[gnue] r7668 - in trunk/gnue-common/src/datasources/drivers/sql: . postgresql

[reinhard]

Author: reinhard
Date: 2005-07-01 03:07:10 -0500 (Fri, 01 Jul 2005)
New Revision: 7668

Added:
   trunk/gnue-common/src/datasources/drivers/sql/postgresql/
   trunk/gnue-common/src/datasources/drivers/sql/postgresql/Behavior.py
Removed:
   trunk/gnue-common/src/datasources/drivers/sql/postgres/
   trunk/gnue-common/src/datasources/drivers/sql/postgresql/Behavior.py
Modified:
   trunk/gnue-common/src/datasources/drivers/sql/postgresql/__init__.py
Log:
Renamed postgres driver back to "official" name postgresql.


Copied: trunk/gnue-common/src/datasources/drivers/sql/postgresql (from rev 
7666, trunk/gnue-common/src/datasources/drivers/sql/postgres)

Deleted: trunk/gnue-common/src/datasources/drivers/sql/postgresql/Behavior.py
===================================================================
--- trunk/gnue-common/src/datasources/drivers/sql/postgres/Behavior.py  
2005-06-28 11:27:42 UTC (rev 7666)
+++ trunk/gnue-common/src/datasources/drivers/sql/postgresql/Behavior.py        
2005-07-01 08:07:10 UTC (rev 7668)
@@ -1,420 +0,0 @@
-# GNU Enterprise Common Library - Schema support for PostgreSQL
-#
-# Copyright 2000-2005 Free Software Foundation
-#
-# This file is part of GNU Enterprise.
-#
-# GNU Enterprise is free software; you can redistribute it
-# and/or modify it under the terms of the GNU General Public
-# License as published by the Free Software Foundation; either
-# version 2, or (at your option) any later version.
-#
-# GNU Enterprise is distributed in the hope that it will be
-# useful, but WITHOUT ANY WARRANTY; without even the implied
-# warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
-# PURPOSE. See the GNU General Public License for more details.
-#
-# You should have received a copy of the GNU General Public
-# License along with program; see the file COPYING. If not,
-# write to the Free Software Foundation, Inc., 59 Temple Place
-# - Suite 330, Boston, MA 02111-1307, USA.
-#
-# $Id$
-
-"""
-Schema support plugin for PostgreSQL backends.
-"""
-
-import os
-
-from gnue.common.apps import errors
-from gnue.common.datasources.drivers import DBSIG2
-from gnue.common.datasources import GSchema
-
-
-# =============================================================================
-# Behavior class
-# =============================================================================
-
-class Behavior (DBSIG2.Behavior):
-  """
-  Behavior class for PostgreSQL backends.
-  """
-
-  # ---------------------------------------------------------------------------
-  # Constructor
-  # ---------------------------------------------------------------------------
-
-  def __init__ (self, *args, **kwargs):
-
-    DBSIG2.Behavior.__init__ (self, *args, **kwargs)
-
-    self._RELKIND = {'v': {'type': 'view',  'name': _("Views")},
-                     'r': {'type': 'table', 'name': _("Tables")}}
-
-    self._maxIdLength   = 31
-    self._alterMultiple = False
-    self._numbers       = [[(4, 'smallint'), (9, 'integer'), (18, 'bigint')],
-                           "numeric (%s,0)", "numeric (%(length)s,%(scale)s)"]
-
-    # Build  typemap: {nativetype: (group, fieldtype)}
-    self._TYPEMAP = {'date'  : ('date', 'date'),
-                     'bool'  : ('boolean', 'boolean'),
-                     'string': ('string', 'string')}
-
-    for item in ['numeric', 'float4', 'float8', 'money', 'int8',
-                 'int2', 'int4', 'serial']:
-      self._TYPEMAP [item] = ('number', 'number')
-
-    for item in ['time', 'reltime']:
-      self._TYPEMAP [item] = ('date', 'time')
-
-    for item in ['timestamp', 'abstime']:
-      self._TYPEMAP [item] = ('date', 'datetime')
-
-    self._type2native.update ({'boolean' : 'boolean',
-                               'datetime': 'timestamp without time zone'})
-
-
-  # ---------------------------------------------------------------------------
-  # Create a new database
-  # ---------------------------------------------------------------------------
-
-  def createDatabase (self):
-    """
