Tom Lane <[EMAIL PROTECTED]> writes:
> Returning to the original problem, it seems to me that comparing "pg_dump
> -s" output is a reasonable way to proceed.
I've actually started checking in a pg_dump -s output file into my CVS tree.
However I prune a few key lines from it. I prune the TOC OID numbers from it,
and anything not owned by the user I'm interested in.
The makefile rule I use looks like:
schema.sql:
pg_dump -U postgres -s user | sed '/^-- TOC entry/d;/^\\connect -
postgres/,/^\\connect - user/d;/^SET search_path/d;/^$$/d;/^--$$/d' > $@
This still suffers from one major deficiency. The order that objects are
outputed isn't necessarily consistent between databases. If I add tables to
the development server but then add them to the production server in a
different order the schema still shows differences even though the objects in
the two databases are identical.
--
greg
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
