On 26/06/2012 16:20, RW wrote: > The staging area is appealing in its own right. I'm less keen > on sub-packages which are going to break update tools. I think it's > very likely that only portmaster would survive.
Yes -- omlettes, eggs etc. There will be some inevitable disruption in
the ports tree to get all these changes in place. However, I hope
people will agree that the end result is worth it.
Staging will be (at least at first) a soft introduction. That is,
patches should go into bsd.ports.mk et al to add staging capability to
the ports tree generally, but nothing will change for an individual port
until the maintainer adds 'USE_STAGE=yes'. In this way, we can test out
some specimen ports thoroughly in live usage, debug the inevitable
problems and work out the best practices for a maintainer to do the
switchover on their ports. Then we have a campaign to switch individual
ports over to staging, much like the ongoing OPTIONSng conversion.
This, we believe, should cause minimal impact to end-users. Eventually
'USE_STAGE=yes' will be made the default and required for all new ports.
Any remaining unconverted ports will have to have 'USE_STAGE=no.'
Introducing sub-ports would be done in a similar gradual manner, but
plans for that are still quite vapourous. I don't share your pessimism
over the fate of all the port updating tools -- if we do this right,
there should be sufficient time and volunteers making patches to get
everything fixed up.
Cheers,
Matthew
--
Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard
Flat 3
PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate
JID: matt...@infracaninophile.co.uk Kent, CT11 9PW
signature.asc
Description: OpenPGP digital signature