-    Create the requested user and database using the tools 'createuser',
-    'createdb' and 'dropuser'. Of course this function should better make use
-    of the template1 database using a connection object.
-    """
-
-    dbname   = self.__connection.parameters.get ('dbname')
-    username = self.__connection.parameters.get ('username', 'gnue')
-    password = self.__connection.parameters.get ('password')
-    host     = self.__connection.parameters.get ('host')
-    port     = self.__connection.parameters.get ('port')
-
-    site = ""
-    if host is not None:
-      site += " --host=%s" % host
-    if port is not None:
-      site += " --port=%s" % port
-
-    # TODO: use a connection object to the template1 database instead of the
-    # shell-scripts. Note: CREATE DATABASE statements must NOT run within a
-    # transaction block, so we cannot use the default connection mechanisms.
-
-    try:
-      os.system (u"dropuser %s%s 2>/dev/null" % (username, site))
-
-    except:
-      pass
-
-    try:
-      createuser = u"createuser %s --createdb --adduser %s" % (site, username)
-      os.system (createuser)
-    except:
-      pass
-
-    createdb = u"createdb %s --owner=%s --encoding=UNICODE %s" \
-        % (site, username, dbname)
-
-    if os.system (createdb):
-      raise errors.ApplicationError, u_("Database creation failed")
-
-
-    self.__connection.manager.loginToConnection (self.__connection)
-
-    if password is not None and password:
-      alterUser = u"ALTER USER %s WITH PASSWORD '%s';" % (username, password)
-      self.__connection.makecursor (alterUser)
-      self.__connection.commit ()
-
-
-  # ---------------------------------------------------------------------------
-  # Read the current connection's schema
-  # ---------------------------------------------------------------------------
-
-  def _readSchema_ (self, parent):
-    """
-    Read the connection's schema and build a GSchema object tree connected to
-    the given parent object (which is of type GSSchema).
-    """
-
-    tables = self.__readTables (parent)
-    fields = self.__readFields (tables)
-    self.__readDefaults (fields)
-    self.__readKeys (tables)
-    self.__readConstraints (tables, fields)
-
-
-  # ---------------------------------------------------------------------------
-  # Read all table-like elements
-  # ---------------------------------------------------------------------------
-
-  def __readTables (self, parent):
-    
-    mapping = {}    # Maps OIDs to GSTable instances
-    tables  = None
-    views   = None
-
-    cmd = u"SELECT c.oid, c.relname, c.relkind " \
-           "FROM pg_class c, pg_namespace n " \
-           "WHERE n.nspname = 'public' AND n.oid = c.relnamespace AND " \
-           "      c.relkind in (%s) " \
-           "ORDER BY c.relname" \
-           % ','.join (["%r" % kind for kind in self._RELKIND.keys ()])
-
-    cursor = self.__connection.makecursor (cmd)
-
-    try:
-      for (oid, relname, relkind) in cursor.fetchall ():
-
-        kind = self._RELKIND [relkind] ['type']
-        properties = {'id': oid, 'name': relname, 'kind': kind}
-
-        if relkind == 'v':
-          if views is None:
-            views = GSchema.GSTables (parent, **self._RELKIND [relkind])
-          master = views
-        else:
-          if tables is None:
-            tables = GSchema.GSTables (parent, **self._RELKIND [relkind])
-          master = tables
-
-        table = GSchema.GSTable (master, **properties)
-
-        # Maintain a temporary mapping from OID's to GSTable instances so
-        # adding fields afterwards runs faster
-        mapping [oid] = table
-
-    finally:
-      cursor.close ()
-
-    return mapping
-
-
-  # ---------------------------------------------------------------------------
-  # Find all fields
-  # ---------------------------------------------------------------------------
-
-  def __readFields (self, tables):
-
-    cmd = u"SELECT attrelid, attname, t.typname, attnotnull, " \
-           "       atthasdef, atttypmod, attnum, attlen " \
-           "FROM pg_attribute a " \
-           "LEFT OUTER JOIN pg_type t ON t.oid = a.atttypid " \
