I have multiple installations of PostgreSQL running on the same box (7.1.x,
7.2.1, etc.) and different code bases running against each installation,
e.g., a 7.1.x cluster, a 7.2.1 cluster, etc. When building DBD::Pg, it is
linked against one of these versioned libraries. In this case, however, I
need the 7.1.x code base to use a DBD::Pg linked against the 7.1.x libraries
while the 7.2.1 code base uses a DBD::Pg linked against the 7.2.1 libraries.
We are kicking around the idea of multiple perl installations, but that
feels like overkill and moves me into configuring a number of details for
perl that I've not had to fuss with before.
Question: Does anyone have a suggestion or experience as to a minimalist and
clean way to setup/build DBD::Pg modules against multiple PG installations?