E.168. Release 7.3.3

E.168.1. Migration to Version 7.3.3

A dump/restore is not required for those running version 7.3.*.

E.168.2. Changes

• Repair sometimes-incorrect computation of StartUpID after a crash

• Avoid slowness with lots of deferred triggers in one transaction (Stephan)

• Don't lock referenced row when UPDATE doesn't change foreign key's value (Jan)

• Use -fPIC not -fpic on Sparc (Tom Callaway)

• Repair lack of schema-awareness in contrib/reindexdb

• Fix contrib/intarray error for zero-element result array (Teodor)

• Ensure createuser script will exit on control-C (Oliver)

• Fix errors when the type of a dropped column has itself been dropped

• CHECKPOINT does not cause database panic on failure in noncritical steps

• Accept 60 in seconds fields of timestamp, time, interval input values

• Issue notice, not error, if TIMESTAMP, TIME, or INTERVAL precision too large

• Fix abstime-to-time cast function (fix is not applied unless you initdb)

• Fix pg_proc entry for timestampt_izone (fix is not applied unless you initdb)

• Make EXTRACT(EPOCH FROM timestamp without time zone) treat input as local time

• 'now'::timestamptz gave wrong answer if timezone changed earlier in transaction

• HAVE_INT64_TIMESTAMP code for time with timezone overwrote its input

• Accept GLOBAL TEMP/TEMPORARY as a synonym for TEMPORARY

• Avoid improper schema-privilege-check failure in foreign-key triggers

• Fix bugs in foreign-key triggers for SET DEFAULT action

• Fix incorrect time-qual check in row fetch for UPDATE and DELETE triggers

• Foreign-key clauses were parsed but ignored in ALTER TABLE ADD COLUMN

• Fix createlang script breakage for case where handler function already exists

• Fix misbehavior on zero-column tables in pg_dump, COPY, ANALYZE, other places

• Fix misbehavior of func_error() on type names containing '%'

• Fix misbehavior of replace() on strings containing '%'

• Regular-expression patterns containing certain multibyte characters failed

• Account correctly for NULLs in more cases in join size estimation

• Avoid conflict with system definition of isblank() function or macro

• Fix failure to convert large code point values in EUC_TW conversions (Tatsuo)

• Fix error recovery for SSL_read/SSL_write calls

• Don't do early constant-folding of type coercion expressions

• Validate page header fields immediately after reading in any page

• Repair incorrect check for ungrouped variables in unnamed joins

• Fix buffer overrun in to_ascii (Guido Notari)

• contrib/ltree fixes (Teodor)

• Fix core dump in deadlock detection on machines where char is unsigned

• Avoid running out of buffers in many-way indexscan (bug introduced in 7.3)

• Fix planner's selectivity estimation functions to handle domains properly

• Fix dbmirror memory-allocation bug (Steven Singer)

• Prevent infinite loop in ln(numeric) due to roundoff error

• GROUP BY got confused if there were multiple equal GROUP BY items

• Fix bad plan when inherited UPDATE/DELETE references another inherited table

• Prevent clustering on incomplete (partial or non-NULL-storing) indexes

• Service shutdown request at proper time if it arrives while still starting up

• Fix left-links in temporary indexes (could make backwards scans miss entries)

• Fix incorrect handling of client_encoding setting in postgresql.conf (Tatsuo)

• Fix failure to respond to pg_ctl stop -m fast after Async_NotifyHandler runs

• Fix SPI for case where rule contains multiple statements of the same type

• Fix problem with checking for wrong type of access privilege in rule query

• Fix problem with EXCEPT in CREATE RULE

• Prevent problem with dropping temp tables having serial columns

• Fix replace_vars_with_subplan_refs failure in complex views

• Fix regexp slowness in single-byte encodings (Tatsuo)

• Allow qualified type names in CREATE CAST and DROP CAST

• Accept SETOF type[], which formerly had to be written SETOF _type

• Fix pg_dump core dump in some cases with procedural languages

• Force ISO datestyle in pg_dump output, for portability (Oliver)

• pg_dump failed to handle error return from lo_read (Oleg Drokin)

• pg_dumpall failed with groups having no members (Nick Eskelinen)

• pg_dumpall failed to recognize --globals-only switch

• pg_restore failed to restore blobs if -X disable-triggers is specified

• Repair intrafunction memory leak in plpgsql

• pltcl's elog command dumped core if given wrong parameters (Ian Harding)

• plpython used wrong value of atttypmod (Brad McLean)

• Fix improper quoting of boolean values in Python interface (D'Arcy)

• Added addDataType() method to PGConnection interface for JDBC

• Fixed various problems with updateable ResultSets for JDBC (Shawn Green)

• Fixed various problems with DatabaseMetaData for JDBC (Kris Jurka, Peter Royal)

• Fixed problem with parsing table ACLs in JDBC

• Better error message for character set conversion problems in JDBC