[FIX] addons: avoid deleting resources after stand-alone base module update

When updating a database, base module is first loaded on its own, and this may cause useless deletion
of resource, because the aut-cleanup process. This process looks for records that are in the base namespace and
have not been processed, and deletes them because they are presumed removed from the newer version. This does not
properly work now that several modules install records (menus, groups) in the base namespace.
This is fixed by skipping the auto-cleanup process during the stand-alone loading of base, which is fine
because it will be processed properly after the rest of the modules are loaded/updated. 
(No specific LP bug reference, but many translation bugs stem partially from this)

bzr revid: odo@openerp.com-20101213161802-ulsece9bb3q1y1fx

bin/addons/__init__.py

@@ -599,8 +599,19 @@ class MigrationManager(object):
 
 log = logging.getLogger('init')
 
-def load_module_graph(cr, graph, status=None, perform_checks=True, **kwargs):
-
+def load_module_graph(cr, graph, status=None, perform_checks=True, skip_cleanup=False, **kwargs):
+    """Migrates+Updates or Installs all module nodes from ``graph``
+       :param graph: graph of module nodes to load
+       :param status: status dictionary for keeping track of progress
+       :param perform_checks: whether module descriptors should be checked for validity (prints warnings
+                              for same cases, and even raise osv_except if certificate is invalid)
+       :param skip_cleanup: whether the auto-cleanup of records should be executed (unlinks any object that
+                            appears to be from one of the updated modules, but have not been loaded during
+                            last loading (i.e. records that seem to have been removed from the module).
+                            This is best left disabled when loading stand-alone modules that could contain
+                            records from dependent modules (i.e. other modules have put records in their
+                            namespace)
+    """
     def process_sql_file(cr, fp):
         queries = fp.read().split(';')
         for query in queries:
@@ -762,7 +773,9 @@ def load_module_graph(cr, graph, status=None, perform_checks=True, **kwargs):
     for model in cr.dictfetchall():
         pool.get('ir.model').instanciate(cr, 1, model['model'], {})
 
-    pool.get('ir.model.data')._process_end(cr, 1, package_todo)
+    if not skip_cleanup:
+        # Cleanup orphan records
+        pool.get('ir.model.data')._process_end(cr, 1, package_todo)
     cr.commit()
 
     return has_updates
@@ -806,7 +819,7 @@ def load_modules(db, force_demo=False, status=None, update_module=False):
         if not graph:
             logger.notifyChannel('init', netsvc.LOG_CRITICAL, 'module base cannot be loaded! (hint: verify addons-path)')
             raise osv.osv.except_osv(_('Could not load base module'), _('module base cannot be loaded! (hint: verify addons-path)'))
-        has_updates = load_module_graph(cr, graph, status, perform_checks=(not update_module), report=report)
+        has_updates = load_module_graph(cr, graph, status, perform_checks=(not update_module), report=report, skip_cleanup=True)
 
         if update_module:
             modobj = pool.get('ir.module.module')