A major security issue fixed in this release, CVE-2013-1899, makes it possible for a connection request containing a database name that begins with "-" to be crafted that can damage or destroy files within a server's data directory. Anyone with access to the port the PostgreSQL server listens on can initiate this request.
Two lesser security fixes are also included in this release: CVE-2013-1900, wherein random numbers generated by contrib/pgcrypto functions may be easy for another database user to guess, and CVE-2013-1901, which mistakenly allows an unprivileged user to run commands that could interfere with in-progress backups. Finally, this release fixes two security issues with the graphical installers for Linux and Mac OS X: insecure passing of superuser passwords to a script, CVE-2013-1903 and the use of predictable filenames in /tmp CVE-2013-1902.
The security issue fixed in this release, CVE-2013-0255, allows a previously authenticated user to crash the server by calling an internal function with invalid arguments. This issue was discovered by independent security researcher Sumit Soni this week and reported via Secunia SVCRP, and we are grateful for their efforts in making PostgreSQL more secure.
Today's update also fixes a performance regression which caused a decrease in throughput when using dynamic queries in stored procedures in version 9.2. Applications which use PL/pgSQL's EXECUTE are strongly affected by this regression and should be updated. Additionally, we have fixed intermittent crashes caused by CREATE/DROP INDEX CONCURRENTLY, and multiple minor issues with replication.
This release is expected to be the final update for version 8.3, which is now End-of-Life (EOL). Users of version 8.3 should plan to upgrade to a later version of PostgreSQL immediately. For more information, see our Versioning Policy.
This update release also contains fixes for many minor issues discovered and patched by the PostgreSQL community in the last two months, including:
* Prevent unnecessary table scans during vacuuming
* Prevent spurious cached plan error in PL/pgSQL
* Allow sub-SELECTs to be subscripted
* Prevent DROP OWNED from dropping databases or tablespaces
* Make ECPG use translated messages
* Allow PL/Python to use multi-table trigger functions (again) in 9.1 and 9.2
* Fix several activity log management issues on Windows
* Prevent autovacuum file truncation from being cancelled by deadlock_timeout
* Make extensions build with the .exe suffix automatically on Windows
* Fix concurrency issues with CREATE/DROP DATABASE
* Reject out-of-range values in to_date() conversion function
* Revert cost estimation for large indexes back to pre-9.2 behavior
* Make pg_basebackup tolerate timeline switches
* Cleanup leftover temp table entries during crash recovery
* Prevent infinite loop when COPY inserts a large tuple into a table with a large fillfactor
* Prevent integer overflow in dynahash creation
* Make pg_upgrade work with INVALID indexes
* Fix bugs in TYPE privileges
* Allow Contrib installchecks to run in their own databases
* Many documentation updates
* Add new timezone "FET".
This update release also contains fixes for many minor issues discovered and patched by the PostgreSQL community in the last month, including several fixes which are specific to version 9.2. These include:
* Prevent psql crash due to incorrectly encoded input data
* Fix pg_regress gmake build issues
* Make sure correct directory is created for extensions
* Fix some issues with buffer locks and VACUUM
* Multiple fixes and improvements for pg_upgrade
* Fix bugs with end-of-recovery when failing over to a standby
* Avoid bogus "out-of-sequence timeline ID" errors in standby-mode
* Don't launch new child processes during shutdown
* Improve the ability of JOINs to use partial indexes
* Fix assorted integer overflow errors
* Eliminate memory leaks in record_out() and record_send()
* Skip searching for subtransaction logs at COMMIT
* Fix WaitLatch() timing issues
* Fix handling of inherited check constraints in ALTER COLUMN TYPE
* Make ALTER EXTENSION SET SCHEMA behave as documented
* Have SEQUENCE SET statements in the "data" section of sectional dumps
* Prevent parser from believing that VIEWs have system columns
* Fix --clean mode for pg_dump
* Prevent hash table corruption on out-of-memory
* Various query planner and executor fixes and improvements
* Multiple documentation updates
* DST updates for seven timezones
The update fixes two potential data corruption issues present in the PostgreSQL 9.1 and 9.2 for any server which has crashed, been shutdown with "immediate", or was failed over to a standby. First, the PostgreSQL development team has discovered a chance of corruption of BTREE and GIN indexes for databases. Second, there is a significant chance of corruption of the visibility map. This update fixes both issues.
We strongly advise users of 9.1 and 9.2 to run VACUUM and/or index rebuilds after applying the update. Please see the 2012-09-24 Update wiki page for detailed instructions.
