[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Attention, packagers! Invalid rebuilds



On Wed, 13 Sep 2006 10:31:28 -0500, Mike McGrath wrote:

> On 9/13/06, Toshio Kuratomi wrote:
> > On Wed, 2006-09-13 at 10:21 +0100, Karanbir Singh wrote:
> > > Michael Schwendt wrote:
> > > > Lots of warnings in today's push of packages in the needsign queue.
> > > >
> > > > There are still packagers who rebuild their packages _without_ increasing
> > > > the "Release" value.
> > >
> > > surely a VCS trigger to check for such ( and notify over email perhaps .
> > > ) is possible at the server end ?
> >
> > If you code this up, I think the check should go in the build code,
> > rather than the VCS.
> >
> > -Toshio
> >
> 
> I seem to remember some discussion about this a few months ago
> concerning the makefile checking during the "make build/plague", which
> isn't perfect but could at least inform the users that they're doing
> something they aren't supposed to.

Probably you confuse this issue with the "make tag" safety-check,
which would protect packagers from tagging uncommitted working-copies.

But a check whether somebody runs "make build" for an old tag would not be
easy to do without a package database and without a history of previously
built packages. Would you want the CVS Makefile to query the repository? ;)

More warnings today, also for the FC-5 branch. In devel:

  autogen
  fortune-firefly
  anjuta-gdl
  nethack-vultures
  kbibtex

Do we still have a mailing-list where every package maintainer must be
subscribed? (e.g. fedora-maintainers?)

Can the "nag-mails" be adjusted to explain what it means "to rebuild"
a package?


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]