Re: NEW processing
On Wednesday 03 December 2008 20:35:11 Clint Adams wrote:
> On Wed, Dec 03, 2008 at 07:52:06PM +0200, George Danchev wrote:
> > I'm afraid that skipping the 3rd thing `trying to reduce the number of
> > bugs in Debian' [1] would lead to a massive waste of time for
> > autobuilders caused by these subsequent uploads meant to bring the
> > package(s) in a technically sane shape, that is what would have been
> > rejected by the ftpmasters in the first place. So, it is much better
> > these to be detected and probably rejected before doing any more harm on
> > their way. Low quality packages won't help users either, nor these users
> > get the finally fixed and brought into relatively sane shape package
> > faster.
>
> No, this is condescending and implies that the ftp-team knows more about
> package quality than the other developers, when that is not true and nor
> should it be true as a condition of holding that role.
Then, why should they check for package namespace being sane and archive being
legal ? Zero checks means NEW being fast as lightning, no ?. OTOH, ftpmasters
might not be familiar with the gory details of a particular piece of
software, but are able to catch the most blatant and redundant packaging
issues like FTBFSIASW, debian/control breakages, and severe policy violations
(mainly FHS) based on their experience. I believe they have seen a lot.
--
pub 4096R/0E4BD0AB 2003-03-18 <people.fccf.net/danchev/key pgp.mit.edu>
Reply to: