Make trac use database namespaces or available equivalents

[anonymous]

trac cannot be installed into a database that is already used by other software when it has the same names. Using schemas (postgresql) or namespaces (or whatever other dbms call it) would prevent this, as it sets some kind of prefix and search path for the data. That would ease the integration of trac into already existing databases, as it will allow to keep trac within the project database

[Jonas Borgström]

Just testing, please ignore — Jonas

[anonymous]

Isn't there support for schemas in PostgreSQLConnector?

from postgres_backend.py, PostgreSQLConnection constructor:

  try:
    self.schema = None
    if 'schema' in params:
       self.schema = params['schema']
       cnx.cursor().execute('SET search_path TO %s', (self.schema,))
  except PGSchemaError:
    cnx.rollback()

and from postgres_backend.py, PostgreSQLConnector.init_db:

       if cnx.schema:
            cursor.execute('CREATE SCHEMA "%s"' % cnx.schema)
            cursor.execute('SET search_path TO %s', (cnx.schema,))

[anonymous]

Documentation is missing regarding how you specify the schema in the connection string.

[anonymous]

This needs to be in the PostgreSQL documentation:

database = postgres://user:pass@server/database?schema=yourschemaname

[Noah Kantrowitz (coderanger) <coderanger@…>]

Replying to anonymous:

This needs to be in the PostgreSQL documentation:

database = postgres://user:pass@server/database?schema=yourschemaname

This site is a wiki, feel free to add things.

[Christian Boos]

Documented since TracEnvironment@17.