and stuffing
commits directly into the pullup queue isn't the right method.

What we *should* do is standardize a set of tags to put in commits to
indicate bug database actions; we don't have to have gnats understand
or act on these directly for them to be useful.

-- 
David A. Holland
dholland%netbsd.org@localhost