On 07/01/2004, at 8:53 AM, Thomas Klausner wrote:

> It's not only for developers uploading binary packages,
> it's also for everyone else who's using binary packages.

   So what about the solution I proposed?

   It should meet the needs of people distributing
binary packages without hampering people using packages
compiled from source.

>> pkgsrc tree to get some work done.  IMHO the danger then
>> up upload packages that were created from an inconsistent
>> pkgsrc tree is much greater this way than without the
>> forced updates.
>
> I'm sorry that this is making you angry, but I don't see
> that as an argument.

   It was not making me angry, but maybe I should write my
first emails after breakfast rather than before ;-)

   Nevertheless, the danger of inconsistent packages is real
and has bitten us before.  And causing people to resort to
the use of inconsistent pkgsrc trees and/or dangerous
package installation practices is not going to help matters.

>> this is not the case.  Pkgsrc is hardly ever in a state
>> where everything compiles.
>
> This is a different argument.

   Yes, and here are probably several more that I cannot
think of at the moment.

   To give you a historic reference: we now have
NO_BIN_ON_FTP and the like for purposes that we used to
(ab)use NO_PACKAGE for.  While the reasons are slightly
different, the principle of separation of technical
and policy requirements remains the same.

   Cheers
       ,
    Rene