Created by Ubuntu Package Importer on 2012-02-28 and last modified on 2013-05-04
Get this branch:
bzr branch lp:ubuntu/oneiric-security/postgresql-9.1
Members of Ubuntu branches can upload to this branch. Log in for directions.

Branch merges

Related bugs

Related blueprints

Branch information

Ubuntu branches
Review team:
Ubuntu Development Team

Recent revisions

15. By Martin Pitt on 2013-04-02

* New upstream security/bug fix release: (LP: #1163184)
  - Fix insecure parsing of server command-line switches.
    A connection request containing a database name that begins with
    "-" could be crafted to damage or destroy files within the server's
    data directory, even if the request is eventually rejected.
  - Reset OpenSSL randomness state in each postmaster child process.
    This avoids a scenario wherein random numbers generated by
    "contrib/pgcrypto" functions might be relatively easy for another
    database user to guess. The risk is only significant when the
    postmaster is configured with ssl = on but most connections don't
    use SSL encryption. [CVE-2013-1900]
  - Make REPLICATION privilege checks test current user not
    authenticated user.
    An unprivileged database user could exploit this mistake to call
    pg_start_backup() or pg_stop_backup(), thus possibly interfering
    with creation of routine backups. [CVE-2013-1901]
  - Fix GiST indexes to not use "fuzzy" geometric comparisons when it's
    not appropriate to do so.
    The core geometric types perform comparisons using "fuzzy"
    equality, but gist_box_same must do exact comparisons, else GiST
    indexes using it might become inconsistent. After installing this
    update, users should "REINDEX" any GiST indexes on box, polygon,
    circle, or point columns, since all of these use gist_box_same.
  - Fix erroneous range-union and penalty logic in GiST indexes that
    use "contrib/btree_gist" for variable-width data types, that is
    text, bytea, bit, and numeric columns.
    These errors could result in inconsistent indexes in which some
    keys that are present would not be found by searches, and also in
    useless index bloat. Users are advised to "REINDEX" such indexes
    after installing this update.
  - Fix bugs in GiST page splitting code for multi-column indexes.
    These errors could result in inconsistent indexes in which some
    keys that are present would not be found by searches, and also in
    indexes that are unnecessarily inefficient to search. Users are
    advised to "REINDEX" multi-column GiST indexes after installing
    this update.
  - See HISTORY/changelog.gz for details about the other bug fixes.

14. By Martin Pitt on 2013-02-05

* New upstream security/bug fix release: (LP: #1116336)
  - Prevent execution of enum_recv from SQL
    The function was misdeclared, allowing a simple SQL command to crash the
    server. In principle an attacker might be able to use it to examine the
    contents of server memory. Our thanks to Sumit Soni (via Secunia SVCRP)
    for reporting this issue. (CVE-2013-0255)
  - See HISTORY/changelog.gz for the other bug fixes.

13. By Jamie Strandboge on 2012-08-16

* New upstream bug fix/security release:
 - Prevent access to external files/URLs via XML entity references
   (Noah Misch, Tom Lane)
   xml_parse() would attempt to fetch external files or URLs as needed
   to resolve DTD and entity references in an XML value, thus allowing
   unprivileged database users to attempt to fetch data with the
   privileges of the database server. While the external data wouldn't
   get returned directly to the user, portions of it could be exposed
   in error messages if the data didn't parse as valid XML; and in any
   case the mere ability to check existence of a file might be useful
   to an attacker. (CVE-2012-3489)
 - Prevent access to external files/URLs via "contrib/xml2"'s
   xslt_process() (Peter Eisentraut)
   libxslt offers the ability to read and write both files and URLs
   through stylesheet commands, thus allowing unprivileged database
   users to both read and write data with the privileges of the
   database server. Disable that through proper use of libxslt's
   security options. (CVE-2012-3488)
   Also, remove xslt_process()'s ability to fetch documents and
   stylesheets from external files/URLs. While this was a documented
   "feature", it was long regarded as a bad idea. The fix for
   CVE-2012-3489 broke that capability, and rather than expend effort
   on trying to fix it, we're just going to summarily remove it.
 - Prevent too-early recycling of btree index pages (Noah Misch)
   When we allowed read-only transactions to skip assigning XIDs, we
   introduced the possibility that a deleted btree page could be
   recycled while a read-only transaction was still in flight to it.
   This would result in incorrect index search results. The
   probability of such an error occurring in the field seems very low
   because of the timing requirements, but nonetheless it should be
 - Fix crash-safety bug with newly-created-or-reset sequences (Tom
   If "ALTER SEQUENCE" was executed on a freshly created or reset
   sequence, and then precisely one nextval() call was made on it, and
   then the server crashed, WAL replay would restore the sequence to a
   state in which it appeared that no nextval() had been done, thus
   allowing the first sequence value to be returned again by the next
   nextval() call. In particular this could manifest for serial
   columns, since creation of a serial column's sequence includes an
 - Fix race condition in enum-type value comparisons (Robert Haas, Tom
   Comparisons could fail when encountering an enum value added since
   the current query started.
 - Fix txid_current() to report the correct epoch when not in hot
   standby (Heikki Linnakangas)
   This fixes a regression introduced in the previous minor release.
 - Prevent selection of unsuitable replication connections as the
   synchronous standby (Fujii Masao)
   The master might improperly choose pseudo-servers such as
   pg_receivexlog or pg_basebackup as the synchronous standby, and
   then wait indefinitely for them.
 - Fix bug in startup of Hot Standby when a master transaction has
   many subtransactions (Andres Freund)
   This mistake led to failures reported as "out-of-order XID
   insertion in KnownAssignedXids".
 - Ensure the "backup_label" file is fsync'd after pg_start_backup()
   (Dave Kerr)
 - Fix timeout handling in walsender processes (Tom Lane)
   WAL sender background processes neglected to establish a SIGALRM
   handler, meaning they would wait forever in some corner cases where
   a timeout ought to happen.
 - Wake walsenders after each background flush by walwriter (Andres
   Freund, Simon Riggs)
   This greatly reduces replication delay when the workload contains
   only asynchronously-committed transactions.