This update release also contains fixes for many minor issues discovered and patched by the PostgreSQL community in the last month, including many fixes for the newly released version 9.2. These include:
fix sorting issue with IN lists and indexes
fix planner failure when combining GROUP BY with window functions
improve selectivity of text searches using prefixes
prevent rescanning of WITH clauses from giving wrong answers
fix PL/Perl crashing issue
reduce bloat for multi-column GiST indexes
time zone data changes for Fiji
disallow Extensions from circular schema assignment
prevent crashes when default_transaction_isolation is set to "serializable"
several minor fixes to pg_upgrade
* CVE-2012-3488: PostgreSQL insecure use of libxslt
* CVE-2012-3489: PostgreSQL insecure use of libxml2
* Updates and corrections to time zone data
* Multiple documentation updates and corrections
* Add limit on max_wal_senders
* Fix dependencies generated during ALTER TABLE ADD CONSTRAINT USING INDEX.
* Correct behavior of unicode conversions for PL/Python
* Fix WITH attached to a nested set operation (UNION/INTERSECT/EXCEPT).
* Fix syslogger so that log_truncate_on_rotation works in the first rotation.
* Only allow autovacuum to be auto-canceled by a directly blocked process.
* Improve fsync request queue operation
* Prevent corner-case core dump in rfree().
* Fix Walsender so that it responds correctly to timeouts and deadlocks
* Several PL/Perl fixes for encoding-related issues
* Make selectivity operators use the correct collation
* Prevent unsuitable slaves from being selected for synchronous replication
* Make REASSIGN OWNED work on extensions as well
* Fix race condition with ENUM comparisons
* Make NOTIFY cope with out-of-disk-space
* Fix memory leak in ARRAY subselect queries
* Reduce data loss at replication failover
* Fix behavior of subtransactions with Hot Standby
Changes:
* Fix incorrect password transformation in contrib/pgcrypto's DES crypt()
function (Solar Designer)
* Ignore SECURITY DEFINER and SET attributes for a procedural language's call
handler (Tom Lane)
* Allow numeric timezone offsets in timestamp input to be up to 16 hours away
from UTC (Tom Lane)
* Fix timestamp conversion to cope when the given time is exactly the last DST
transition time for the the current timezone (Tom Lane)
* Fix text to name and char to name casts to perform string truncation
correctly in multibyte encodings (Karl Schnaitter)
* Fix memory copying bug in to_tsquery() (Heikki Linnakangas)
* Fix planner's handling of outer PlaceHolderVars within subqueries (Tom Lane)
* Fix slow session startup when pg_attribute is very large (Tom Lane)
* Ensure sequential scans check for query cancel reasonably often (Merlin
Moncure)
* Ensure the Windows implementation of PGSemaphoreLock() clears
ImmediateInterruptOK before returning (Tom Lane)
* Show whole-row variables safely when printing views or rules (Abbas Butt, Tom
Lane)
* Fix COPY FROM to properly handle null marker strings that correspond to
invalid encoding (Tom Lane)
* Ensure autovacuum worker processes perform stack depth checking properly
(Heikki Linnakangas)
* Fix logging collector to not lose log coherency under high load (Andrew
Dunstan)
* Fix logging collector to ensure it will restart file rotation after receiving
SIGHUP (Tom Lane)
* Fix WAL replay logic for GIN indexes to not fail if the index was
subsequently dropped (Tom Lane)
* Fix memory leak in PL/pgSQL's RETURN NEXT command (Joe Conway)
* Fix PL/pgSQL's GET DIAGNOSTICS command when the target is the function's
first variable (Tom Lane)
* Fix potential access off the end of memory in psql's expanded display (\x)
mode (Peter Eisentraut)
* Fix several performance problems in pg_dump when the database contains many
objects (Jeff Janes, Tom Lane)
* Fix contrib/dblink's dblink_exec() to not leak temporary database connections
upon error (Tom Lane)
* Fix contrib/dblink to report the correct connection name in error messages
(Kyotaro Horiguchi)
* Update time zone data files to tzdata release 2012c for DST law changes in
Antarctica, Armenia, Chile, Cuba, Falkland Islands, Gaza, Haiti, Hebron,
Morocco, Syria, and Tokelau Islands; also historical corrections for Canada.
