firebird-mirror/doc/WhatsNew

  **************
  * v1.5 Alpha 1
  **************

    * Fixed unregistered bug.
        A well-known "decompression overran buffer" error appears during the schema creation.

    * New ROWS_AFFECTED system variable (SF #451927).
        Return number of rows affected by the last INSERT/UPDATE/DELETE statement.
      Notes:
        1. Available in PSQL only.
        2. For any other statement than INSERT/UPDATE/DELETE, result is always zero.

    * Dynamic exception messages (SF #446240).
        Allow to throw an exception with another message than this exception has been created with.
      Syntax:
        EXCEPTION name [value];

    * New SQLCODE and GDSCODE system variables (SF #547383).
        Provide an access to a code of the catched error within the WHEN-block.
      Notes:
        1. Available in PSQL only.
        2. Makes sense in WHEN-block only, in other places returns 0 (success).

    * Exception re-initiate semantics.
        Allow an already catched exception to be re-thrown from the WHEN-block.
      Syntax:
        EXCEPTION;
      Note:
        Makes sense in WHEN-block only, in other places evaluates to no-op.

    * Fixed unregistered bug.
        The server crashes during the garbage collection under heavy load.

    * Deferred metadata compilation.
        Solve a lot of reasons of the well-known "object in use" error.

    * New NULL order handling.
        Allow user-defined ordering of NULLs.
      Syntax:
        [ORDER BY <order_list>]
        <order_list> = {col | int} [COLLATE collation]
          [ASC[ENDING] | DESC[ENDING]] [NULLS {FIRST | LAST}]
          [, <order_list> ...]
      Note:
        Default behaviour is NULLS LAST.

    * Fixed unregistered bug.
        gstat shows wrong value for maxdup element.

    * New registry key is used on win32.
        Currently it's SOFTWARE\FirebirdSQL\Firebird.

    * User-defined constraint index names (SF #451925).
        Allow an index name to be either constraint name or user-defined name.
      Syntax:
        <col_constraint> = [CONSTRAINT constraint]
          {UNIQUE [<constraint_index>]
          | PRIMARY KEY [<constraint_index>]
          | REFERENCES other_table [( other_col [, other_col ...])]
              [ON DELETE {NO ACTION|CASCADE|SET DEFAULT|SET NULL}]
              [ON UPDATE {NO ACTION|CASCADE|SET DEFAULT|SET NULL}]
              [<constraint_index>]
          | CHECK ( <search_condition>)}
        <tconstraint> = [CONSTRAINT constraint]
          {{PRIMARY KEY | UNIQUE} ( col [, col ...]) [<constraint_index>]
          | FOREIGN KEY ( col [, col ...]) REFERENCES other_table
              [ON DELETE {NO ACTION|CASCADE|SET DEFAULT|SET NULL}]
              [ON UPDATE {NO ACTION|CASCADE|SET DEFAULT|SET NULL}]
              [<constraint_index>]
          | CHECK ( <search_condition>)}
        <constraint_index> = USING [ASC[ENDING] | DESC[ENDING]] INDEX name
      Note:
        Index is named the same way as a constraint by default.

    * New RECREATE VIEW statement.
        A shorthand for DROP VIEW / CREATE VIEW couple of statements.
      Syntax:
        RECREATE VIEW name <view_definition>;

    * Fixed unregistered bug.
        Trigger which name starts with 'RDB$' cannot be altered or dropped at all.

    * Changed file names.
        Renamed distribution files to make sure we're Firebird. Now they're fbserver, fbclient, firebird.msg etc.
      Note:
        The client library is fbclient now and it should be used in all new FB-based projects. gds32 contains nothing but redirected exports and is provided for compatibility only.

    * Minor ODS upgrade.
        Added new system indices (RDB$INDEX_41, RDB$INDEX_42, RDB$INDEX_43), now ODS version is 10.1.

    * New CREATE OR ALTER statement (SF #451935).
        Allow either creating or altering a database object depending on its existance.
      Syntax:
        CREATE OR ALTER name <object_definition>;
      Notes:
        1. Applicable to SPs/triggers/views only.
        2. The statement has the same meaning as CREATE OR REPLACE one in Oracle.

    * Fixed unregistered bug.
        Broken dependencies (like DB$34) appear in the database after metadata changes.

    * New ALTER VIEW statement (SF #490261).
        Allow altering an existing view.
      Syntax:
        ALTER VIEW name [(<view_col> [, <view_col> ... ])] AS <select>;

    * Enhanced declaration of local variables.
        Simplify syntax and allow declaring and defining variable at the same time.
      Syntax:
        DECLARE [VARIABLE] name <variable_type> [{'=' | DEFAULT} value];
      Example:
        DECLARE my_var INTEGER = 123;

    * Disabled BREAK statement for triggers (like EXIT) due to known internal limitations.

    * Enhanced grouping (SF #555839, #546274).
        Allow to GROUP BY internal functions and subqueries. Also allow to GROUP BY ordinal (i.e. column position).

