Re: Will the amd64 port be rejected because of the 98% rule?

[Goswin von Brederlow <brederlo@informatik.uni-tuebingen.de>]

Peter Samuelson <peter@p12n.org> writes:

> [John Hasler]
>> Make it 98% of the packages buildable on the accepted port with the
>> highest build percentage.
>
> That's not fair either, unless you require all packages to be autobuilt
> (which is a thread we don't need to duplicate now).  There have always
> been a significant number of packages which FTBFS except on the
> developer's own machine, usually i386.  This artificially inflates the
> percentage on the "port with the highest build percentage".

That would be a FTBFS bug and RC. That would automaticaly exclude that
package from testing, causing its removal if the problem can't be
solved.

I would hope the 98% cut will be calculated from testing and not
unstable in which case such uploads would not skew the result.

> In the end the decision to apply this benchmark will be manual and
> arbitrary; it's not like a precise way to measure it really matters.
> As stated elsewhere, at present, nobody except i386 comes even close to
> 98%.

I see an increase of amd64+source uploads once amd64 gets added to
debian and a respective decrease for i386. Maybe that evens it out a
bit more.

By the way, i386 does not make the cut according to the vancouver
prospect due to the number of buildds required. So are we left with 0
archs in etch? :) That will certainly speed up the release.

MfG
        Goswin