Re: Work done on the WNPP front
David Moreno Garza <email@example.com> writes:
> A fourth shoot was intended a couple of minutes ago on the ITA bugs. The
> ITAs were a little bit different. Since they represent existing
> packages, the mechanism used here was intended to close bugs, but
> retitling into O. ITAs are a very sensitive part of the WNPP chunk. An
> automatic retitling for ITAs with an inactivity greater than 250 days
> was launched, but it didn't return any bug retitled, which is great
> since at least this shows there is some activity on the BTS regarding
> these bugs.
You can eliminate most of the noise by taking into account only messages
from the prospective adopter (i.e. the owner of the WNPP bug). See
~vela/ita-age.pl on merkel for a sample implementation; as of today it
lists 20 ITAs older than 250 days.
However, there is an important gotcha with retitling ITAs: currently we
have no reliable way to determine whether the package was previously up
for adoption (RFA) or orphaned (O).
The natural solution is to migrate WNPP bugs to BTS tags. If I'm
remembering correctly, Anthony Towns talked on IRC about using
orphaned/up-for-adoption tags combined with pending to mark ITAs.
The expiry script would then simply remove the pending tag.
This would also prevent merged bugs from getting into conflicting states
since (unlike titles) they share a single set of tags (see #310181 for
an example of the problem).