That said, there is an alternative, which we can use if the situation
continues like that for a long duration, and that would be to ask for
the removal of Stefano's libpgsql (or whatever, i never remember the
name of this one) package from testing. If we were to do this, and the
package would be removed in timely fashion, then all ocaml stuff would
enter testing the next day, and the libpgsql library will enter testing
the day after postgresql becomes a valid candidate.
This course of action would cause problem if :
o Stefano thinks it is a bad idea :)))
o Someone following testing is really using libpgsql on an arch where
postgresql is not built (sparc i think). The case where he does on
i386 could simply be solved by installing the unstable version of
postgresql and libpgsql.
So, what do you think about this plans ? Do we wait (a few more days) or
do we ask for the removal of libpgsql from testing ? I think it is
stefano who has to take the decision about this, based on what he sees
best for the libpgsql users.