-           "WHERE attnum >= 0 AND attisdropped = False " \
-           "ORDER BY attrelid, attnum"
-
-    cursor = self.__connection.makecursor (cmd)
-    fields = None
-    result = {}
-
-    try:
-      for rs in cursor.fetchall ():
-        (relid, name, typename, notnull, hasdef, typemod, attnum, attlen) = rs
-
-        # only process attributes from tables we've listed before
-        if not relid in tables:
-          continue
-
-        attrs = {'id'        : "%s.%s" % (relid, attnum),
-                 'name'      : name,
-                 'nativetype': typename,
-                 'nullable'  : hasdef or not notnull}
-
-        if typename.lower () in self._TYPEMAP:
-          (group, attrs ['type']) = self._TYPEMAP [typename.lower ()]
-        else:
-          (group, attrs ['type']) = self._TYPEMAP ['string']
-
-        if group == 'number':
-          if typemod != -1:
-            value = typemod - 4
-            attrs ['length']    = value >> 16
-            attrs ['precision'] = value &  0xFFFF
-
-          elif attlen > 0:
-            attrs ['length'] = len ("%s" % 2L ** (attlen * 8))
-
-        elif typemod != -1:
-          attrs ['length'] = typemod - 4
-
-        # Remove obsolete attributes
-        if group in ['date', 'boolean']:
-          for item in ['length', 'precision']:
-            if item in attrs:
-              del attrs [item]
-
-        elif group in ['string']:
-          if 'precision' in attrs:
-            del attrs ['precision']
-
-        table = tables [relid]
-        fields = table.findChildOfType ('GSFields')
-        if fields is None:
-          fields = GSchema.GSFields (table)
-
-        result [attrs ['id']] = GSchema.GSField (fields, **attrs)
-
-    finally:
-      cursor.close ()
-
-    return result
-
-
-  # ---------------------------------------------------------------------------
-  # Read defaults and apply them to the given fields
-  # ---------------------------------------------------------------------------
-
-  def __readDefaults (self, fields):
-
-    cmd = u"SELECT adrelid, adnum, adsrc FROM pg_attrdef ORDER BY adrelid"
-
-    cursor = self.__connection.makecursor (cmd)
-
-    try:
-      for (relid, fieldnum, source) in cursor.fetchall ():
-        field = fields.get ("%s.%s" % (relid, fieldnum))
-        
-        # Skip all defaults of not listed fields
-        if field is None:
-          continue
-
-        if source [:8] == 'nextval(':
-          field.defaultwith = 'serial'
-
-        elif source == 'now()':
-          field.defaultwith = 'timestamp'
-
-        else:
-          field.defaultwith = 'constant'
-          field.default = source.split ('::') [0].strip ("'")
-
-    finally:
-      cursor.close ()
-
-
-  # ---------------------------------------------------------------------------
-  # Read all indices and associate them with their table/view
-  # ---------------------------------------------------------------------------
-
-  def __readKeys (self, tables):
-
-    cmd = u"SELECT indrelid, indkey, indisunique, indisprimary, c.relname " \
-           "FROM pg_index i LEFT OUTER JOIN pg_class c ON c.oid = indexrelid"
-
-    cursor = self.__connection.makecursor (cmd)
-
-    try:
-      for (relid, fieldvec, isUnique, isPrimary, name) in cursor.fetchall ():
-
-        # Skip functional indices. A functional index is an index that is built
-        # upon a fuction manipulating a field upper(userid) vs userid
-        fields = [int (i) - 1 for i in fieldvec.split ()]
-        if not fields:
-          continue
-
-        # only process keys of listed tables
-        table = tables.get (relid)
-        if table is None:
-          continue
-
-        if isPrimary:
-          index  = GSchema.GSPrimaryKey (table, name = name)
-          fClass = GSchema.GSPKField
-        else:
-          indices = table.findChildOfType ('GSIndexes')
-          if indices is None:
-            indices = GSchema.GSIndexes (table)
-
-          index  = GSchema.GSIndex (indices, unique = isUnique, name = name)
-          fClass = GSchema.GSIndexField
-
