Re: Attempted "testing propagation status" report
Neil Schemenauer wrote:
> Say 'A' depends on 'B' with a matching version number. In testing
> we have:
> A_1 B_1
> and in unstable:
> A_1 B_2
> We can't move B into testing since it's waiting for A. Now someone
> uploads a new version of B,
> A_1 B_3
> and shortly after someone finally uploads A,
> A_2 B_3
> it would be nice if the testing scripts could use B_2 to update the
> testing distribution.
One problem with this approach (which I once suggested myself ) is that the
combination A_2 and B_2 will never have been tested by anyone before it enters
testing, since the combination never existed in unstable.
This means that while getting the packages into testing quicker, it also risks
making testing more buggy.