* Require execute permission on the trigger function for CREATE TRIGGER
* Convert newlines to spaces in names written in pg_dump comments
* Fix btree index corruption from insertions concurrent with vacuuming
* Allow non-existent values for some settings in ALTER USER/DATABASE SET
* Track the OID counter correctly during WAL replay, even when it wraps around
* Fix regular expression back-references with * attached
* Fix recently-introduced memory leak in processing of inet/cidr values
* Avoid double close of file handle in syslogger on Windows
* Fix I/O-conversion-related memory leaks in plpgsql
* Improve pg_dump's handling of inherited table columns
* Fix pg_restore's direct-to-database mode for INSERT-style table data
* Fix error in contrib/intarray's int[] & int[] operator
* Fix error detection in contrib/pgcrypto's encrypt_iv() and decrypt_iv()
* Fix one-byte buffer overrun in contrib/test_parser
* Use __sync_lock_test_and_set() for spinlocks on ARM, if available
* Use -fexcess-precision=standard option when building with gcc versions that
accept it
* Allow use of threaded Python on FreeBSD
* Fix bugs in information_schema.referential_constraints view
* Correct collations for citext columns and indexes
* Prevent possible crash when joining to a scalar function
* Prevent transitory data corruption of GIN indexes after a crash
* Prevent data corruption on TOAST columns when copying data
* Fix failures during hot standby startup
* Correct another "variable not found in subplan target list" bug
* Fix bug with sorting on aggregate expressions in windowing functions
* Multiple bug fixes for pg_upgrade
* Change Foreign Key creation order to better support self-referential keys
* Multiple bug fixes to CREATE EXTENSION
* Ensure that function return type and data returned from PL/perl agree
* Ensure that PL/perl strings are always UTF-8
* Assorted bug fixes for various Extensions
* Updates to the time zone database, particularly to CST6
* Fix bugs in indexing of in-doubt HOT-updated tuples
* Fix multiple bugs in GiST index page split processing
* Fix possible buffer overrun in tsvector_concat()
* Fix crash in xml_recv when processing a "standalone" parameter
* Make pg_options_to_table return NULL for an option with no value
* Avoid possibly accessing off the end of memory in "ANALYZE" and in
SJIS-2004 encoding conversion
* Prevent intermittent hang in interactions of startup process with
bgwriter process
* Fix race condition in relcache init file invalidation
* Fix memory leak at end of a GiST index scan
* Fix incorrect memory accounting (leading to possible memory bloat)
in tuplestores supporting holdable cursors and plpgsql's RETURN
NEXT command
* Fix performance problem when constructing a large, lossy bitmap
* Fix join selectivity estimation for unique columns
* Fix nested PlaceHolderVar expressions that appear only in
sub-select target lists
* Allow nested EXISTS queries to be optimized properly
* Fix array- and path-creating functions to ensure padding bytes are zeroes
* Fix "EXPLAIN" to handle gating Result nodes within inner-indexscan subplans
* Work around gcc 4.6.0 bug that breaks WAL replay
* Fix dump bug for VALUES in a view
* Disallow SELECT FOR UPDATE/SHARE on sequences
This operation doesn't work as expected and can lead to failures.
* Fix "VACUUM" so that it always updates pg_class.reltuples/relpages
* more...
* fix build issues on HP-UX and Itanium
* update time zone files for recent time zone changes
* change SQLSTATE for Hot Standby warnings
* prevent bgwriter hang during recovery
* prevent recursive composite type creation
* disallow dropping tables whose triggers are still pending
* allow use of "replication" as a user name
* prevent a crash during GEQO planning
* improve join plans for tables with missing statistics
* fix error with SELECT FOR UPDATE in a subselect
* close PL/python array slice memory leak
* allow SSL connections for users with no home directory
* Avoid failures when "EXPLAIN" tries to display a simple-form CASE
expression.
* Fix assignment to an array slice that is before the existing range
of subscripts.
* Avoid unexpected conversion overflow in planner for very distant
date values.
* Fix pg_restore's text output for large objects (BLOBs) when
standard_conforming_strings is on.
* Fix erroneous parsing of tsquery values containing ... &
!(subexpression) | ...
* Fix buffer overrun in "contrib/intarray"'s input function for the
query_int type.
* Fix bug in "contrib/seg"'s GiST picksplit algorithm.