 - Fix LISTEN/NOTIFY to cope better with I/O problems, such as out of
   disk space (Tom Lane)
   After a write failure, all subsequent attempts to send more NOTIFY
   messages would fail with messages like "Could not read from file
   "pg_notify/nnnn" at offset nnnnn: Success".
 - Only allow autovacuum to be auto-canceled by a directly blocked
   process (Tom Lane)
   The original coding could allow inconsistent behavior in some
   cases; in particular, an autovacuum could get canceled after less
   than deadlock_timeout grace period.
 - Improve logging of autovacuum cancels (Robert Haas)
 - Fix log collector so that log_truncate_on_rotation works during the
   very first log rotation after server start (Tom Lane)
 - Fix WITH attached to a nested set operation
 - Ensure that a whole-row reference to a subquery doesn't include any
   extra GROUP BY or ORDER BY columns (Tom Lane)
 - Fix dependencies generated during ALTER TABLE ... ADD CONSTRAINT
   USING INDEX (Tom Lane)
   This command left behind a redundant pg_depend entry for the index,
   which could confuse later operations, notably ALTER TABLE ... ALTER
   COLUMN TYPE on one of the indexed columns.
 - Fix "REASSIGN OWNED" to work on extensions (Alvaro Herrera)
 - Disallow copying whole-row references in CHECK constraints and
   index definitions during "CREATE TABLE" (Tom Lane)
   This situation can arise in "CREATE TABLE" with LIKE or INHERITS.
   The copied whole-row variable was incorrectly labeled with the row
   type of the original table not the new one. Rejecting the case
   seems reasonable for LIKE, since the row types might well diverge
   later. For INHERITS we should ideally allow it, with an implicit
   coercion to the parent table's row type; but that will require more
   work than seems safe to back-patch.
 - Fix memory leak in ARRAY(SELECT ...) subqueries (Heikki
   Linnakangas, Tom Lane)
 - Fix planner to pass correct collation to operator selectivity
   estimators (Tom Lane)
   This was not previously required by any core selectivity estimation
   function, but third-party code might need it.
 - Fix extraction of common prefixes from regular expressions (Tom
   The code could get confused by quantified parenthesized
   subexpressions, such as ^(foo)?bar. This would lead to incorrect
   index optimization of searches for such patterns.
 - Fix bugs with parsing signed "hh":"mm" and "hh":"mm":"ss" fields in
   interval constants (Amit Kapila, Tom Lane)
 - Fix pg_dump to better handle views containing partial GROUP BY
   lists (Tom Lane)
   A view that lists only a primary key column in GROUP BY, but uses
   other table columns as if they were grouped, gets marked as
   depending on the primary key. Improper handling of such primary key
   dependencies in pg_dump resulted in poorly-ordered dumps, which at
   best would be inefficient to restore and at worst could result in
   outright failure of a parallel pg_restore run.
 - In PL/Perl, avoid setting UTF8 flag when in SQL_ASCII encoding
   (Alex Hunsaker, Kyotaro Horiguchi, Alvaro Herrera)
 - Use Postgres' encoding conversion functions, not Python's, when
   converting a Python Unicode string to the server encoding in
   PL/Python (Jan Urbanski)
   This avoids some corner-case problems, notably that Python doesn't
   support all the encodings Postgres does. A notable functional
   change is that if the server encoding is SQL_ASCII, you will get
   the UTF-8 representation of the string; formerly, any non-ASCII
   characters in the string would result in an error.
 - Fix mapping of PostgreSQL encodings to Python encodings in
   PL/Python (Jan Urbanski)
 - Report errors properly in "contrib/xml2"'s xslt_process() (Tom
 - Update time zone data files to tzdata release 2012e for DST law
   changes in Morocco and Tokelau

12. By Martin Pitt on 2012-06-04

* New upstream bug fix/security release: (LP: #1008317)
  - Fix incorrect password transformation in "contrib/pgcrypto"'s DES
    crypt() function.
    If a password string contained the byte value 0x80, the remainder
    of the password was ignored, causing the password to be much weaker
    than it appeared. With this fix, the rest of the string is properly
    included in the DES hash. Any stored password values that are
    affected by this bug will thus no longer match, so the stored
    values may need to be updated. (CVE-2012-2143)
  - Ignore SECURITY DEFINER and SET attributes for a procedural
    language's call handler. Applying such attributes to a call handler
    could crash the server. (CVE-2012-2655)
  - Make "contrib/citext"'s upgrade script fix collations of citext
    arrays and domains over citext.
    Release 9.1.2 provided a fix for collations of citext columns and
    indexes in databases upgraded or reloaded from pre-9.1
    installations, but that fix was incomplete: it neglected to handle
    arrays and domains over citext. This release extends the module's
    upgrade script to handle these cases. As before, if you have
    already run the upgrade script, you'll need to run the collation
    update commands by hand instead. See the 9.1.2 release notes for
    more information about doing this.
  - Allow numeric timezone offsets in timestamp input to be up to 16
    hours away from UTC. Some historical time zones have offsets larger than
    15 hours, the previous limit. This could result in dumped data values
    being rejected during reload.
  - Fix timestamp conversion to cope when the given time is exactly the
    last DST transition time for the current timezone.
    This oversight has been there a long time, but was not noticed
    previously because most DST-using zones are presumed to have an
    indefinite sequence of future DST transitions.
  - Fix text to name and char to name casts to perform string
    truncation correctly in multibyte encodings.
  - Fix memory copying bug in to_tsquery().
  - Ensure txid_current() reports the correct epoch when executed in
    hot standby.
  - Fix planner's handling of outer PlaceHolderVars within subqueries.
    This bug concerns sub-SELECTs that reference variables coming from
    the nullable side of an outer join of the surrounding query. In
    9.1, queries affected by this bug would fail with "ERROR:
    Upper-level PlaceHolderVar found where not expected". But in 9.0
    and 8.4, you'd silently get possibly-wrong answers, since the value
    transmitted into the subquery wouldn't go to null when it should.
  - Fix planning of UNION ALL subqueries with output columns that are
    not simple variables.
    Planning of such cases got noticeably worse in 9.1 as a result of a
    misguided fix for "MergeAppend child's targetlist doesn't match
    MergeAppend" errors. Revert that fix and do it another way.
  - Fix slow session startup when pg_attribute is very large.
    If pg_attribute exceeds one-fourth of shared_buffers, cache
    rebuilding code that is sometimes needed during session start would
    trigger the synchronized-scan logic, causing it to take many times
    longer than normal. The problem was particularly acute if many new
    sessions were starting at once.
  - Ensure sequential scans check for query cancel reasonably often.
    A scan encountering many consecutive pages that contain no live
    tuples would not respond to interrupts meanwhile.
  - Ensure the Windows implementation of PGSemaphoreLock() clears
    ImmediateInterruptOK before returning.
    This oversight meant that a query-cancel interrupt received later
    in the same query could be accepted at an unsafe time, with
    unpredictable but not good consequences.
  - Show whole-row variables safely when printing views or rules.
    Corner cases involving ambiguous names (that is, the name could be
    either a table or column name of the query) were printed in an
    ambiguous way, risking that the view or rule would be interpreted
    differently after dump and reload. Avoid the ambiguous case by
    attaching a no-op cast.
  - Fix "COPY FROM" to properly handle null marker strings that
    correspond to invalid encoding.
    A null marker string such as E'\\0' should work, and did work in
    the past, but the case got broken in 8.4.
  - Fix "EXPLAIN VERBOSE" for writable CTEs containing RETURNING
  - Fix "PREPARE TRANSACTION" to work correctly in the presence of
    advisory locks.
    Historically, "PREPARE TRANSACTION" has simply ignored any
    session-level advisory locks the session holds, but this case was
    accidentally broken in 9.1.
  - Fix truncation of unlogged tables.
  - Ignore missing schemas during non-interactive assignments of
    This re-aligns 9.1's behavior with that of older branches.
    Previously 9.1 would throw an error for nonexistent schemas
    mentioned in search_path settings obtained from places such as
  - Fix bugs with temporary or transient tables used in extension
    This includes cases such as a rewriting "ALTER TABLE" within an
    extension update script, since that uses a transient table behind
    the scenes.
  - Ensure autovacuum worker processes perform stack depth checking
    Previously, infinite recursion in a function invoked by
    auto-"ANALYZE" could crash worker processes.
  - Fix logging collector to not lose log coherency under high load.
    The collector previously could fail to reassemble large messages if
    it got too busy.
  - Fix logging collector to ensure it will restart file rotation after
    receiving SIGHUP.
  - Fix "too many LWLocks taken" failure in GiST indexes.
  - Fix WAL replay logic for GIN indexes to not fail if the index was
    subsequently dropped.
  - Correctly detect SSI conflicts of prepared transactions after a
  - Avoid synchronous replication delay when committing a transaction
    that only modified temporary tables.
    In such a case the transaction's commit record need not be flushed
    to standby servers, but some of the code didn't know that and
    waited for it to happen anyway.
  - Fix error handling in pg_basebackup.
  - Fix walsender to not go into a busy loop if connection is
  - Fix memory leak in PL/pgSQL's "RETURN NEXT" command.
  - Fix PL/pgSQL's "GET DIAGNOSTICS" command when the target is the
    function's first variable.
  - Ensure that PL/Perl package-qualifies the _TD variable.
    This bug caused trigger invocations to fail when they are nested
    within a function invocation that changes the current package.
  - Fix PL/Python functions returning composite types to accept a
    string for their result value.
    This case was accidentally broken by the 9.1 additions to allow a
    composite result value to be supplied in other formats, such as
  - Fix potential access off the end of memory in psql's expanded
    display ("\x") mode.
  - Fix several performance problems in pg_dump when the database
    contains many objects.
    pg_dump could get very slow if the database contained many schemas,
    or if many objects are in dependency loops, or if there are many
    owned sequences.
  - Fix memory and file descriptor leaks in pg_restore when reading a
    directory-format archive.
  - Fix pg_upgrade for the case that a database stored in a non-default
    tablespace contains a table in the cluster's default tablespace.
  - In ecpg, fix rare memory leaks and possible overwrite of one byte
    after the sqlca_t structure.
  - Fix "contrib/dblink"'s dblink_exec() to not leak temporary database
    connections upon error.
  - Fix "contrib/dblink" to report the correct connection name in error
  - Fix "contrib/vacuumlo" to use multiple transactions when dropping
    many large objects.
    This change avoids exceeding max_locks_per_transaction when many
    objects need to be dropped. The behavior can be adjusted with the
    new -l (limit) option.

11. By Martin Pitt on 2012-02-27

* New upstream security/bug fix release: (LP: #941912)
  - Require execute permission on the trigger function for "CREATE
    This missing check could allow another user to execute a trigger
    function with forged input data, by installing it on a table he
    owns. This is only of significance for trigger functions marked
    SECURITY DEFINER, since otherwise trigger functions run as the
    table owner anyway. (CVE-2012-0866)
  - Remove arbitrary limitation on length of common name in SSL
    Both libpq and the server truncated the common name extracted from
    an SSL certificate at 32 bytes. Normally this would cause nothing
    worse than an unexpected verification failure, but there are some
    rather-implausible scenarios in which it might allow one
    certificate holder to impersonate another. The victim would have to
    have a common name exactly 32 bytes long, and the attacker would
    have to persuade a trusted CA to issue a certificate in which the
    common name has that string as a prefix. Impersonating a server
    would also require some additional exploit to redirect client
    connections. (CVE-2012-0867)
  - Convert newlines to spaces in names written in pg_dump comments.
    pg_dump was incautious about sanitizing object names that are
    emitted within SQL comments in its output script. A name containing
    a newline would at least render the script syntactically incorrect.
    Maliciously crafted object names could present a SQL injection risk
    when the script is reloaded. (CVE-2012-0868)
  - Fix btree index corruption from insertions concurrent with
    An index page split caused by an insertion could sometimes cause a
    concurrently-running "VACUUM" to miss removing index entries that
    it should remove. After the corresponding table rows are removed,
    the dangling index entries would cause errors (such as "could not
    read block N in file ...") or worse, silently wrong query results
    after unrelated rows are re-inserted at the now-free table
    locations. This bug has been present since release 8.2, but occurs
    so infrequently that it was not diagnosed until now. If you have
    reason to suspect that it has happened in your database, reindexing
    the affected index will fix things.
  - Fix transient zeroing of shared buffers during WAL replay.
    The replay logic would sometimes zero and refill a shared buffer,
    so that the contents were transiently invalid. In hot standby mode
    this can result in a query that's executing in parallel seeing
    garbage data. Various symptoms could result from that, but the most
    common one seems to be "invalid memory alloc request size".
  - Fix handling of data-modifying WITH subplans in READ COMMITTED
    A WITH clause containing "INSERT"/"UPDATE"/"DELETE" would crash if
    the parent "UPDATE" or "DELETE" command needed to be re-evaluated
    at one or more rows due to concurrent updates in READ COMMITTED
  - Fix corner case in SSI transaction cleanup.
    When finishing up a read-write serializable transaction, a crash
    could occur if all remaining active serializable transactions are
  - Fix postmaster to attempt restart after a hot-standby crash.
    A logic error caused the postmaster to terminate, rather than
    attempt to restart the cluster, if any backend process crashed
    while operating in hot standby mode.
  - Fix "CLUSTER"/"VACUUM FULL" handling of toast values owned by
    recently-updated rows.
    This oversight could lead to "duplicate key value violates unique
    constraint" errors being reported against the toast table's index
    during one of these commands.
  - Update per-column permissions, not only per-table permissions, when
    changing table owner.
    Failure to do this meant that any previously granted column
    permissions were still shown as having been granted by the old
    owner. This meant that neither the new owner nor a superuser could
    revoke the now-untraceable-to-table-owner permissions.
  - Support foreign data wrappers and foreign servers in "REASSIGN
    This command failed with "unexpected classid" errors if it needed
    to change the ownership of any such objects.
  - Allow non-existent values for some settings in "ALTER USER/DATABASE
    Allow default_text_search_config, default_tablespace, and
    temp_tablespaces to be set to names that are not known. This is
    because they might be known in another database where the setting
    is intended to be used, or for the tablespace cases because the
    tablespace might not be created yet. The same issue was previously
    recognized for search_path, and these settings now act like that
  - Fix "unsupported node type" error caused by COLLATE in an "INSERT"
  - Avoid crashing when we have problems deleting table files
    Dropping a table should lead to deleting the underlying disk files
    only after the transaction commits. In event of failure then (for
    instance, because of wrong file permissions) the code is supposed
    to just emit a warning message and go on, since it's too late to
    abort the transaction. This logic got broken as of release 8.4,
    causing such situations to result in a PANIC and an unrestartable
  - Recover from errors occurring during WAL replay of "DROP
    Replay will attempt to remove the tablespace's directories, but
    there are various reasons why this might fail (for example,
    incorrect ownership or permissions on those directories). Formerly
    the replay code would panic, rendering the database unrestartable
    without manual intervention. It seems better to log the problem and
    continue, since the only consequence of failure to remove the
    directories is some wasted disk space.
  - Fix race condition in logging AccessExclusiveLocks for hot standby.
    Sometimes a lock would be logged as being held by "transaction
    zero". This is at least known to produce assertion failures on
    slave servers, and might be the cause of more serious problems.
  - Track the OID counter correctly during WAL replay, even when it
    wraps around.
  - Prevent emitting misleading "consistent recovery state reached" log
    message at the beginning of crash recovery.
  - Fix initial value of pg_stat_replication.replay_location.
  - Fix regular expression back-references with - attached.
    Rather than enforcing an exact string match, the code would
    effectively accept any string that satisfies the pattern
    sub-expression referenced by the back-reference symbol.
    A similar problem still afflicts back-references that are embedded
    in a larger quantified expression, rather than being the immediate
    subject of the quantifier. This will be addressed in a future
    PostgreSQL release.
  - Fix recently-introduced memory leak in processing of inet/cidr
  - Fix planner's ability to push down index-expression restrictions
    through UNION ALL.
  - Fix planning of WITH clauses referenced in "UPDATE"/"DELETE" on an
    inherited table.
    This bug led to "could not find plan for CTE" failures.
  - Fix GIN cost estimation to handle column IN (...) index conditions.
    This oversight would usually lead to crashes if such a condition
    could be used with a GIN index.
  - Fix dangling pointer after "CREATE TABLE AS"/"SELECT INTO" in a
    SQL-language function.
    In most cases this only led to an assertion failure in
    assert-enabled builds, but worse consequences seem possible.
  - Fix I/O-conversion-related memory leaks in plpgsql.
  - Work around bug in perl's SvPVutf8() function.
    This function crashes when handed a typeglob or certain read-only
    objects such as $^V. Make plperl avoid passing those to it.
  - In pg_dump, don't dump contents of an extension's configuration
    tables if the extension itself is not being dumped.
  - Improve pg_dump's handling of inherited table columns.
    pg_dump mishandled situations where a child column has a different
    default expression than its parent column. If the default is
    textually identical to the parent's default, but not actually the
    same (for instance, because of schema search path differences) it
    would not be recognized as different, so that after dump and
    restore the child would be allowed to inherit the parent's default.
    Child columns that are NOT NULL where their parent is not could
    also be restored subtly incorrectly.
  - Fix pg_restore's direct-to-database mode for INSERT-style table data.
    Direct-to-database restores from archive files made with
    "--inserts" or "--column-inserts" options fail when using
    pg_restore from a release dated September or December 2011, as a
    result of an oversight in a fix for another problem. The archive
    file itself is not at fault, and text-mode output is okay.
  - Teach pg_upgrade to handle renaming of plpython's shared library.
    Upgrading a pre-9.1 database that included plpython would fail
    because of this oversight.
  - Allow pg_upgrade to process tables containing regclass columns.
    Since pg_upgrade now takes care to preserve pg_class OIDs, there
    was no longer any reason for this restriction.
  - Make libpq ignore ENOTDIR errors when looking for an SSL client
    certificate file.
    This allows SSL connections to be established, though without a
    certificate, even when the user's home directory is set to
    something like /dev/null.
  - Fix some more field alignment issues in ecpg's SQLDA area.
  - Allow AT option in ecpg DEALLOCATE statements.
    The infrastructure to support this has been there for awhile, but
    through an oversight there was still an error check rejecting the
  - Do not use the variable name when defining a varchar structure in
  - Fix "contrib/auto_explain"'s JSON output mode to produce valid JSON.
  - Fix error in "contrib/intarray"'s int[] & int[] operator.
    If the smallest integer the two input arrays have in common is 1,
    and there are smaller values in either array, then 1 would be
    incorrectly omitted from the result.
  - Fix error detection in "contrib/pgcrypto"'s encrypt_iv() and
    These functions failed to report certain types of invalid-input
    errors, and would instead return random garbage values for
    incorrect input.
  - Fix one-byte buffer overrun in "contrib/test_parser".
    The code would try to read one more byte than it should, which
    would crash in corner cases. Since "contrib/test_parser" is only
    example code, this is not a security issue in itself, but bad
    example code is still bad.
  - Use __sync_lock_test_and_set() for spinlocks on ARM, if available.
    This function replaces our previous use of the SWPB instruction,
    which is deprecated and not available on ARMv6 and later. Reports
    suggest that the old code doesn't fail in an obvious way on recent
    ARM boards, but simply doesn't interlock concurrent accesses,
    leading to bizarre failures in multiprocess operation.
  - Use "-fexcess-precision=standard" option when building with gcc
    versions that accept it.
    This prevents assorted scenarios wherein recent versions of gcc
    will produce creative results.
  - Allow use of threaded Python on FreeBSD (Chris Rees)
    Our configure script previously believed that this combination
    wouldn't work; but FreeBSD fixed the problem, so remove that error
* Drop 00git_inet_cidr_unpack.patch, 01-armel-tas.patch: Applied upstream.

10. By Martin Pitt on 2011-09-26

* New upstream bug fix release:
  - Make pg_options_to_table return NULL for an option with no value.
    Previously such cases would result in a server crash.
  - Fix memory leak at end of a GiST index scan.
    Commands that perform many separate GiST index scans, such as
    verification of a new GiST-based exclusion constraint on a table
    already containing many rows, could transiently require large
    amounts of memory due to this leak.
  - Fix explicit reference to pg_temp schema in "CREATE TEMPORARY
    TABLE". This used to be allowed, but failed in 9.1.0.

9. By Martin Pitt on 2011-09-12

* Final 9.1 release.
* 02-relax-sslkey-permscheck.patch, 50-per-version-dirs.patch: Refresh to
  apply cleanly.
* debian/control: Tighten the dependencies of the -pl* extensions/-contrib
  to postgresql-9.1 to the same binary version. (Closes: #640335)

8. By Martin Pitt on 2011-08-27

* debian/watch: Fix for mangling ~rc, thanks Peter Eisentraut.
  (Closes: #639357)
* debian/control: Add versionless Provides: to the PL* extensions, as per
  request from Christoph Berg.
* debian/control: Add "Replaces: postgresql-9.0-dbg" to fix file conflict.
  (Closes: #639258)
* debian/control: Drop the versionless metapackages, they are built from
  postgresql-common now. This behaves better with backports. Thanks to
  Christoph Berg for the suggestion.

7. By Martin Pitt on 2011-08-25

debian/control: Build the versionless metapackages again, and point them
to 9.1.

6. By Martin Pitt on 2011-07-14

* New upstream beta release.
  - Works around gcc 4.6.0 bug. (Closes: #633086)

  Note that this does not change the data format since Beta 2, so no need
  to dump/reload clusters.

Branch metadata

Branch format:
Branch format 7
Repository format:
Bazaar repository format 2a (needs bzr 1.16 or later)
Stacked on:
This branch contains Public information 
Everyone can see this information.