Release Notes

Author : Ashley G Trüter
Updated: 14-Jan-2006
Purpose: This document describes changes made for each release.

Contents:

1. Release 2.0.2

This release corrects a critical flaw with multi-column DISTINCT use. Users of previous versions are strongly advised to upgrade to this version.

1.1 DISTINCT clause with multiple columns

Due to the way in which REBOL handles unique/skip, RebDB SELECT statements that used the DISTINCT clause and returned more than one column may have returned a subset of the correct rows. The following REBOL code demonstrates this problem:

>> unique/skip [1 "A" 1 "B"] 2
== [1 "A"]
>> unique/skip ["A" 1 "B" 1] 2
== ["A" 1 "B" 1]

RebDB v2.0.2 no longer uses unique/skip.

1.2 Compatibility

RebDB now requires Core 2.6.0 (included with View 1.3) or greater.

1.3 Lines

By default, RebDB data files are now formatted with one row of data per line. You can revert to the old behaviour by setting db/lines?: false, or using the new set lines option from the RebDB client (see below).

1.4 SQL client set options

1.4.1 Browser

The set browser option has been removed as the REBOL set-browser-path function has been depreciated.

1.4.2 Lines

RebDB's line mode can now be toggled on or off with the set lines option.

1.4.3 Path

RebDB's base data directory can now be set/changed with the set path option.

1.4.4 Help

The help command has been updated to reflect the above changes.

2. Release 2.0.1

This release is a maintenance release with one major enhancement: automatic log recovery.

2.1 Automatic log recovery

RebDB now automatically applies all log files at startup. You can see this behaviour for yourself by issuing the following statements:

SQL> create test [col]
SQL> insert into test values [0]
SQL> exit

and then restart the SQL client to see the following replay messages:

Replaying 1 change(s).
insert into test values [0]
Replay completed.

2.2 SQL client error messages

Underlying REBOL errors (e.g. math overflow) are now trapped and displayed correctly. The following statements demonstrate this:

SQL> create test [col]
SQL> insert into test values [2000000000]
SQL> insert into test values [2000000000]
SQL> select sum col from test