The catalog pg_pltemplate stores "template" information for procedural languages. A template for a language allows the language to be created in a particular database by a simple CREATE LANGUAGE command, with no need to specify implementation details.
Unlike most system catalogs, pg_pltemplate is shared across all databases of a cluster: there is only one copy of pg_pltemplate per cluster, not one per database. This allows the information to be accessible in each database as it is needed.
Table 43-26. pg_pltemplate Columns
|tmplname||name||Name of the language this template is for|
|tmpltrusted||boolean||True if language is considered trusted|
|tmplhandler||text||Name of call handler function|
|tmplvalidator||text||Name of validator function, or NULL if none|
|tmpllibrary||text||Path of shared library that implements language|
|tmplacl||aclitem||Access privileges for template (not yet used)|
There are not currently any commands that manipulate procedural language templates; to change the built-in information, a superuser must modify the table using ordinary INSERT, DELETE, or UPDATE commands. It is likely that a future release of PostgreSQL will offer commands to change the entries in a cleaner fashion.
When implemented, the tmplacl field will provide access control for the template itself (i.e., the right to create a language using it), not for the languages created from the template.