* Force the default wal_sync_method to be fdatasync on Linux
* Fix assorted bugs in WAL replay logic for GIN indexes
* Fix recovery from base backup when the starting checkpoint WAL
record is not in the same WAL segment as its redo point
* Fix persistent slowdown of autovacuum workers when multiple workers
remain active for a long time
* Add support for detecting register-stack overrun on IA64
* Add a check for stack overflow in copyObject()
* Fix detection of page splits in temporary GiST indexes
* Fix error checking during early connection processing
* Improve efficiency of window functions
* Avoid memory leakage while "ANALYZE"'ing complex index expressions
* Ensure an index that uses a whole-row Var still depends on its table
* Do not "inline" a SQL function with multiple OUT parameters
* Behave correctly if ORDER BY, LIMIT, FOR UPDATE, or WITH is
attached to the VALUES part of INSERT ... VALUES
* Fix constant-folding of COALESCE() expressions
* Fix postmaster crash when connection acceptance (accept() or one of
the calls made immediately after it) fails, and the postmaster was
compiled with GSSAPI support
* Fix missed unlink of temporary files when log_temp_files is active
* Add print functionality for InhRelation nodes
* Fix incorrect calculation of distance from a point to a horizontal
line segment
* Fix incorrect calculation of transaction status in ecpg
* Fix PL/pgSQL's handling of "simple" expressions to not fail in
recursion or error-recovery cases
* Fix PL/Python's handling of set-returning functions
* Fix bug in "contrib/cube"'s GiST picksplit algorithm
* Don't emit "identifier will be truncated" notices in
"contrib/dblink" except when creating new connections
* Fix potential coredump on missing public key in "contrib/pgcrypto"
* Fix memory leak in "contrib/xml2"'s XPath query functions
* Update time zone data files to tzdata release 2010o for DST law
changes in Fiji and Samoa; also historical corrections for Hong
Kong.
* Use a separate interpreter for each calling SQL userid in PL/Perl and PL/Tcl
* Prevent possible crashes in pg_get_expr() by disallowing it from being called
with an argument that is not one of the system catalog columns it's intended
to be used with
* Treat exit code 128 (ERROR_WAIT_NO_CHILDREN) as non-fatal on Windows
* Fix incorrect placement of placeholder evaluation
* Fix possible duplicate scans of UNION ALL member relations
* Fix "cannot handle unplanned sub-select" error
* Fix mishandling of whole-row Vars that reference a view or sub-select and
appear within a nested sub-select
* Fix mishandling of cross-type IN comparisons
* Fix computation of "ANALYZE" statistics for tsvector columns
* Improve planner's estimate of memory used by array_agg(), string_agg(), and
similar aggregate functions
* Fix failure to mark cached plans as transient
* Reduce PANIC to ERROR in some occasionally-reported btree failure cases, and
provide additional detail in the resulting error messages
* Fix incorrect search logic for partial-match queries with GIN indexes
* Prevent show_session_authorization() from crashing within autovacuum
processes
* Defend against functions returning setof record where not all the returned
rows are actually of the same rowtype
* Fix possible corruption of pending trigger event lists during subtransaction
rollback
* Fix possible failure when hashing a pass-by-reference function result
* Improve merge join's handling of NULLs in the join columns
* Take care to fsync the contents of lockfiles (both "postmaster.pid" and the
socket lockfile) while writing them
* Avoid recursion while assigning XIDs to heavily-nested subtransactions
* Avoid holding open old WAL segments in the walwriter process
* Fix log_line_prefix's %i escape, which could produce junk early in backend
startup
* Prevent misinterpretation of partially-specified relation options
for TOAST tables
* Fix inheritance count tracking in "ALTER TABLE ... ADD CONSTRAINT"
* Fix possible data corruption in "ALTER TABLE ... SET TABLESPACE" when
archiving is enabled
* Allow "CREATE DATABASE" and "ALTER DATABASE ... SET TABLESPACE" to be
interrupted by query-cancel
* Improve "CREATE INDEX"'s checking of whether proposed index
expressions are immutable
* Fix "REASSIGN OWNED" to handle operator classes and families
* Fix possible core dump when comparing two empty tsquery values
* Fix LIKE's handling of patterns containing % followed by _
* Enforce restrictions in plperl using an opmask applied to the whole
interpreter, instead of using "Safe.pm"
* Prevent PL/Tcl from executing untrustworthy code from pltcl_modules
* Fix data corruption during WAL replay of ALTER ... SET TABLESPACE
* Fix possible crash if a cache reset message is received during
rebuild of a relcache entry
* Apply per-function GUC settings while running the language
validator for the function
* Do constraint exclusion for inherited "UPDATE" and "DELETE" target
tables when constraint_exclusion = partition
* Do not allow an unprivileged user to reset superuser-only parameter
settings
* Avoid possible crash during backend shutdown if shutdown occurs
when a CONTEXT addition would be made to log entries
* Fix erroneous handling of %r parameter in recovery_end_command
* Ensure the archiver process responds to changes in archive_command
as soon as possible
* Fix pl/pgsql's CASE statement to not fail when the case expression
is a query that returns no rows
* Update pl/perl's "ppport.h" for modern Perl versions
* Fix assorted memory leaks in pl/python
* Handle empty-string connect parameters properly in ecpg
* Prevent infinite recursion in psql when expanding a variable that
refers to itself
* Fix psql's \copy to not add spaces around a dot within \copy
(select ...)
