Various columns in the Slony schema that store relation names use the "name" type. According to PostgreSQL docs: http://www.postgresql.org/docs/9.0/static/datatype-character.html "The name type exists only for the storage of identifiers in the internal system catalogs and is not intended for use by the general user. Its length is currently defined as 64 bytes (63 usable characters plus terminator) but should be referenced using the constant NAMEDATALEN in C source code." There is some room to quibble over whether Slony is an "internal system" thing or a "general user," but it has been observed that the use of the name type adversely affects attempts to run pg_upgrade against Slony nodes. http://lists.slony.info/pipermail/slony1-general/2011-April/011620.html It seems as though we should switch from "name" to "text".
When switching to "text", will we have to add explicit casting to all the queries that use those columns?
I just made the changes on sl_table/sl_sequence, and ran one of the old regression tests. No problems encountered. https://github.com/cbbrowne/slony1-engine/tree/bug203 I did not do any casting changes, and experienced no problems, so I don't expect any explicit casting should be necessary.