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

Re: NEW processing

> It surprises me that the only solution to that problem seems to be to
> add more people to the FTP team, so that the processing bandwidth will
> improve.

That is the only solution which will help.

> That's not true.

Suuuure. And the world dies tomorrow.

> We imposed that reviewing step to ourselves, and, if it's doing more
> harm (by slowing down development and annoying contributors) than good
> (by detecting mistakes and improving Debian's overall quality), we
> could simply decide to drop it. (or to drop it partially, for some
> categories of uploads).

Yeah, sure. And open the archive to more legal problems than people can
If we wouldnt have the current checks for licenses stuff, the archive
would be a mess of undistributable and illegal stuff. We get more "All
rights reserved", "CONFIDENTIAL, FOR INTERNAL USE ONLY" and "Provided
under NDA, if you distribute this we will sue you" (in much more words)
than is ever good.
And there is the usual forgotten license/copyright holder too.

And the checks for that take about 98% of the time in NEW.

Packages that only add new binary components are already sorted above
packages that have completly new source, to decrease their time in the
queue, as their checks are much faster done than a complete source
review. But even those have a little review from us, enough people
manage to even get those done wrong. Love empty packages? soname changes
like to do that to people, for example.

Yes, the only good solution is to add more people. Luckily we got
another volunteer during the ftpmaster meeting in Extremadura, someone
that not only complains on lists but actually wants to do work.

Note that we currently are working on integrating lintian into dak in a
way that lets us autoreject on selected lintian tags. That will help NEW
a little too, even if NEW is the smallest driving force for this change.
But in the same change we *can* go and reject all NEW packages that show
*any* (unoverridden) lintian E:, that would be a large batch already.

bye, Joerg
* wiggy just looking at gforge-inject output
<wiggy> last year I could not run it for months and still not see any new users
<wiggy> it just added 19 new users
<Mithrandir> it's this bloody active new DAM we've managed to get.

Reply to: