We are pleased to announce general availability of Slony-I 1.2.11.
Note that both sources and documentation are available; see
<http://slony.info/>

Fixes are the following:

    * Add in tools/mkservice scripts previously added to CVS HEAD
    * During subscription, do UPDATE to pg_class.relhasindex *after* the TRUNCATE because, in 8.2+, TRUNCATE resets this attribute
    * Fixed a problem with the setsync tracking with Log Shipping in cases where slon does an internal restart (thereby rereading the pset.ssy_seqno) and ignoring non-SYNC events because those don't change the sl_setsync table.
    * More explicit type casting of text objects for compatibility with PostgreSQL 8.3
    * Fixed problem with DDL SCRIPT statement parser: it wasn't 'quoting' semicolons inside parentheses (this notably occurs in CREATE RULE).
    * Fixed problem with DDL SCRIPT statement submission; it was interpreting the statement as a format string, which would have ill effects in the presence of things that are interpreted such as format strings (%d, %f, %s) and \backslashed things like \\, \n.
    * Further DDL Script issue: non-terminated statement at the end (e.g. - without trailing semicolon ";") would get omitted.
    * Typo fix: when trying to disable a node, the logs would report "enableNode" rather than "disableNode". Fixed.
    * Add usage/version options to help output in slon.
    * Fix archive logging for replicated sequences.
    * Fix to log shipping - added another table, sl_archive_counter where the log writing slon simply tracks when it wrote the last offline archive file and maintains a counter. This counter is now tracked in the offline replica and must increment gap free.
    * Change the filenames of archive logs to be based on internal archive tracking number. This makes it easy for the mechanism applying archives to figure out what needs to be applied next - just look in sl_archive_tracking.
    * Fix log shipping test to accomodate the new tracking scheme, and update documentation to describe this better. 

The testing regimen included running all the tests in the test bed
against instances of Slony-I running against PostgreSQL versions 7.4,
8.0, 8.1, 8.2, as well as CVS HEAD (forthcoming as 8.3). There was
also (against 7.4) a test done to verify that UPGRADE FUNCTIONS would
bring version 1.1.x to 1.2.11.
-- 
let name="cbbrowne" and tld="cbbrowne.com" in String.concat "@" [name;tld];;
http://linuxfinances.info/info/nonrdbms.html
Why  are  they  called   buildings,  when  they're  already  finished?
Shouldn't they be called builts?