-        fieldList = table.findChildrenOfType ('GSField', False, True)
-        for find in fields:
-          fClass (index, name = fieldList [find].name)
-
-    finally:
-      cursor.close ()
-
-
-  # ---------------------------------------------------------------------------
-  # Read all constraints
-  # ---------------------------------------------------------------------------
-
-  def __readConstraints (self, tables, fields):
-
-    cmd = u"SELECT conname, conrelid, confrelid, conkey, confkey, contype " \
-           "FROM pg_constraint WHERE contype in ('f', 'u')"
-
-    cursor = self.__connection.makecursor (cmd)
-    try:
-      for (name, relid, fkrel, key, fkey, ctype) in cursor.fetchall ():
-        table   = tables.get (relid)
-
-        if ctype == 'f':
-          fktable = tables.get (fkrel)
-
-          # We need both ends of a relation to be a valid constraint
-          if table is None or fktable is None:
-            continue
-
-          parent = table.findChildOfType ('GSConstraints')
-          if parent is None:
-            parent = GSchema.GSConstraints (table)
-
-          constr = GSchema.GSForeignKey (parent, name = name,
-                                                 references = fktable.name)
-
-          kp  = isinstance (key, basestring) and key [1:-1].split (',') or key
-          fkp = isinstance (fkey, basestring) and fkey [1:-1].split(',') or 
fkey
-
-          k = [fields ["%s.%s" % (relid, i)].name for i in kp]
-          f = [fields ["%s.%s" % (fkrel, i)].name for i in fkp]
-
-          for (name, refname) in zip (k, f):
-            GSchema.GSFKField (constr, name = name, references = refname)
-
-        # Unique-Constraint
-        elif ctype == 'u':
-          parent = table.findChildOfType ('GSConstraints') or \
-              GSchema.GSConstraints (table)
-          constr = GSchema.GSUnique (parent, name = name)
-          kp = isinstance (key, basestring) and key [1:-1].split (',') or key
-
-          for name in [fields ["%s.%s" % (relid, i)].name for i in kp]:
-            GSchema.GSUQField (constr, name = name)
-
-          # Ok, since we know PostgreSQL automatically creates a unique index
-          # of the same name, we drop that index since it would only confuse a
-          # later diff
-          for ix in table.findChildrenOfType ('GSIndex', False, True):
-            if ix.name == constr.name:
-              parent = ix.getParent ()
-              parent._children.remove (ix)
-              ix.setParent (None)
-
-    finally:
-      cursor.close ()
-
-
-  # ---------------------------------------------------------------------------
-  # Handle special defaults
-  # ---------------------------------------------------------------------------
-
-  def _defaultwith_ (self, code, field):
-    """
-    Create a sequence for 'serials' and set the default for 'timestamps'.
-
-    @param code: code-triple to get the result
-    @param field: GSField instance of the field having the default
-    """
-
-    if field.defaultwith == 'serial':
-      seq = self._getSequenceName_ (field)
-      code [0].append (u"CREATE SEQUENCE %s" % seq)
-      field.default = "DEFAULT nextval ('%s')" % seq
-
-    elif field.defaultwith == 'timestamp':
-      field.default = "DEFAULT now()"

Copied: trunk/gnue-common/src/datasources/drivers/sql/postgresql/Behavior.py 
(from rev 7667, 
trunk/gnue-common/src/datasources/drivers/sql/postgres/Behavior.py)

Modified: trunk/gnue-common/src/datasources/drivers/sql/postgresql/__init__.py
===================================================================
--- trunk/gnue-common/src/datasources/drivers/sql/postgres/__init__.py  
2005-06-28 11:27:42 UTC (rev 7666)
+++ trunk/gnue-common/src/datasources/drivers/sql/postgresql/__init__.py        
2005-07-01 08:07:10 UTC (rev 7668)
@@ -30,7 +30,7 @@
 # Define aliases for this plugin
 # =============================================================================
 
-__pluginalias__ = ['pgsql', 'postgresql']
+__pluginalias__ = ['pgsql', 'postgres']
 
 
 # =============================================================================