Version 3.4 of DES adds to previous version (3.3.2):

• Enhancements:
• New version of ACIDE 0.12, including as upgrades:
• New PDG panel, displayed either when a /pdg command is issued or on demand in the View menu. Display properties of this panel have been added to the application menu
• Relocatable panels. A panel can be dragged to another panel position, interchanging both panels' locations
  RA extended projection includes also the possibility of renamings
• Added listing of offending rules for set variables unsafety
• Added mode information for set variables unsafety
• Creation of incorrect views no longer displays the generic error message
• Some incorrect SQL statements due to aggregates are catched
• Constraints are checked along updates
• Datalog division operator can be nested
• Changes:
• Submitting an order_by related query overrides current order_answer flag. There is no need to submit /order_answer off to make sorted listings work
• Duplicate elimination (duplicates disabled and distinct operator) considers different nulls as the same null, therefore behaving more closely to SQL
• Variable order in compiled rules
• Fixed bugs:
• The division operation was incorrect in RA and SQL
• Aggregates included extra tuples along nested hypothetical assumptions
• Removed undeclared predicate warning for true/0 when occurring as a condition
• The predicate group_by was functioning incorrectly in some cases, depending on previous tabled results
• Undefined predicate warning removed for false/0
• SQL NOT EXISTS clause was incorrectly translated by adding a top-1 goal
• Computing the order_by predicate was omitted for /order_answer on. So, a query as top(1,order_by(t(X),[X])) might deliver an incorrect answer
• In multiline mode, when processing a file containing a tab and an SQL line remark (--), then the next line was considered as a part of the remark
• Built-ins were tried to be made persistent for persistent rules including them
• Retracting an IDB rule for a persistent predicate did not refresh the database schema
• Assuming IDB rules in an hypothetical implication lead to exceptions in some cases
• The combination of top and distinct metapredicates produced extra tuples
• Translation of the Datalog division operator lead to remove head variables not involved in the operation
• Anonymous variables in consulted rules were given names

Previous versions