    * New COALESCE internal function (SF #451917).
        Allow a column value to be calculated by a number of expressions, the first expression returning a non NULL value is returned as the column value.
      Syntax:
        COALESCE (value {, value} ... )
      Notes:
        1. COALESCE (V1, V2) is equivalent to the following case specification:
             CASE WHEN V1 IS NOT NULL THEN V1 ELSE V2 END
        2. COALESCE (V1, V2, ..., Vn), for n >= 3, is equivalent to the following case specification:
             CASE WHEN V1 IS NOT NULL THEN V1 ELSE COALESCE (V2, ..., Vn) END
        3. The function has the same meaning as NVL one in Oracle.
      Example:
        SELECT
          PROJ_NAME AS Projectname,
          COALESCE(e.FULL_NAME, '[> not assigned <]') AS Employeename
        FROM
          PROJECT p LEFT JOIN EMPLOYEE e ON (e.EMP_NO = p.TEAM_LEADER)

    * New NULLIF internal function (SF #451917).
        Return a NULL value for a sub-expression if it has a specific value, otherwise return the value of the sub-expression.
      Syntax:
        NULLIF (value, value)
      Note:
        NULLIF (V1, V2) is equivalent to the following case specification:
          CASE WHEN V1 = V2 THEN NULL ELSE V1 END
      Example:
        UPDATE PRODUCTS
          SET STOCK = NULLIF(STOCK, 0)

    * New CASE internal function (SF #451917).
        Allow the result of a column to be determined by a the results of a case expression.
      Syntax:
        <case_specification> = <simple_case> | <searched_case>

        <simple_case> = CASE value <simple_when_clause> ... [ELSE value] END
        <simple_when_clause> = WHEN value THEN value

        <searched_case> = CASE <searched_when_clause> ... [ELSE value] END
        <searched_when_clause> = WHEN <search_condition> THEN value
      Note:
        The function has the same meaning as DECODE one in Oracle.
      Examples:
        1. Simple case:
          SELECT
            o.ID,
            o.Description,
            CASE o.Status
              WHEN 1 THEN 'confirmed'
              WHEN 2 THEN 'in production'
              WHEN 3 THEN 'ready'
              WHEN 4 THEN 'shipped'
              ELSE 'unknown status ''' || o.Status || ''''
            END
          FROM
            Orders o
        2. Searched case:
          SELECT
            o.ID,
            o.Description,
            CASE
              WHEN (o.Status IS NULL) THEN 'new'
              WHEN (o.Status = 1) THEN 'confirmed'
              WHEN (o.Status = 3) THEN 'in production'
              WHEN (o.Status = 4) THEN 'ready'
              WHEN (o.Status = 5) THEN 'shipped'
              ELSE 'unknown status ''' || o.Status || ''''
            END
          FROM
            Orders o

    * Fixed bug SF #545725.
        Automatic/background sweep hangs.

    * Fixed unregistered bug.
        The server crashes when XSQLDA structures are prepared for not all statement parameters.

    * Enabled support for empty BEGIN...END blocks.

    * Partially fixed bug SF #567931.
        Metadata security hole.

    * Fixed unregistered bug.
        INT64 arrays don't work.

    * Fixed bug SF #437859.
        Execute procedure and string concat.
      Note:
        This change allows any expression to be used as a SP parameter.

    * Fixed bug SF #562417.
        Aggregate concatenated empty char.

    * New BIGINT datatype (SF #446206).
        Allow native SQL usage of 64-bit exact numerics.
      Note:
        Available in dialect 3 only.

    * Universal triggers (SF #451922).
        Allow one trigger to be fired for a number of action types.
      Syntax:
        CREATE TRIGGER name FOR table
          [ACTIVE | INACTIVE]
          <trigger_action_prefix>
          <trigger_action_suffix> [OR <trigger_action_suffix>] [OR <trigger_action_suffix>]
          [POSITION number]
          AS <trigger_body>
        <trigger_action_prefix> = {BEFORE | AFTER}
        <trigger_action_suffix> = {DELETE | INSERT | UPDATE}
      Example:
        CREATE TRIGGER my_trigger FOR my_table BEFORE INSERT OR UPDATE
          AS BEGIN
            IF (NEW.DOC_ID IS NULL) THEN
              EXCEPTION my_exception;
          END

    * New CONNECTION_ID and TRANSACTION_ID system variables (SF #446238, #446243).
        Return appropriate internal identifier stored on the database header page.
      Notes:
        1. Available in SQL/PSQL.
        2. The values are reset after a database is restored.

    * Server-side database aliases (SF #446180).
        Any database can be attached to using an "alias" name instead of its physical pathname.
        The list of known database aliases is stored in aliases.conf file under the server installation root.
      Example:
        alias entry in the configuration file: my_database = d:\dbs\my\database.gdb
        connection string: localhost:my_database

    * In-memory sorting.
        If SORT plan is used for a SQL statement, the sorting is done in memory.
        If there's not enough memory for this operation, old temporary file based method is used.

    * Fixed bug SF #538201.
        Crash with extract from null as date.

    * "EXECUTE VARCHAR" statement (SF #446256).
        Allow execution of dynamic SQL statements in SPs/triggers.
      Syntax:
        EXECUTE VARCHAR value;
      Notes:
        1. Available in PSQL only.
        2. Currently cannot return values, so SELECT statements are not allowed.
        3. Recursion level is limited to 500 (hardcoded).
      Example:
        EXECUTE VARCHAR my_var;