* Avoid formatting failure in psql when running in a locale context
that doesn't match the client_encoding
* Fix unnecessary "GIN indexes do not support whole-index scans"
errors for unsatisfiable queries using "contrib/intarray" operators
* Ensure that "contrib/pgstattuple" functions respond to cancel
interrupts promptly
* Make server startup deal properly with the case that shmget()
returns EINVAL for an existing shared memory segment
* Avoid possible crashes in syslogger process on Windows
* Deal more robustly with incomplete time zone information in the
Windows registry
* Update the set of known Windows time zone names
* Update time zone data files to tzdata release 2010j for DST law
changes in Argentina, Australian Antarctic, Bangladesh, Mexico,
Morocco, Pakistan, Palestine, Russia, Syria, Tunisia; also
historical corrections for Taiwan.
Also, add PKST (Pakistan Summer Time) to the default set of
timezone abbreviations.
* Protect against indirect security threats caused by index functions
changing session-local state
* Reject SSL certificates containing an embedded null byte in the
common name (CN) field
* Fix hash index corruption
* Fix possible crash during backend-startup-time cache initialization
* Avoid crash on empty thesaurus dictionary
* Prevent signals from interrupting VACUUM at unsafe times
* Fix possible crash due to integer overflow in hash table size
calculation
* Fix crash if a DROP is attempted on an internally-dependent object
* Fix very rare crash in inet/cidr comparisons
* Ensure that shared tuple-level locks held by prepared transactions
are not ignored
* Fix premature drop of temporary files used for a cursor that is
accessed within a subtransaction
* Fix memory leak in syslogger process when rotating to a new CSV
logfile
* Fix memory leak in postmaster when re-parsing "pg_hba.conf"
* Fix Windows permission-downgrade logic
* Make FOR UPDATE/SHARE in the primary query not propagate into WITH
queries
* Fix bug with a WITH RECURSIVE query immediately inside another one
* Fix concurrency bug in hash indexes
* Fix incorrect logic for GiST index page splits, when the split
depends on a non-first column of the index
* Fix wrong search results for a multi-column GIN index with
fastupdate enabled
* More...
* Fix WAL page header initialization at the end of archive recovery
* Fix "cannot make new WAL entries during recovery" error
* Fix problem that could make expired rows visible after a crash
* Disallow "RESET ROLE" and "RESET SESSION AUTHORIZATION" inside
security-definer functions
* Make "LOAD" of an already-loaded loadable module into a no-op
* Make window function PARTITION BY and ORDER BY items always be
interpreted as simple expressions
* Fix several errors in planning of semi-joins
* Fix handling of whole-row references to subqueries that are within
an outer join
* Fix Windows shared-memory allocation code
* Fix locale handling with plperl
* Fix handling of reloptions to ensure setting one option doesn't
force default values for others
* Ensure that a "fast shutdown" request will forcibly terminate open
sessions, even if a "smart shutdown" was already in progress
* Avoid memory leak for array_agg() in GROUP BY queries
* Treat to_char(..., 'TH') as an uppercase ordinal suffix with 'HH'/'HH12'
monitoring tools and commands. Each user has their own favorite features which
will make day-to-day work with PostgreSQL easier and more productive for them.
Among the most popular enhancements are:
* Parallel Database Restore, speeding up recovery from backup up to 8 times
* Per-Column Permissions, allowing more granular control of sensitive data
* Per-database Collation Support, making PostgreSQL more useful in
multi-lingual environments
* In-place Upgrades through pg_migrator (beta), enabling upgrades from 8.3 to
8.4 without extensive downtime
* New Query Monitoring Tools, giving administrators more insight into query
activity
* Greatly Reduced VACUUM Overhead through the Visibility Map
* New Monitoring Tools for current queries, query load and deadlocks
Version 8.4 also makes data analysis easier through the advanced ANSI SQL2003
features of windowing functions, common table expressions and recursive queries.
Enhancements to stored procedures, such as default parameters and variadic
parameters, make database server programming simpler and more compact.
Of course, there are also performance improvements included in this version.
