PostgreSQL Weekly News – March 13 2011
== PostgreSQL Weekly News – March 13 2011 ==
PostgreSQL 9.1alpha4 released, with lots of big, new goodies!
This is the last planned alpha for 9.1. You can get it here:
New survey: which 9.1 feature are you most excited about?
== PostgreSQL Jobs for March ==
== PostgreSQL Local ==
PostgreSQL Conference East 2011: NYC March 22nd – 25th
There will be a large PostgreSQL presence at the MySQL Conference and
Expo, April 11-14, 2011 in Santa Clara, California.
Open Database Camp will be on May 7-9, 2011 in Sardinia, Italy
PGCon will be May 19-20, 2011 at the University of Ottawa, preceded by
two days of tutorials on May 17-18.
== PostgreSQL in the News ==
Planet PostgreSQL: http://planet.postgresql.org/
PostgreSQL Weekly News is brought to you this week by David Fetter
Submit news and announcements by Sunday at 3:00pm Pacific time.
Please send English language ones to firstname.lastname@example.org, German language
to email@example.com, Italian language to firstname.lastname@example.org. Spanish language
== Reviews ==
== Applied Patches ==
Heikki Linnakangas pushed:
– Silence compiler warning about undefined function when compiling
– Begin error message with lower-case letter.
– Document the DEFERRABLE option in SET TRANSACTION command. Kevin
– If recovery_target_timeline is set to ‘latest’ and standby mode is
enabled, periodically rescan the archive for new timelines, while
waiting for new WAL segments to arrive. This allows you to set up a
standby server that follows the TLI change if another standby server
is promoted to master. Before this, you had to restart the standby
server to make it notice the new timeline. This patch only scans
the archive for TLI changes, it won’t follow a TLI change in
streaming replication. That is much needed too, but it would be a
much bigger patch than I dare to sneak in this late in the release
cycle. There was discussion on improving the sanity checking of the
WAL segments so that the system would notice more reliably if the
new timeline isn’t an ancestor of the current one, but that is not
included in this patch. Reviewed by Fujii Masao.
– Truncate predicate lock manager’s SLRU lazily at checkpoint. That’s
safer than doing it aggressively whenever the tail-XID pointer is
advanced, because this way we don’t need to do it while holding
SerializableXactHashLock. This also fixes bug #5915 spotted by
YAMAMOTO Takashi, and removes an obsolete comment spotted by Kevin
– Don’t throw a warning if vacuum sees PD_ALL_VISIBLE flag set on a
page that contains newly-inserted tuples that according to our
OldestXmin are not yet visible to everyone. The value returned by
GetOldestXmin() is conservative, and it can move backwards on
repeated calls, so if we see that contradiction between the
PD_ALL_VISIBLE flag and status of tuples on the page, we have to
assume it’s because an earlier vacuum calculated a higher OldestXmin
value, and all the tuples really are visible to everyone. We have
received several reports of this bug, with the “PD_ALL_VISIBLE flag
was incorrectly set in relation …” warning appearing in logs. We
were finally able to hunt it down with David Gould’s help to run
extra diagnostics in an environment where this happened frequently.
Also reword the warning, per Robert Haas’ suggestion, to not imply
that the PD_ALL_VISIBLE flag is necessarily at fault, as it might
also be a symptom of corruption on a tuple header. Backpatch to
8.4, where the PD_ALL_VISIBLE flag was introduced.
– Fix overly strict assertion in SummarizeOldestCommittedSxact().
There’s a race condition where SummarizeOldestCommittedSxact() is
called even though another backend already cleared out all finished
sxact entries. That’s OK, RegisterSerializableTransactionInt() can
just retry getting a news xact slot from the available-list when
that happens. Reported by YAMAMOTO Takashi, bug #5918.
– Fix bugs in the isolation tester flex rules. Tom Lane pointed out
that it was giving a warning: “-s option given but default rule can
be matched”. That was because there was no rule to handle newline in
a quoted string. I made that throw an error. Also, line number
tracking was broken, giving incorrect line number on error. Fixed
– In ecpg preprocessor, don’t try to look up constants in the test for
variable hiding. A constant is not a variable. It worked in most
cases by accident, because we add constants to the global list of
variables (why?), but float constants like 1.23 were interpreted as
struct field references, and not found. Backpatch to 9.0, where the
test for variable hiding was added.
Tom Lane pushed:
– Zero out vacuum_count and related counters in pgstat_recv_tabstat().
This fixes an oversight in commit 946045f04d11d246a834b917a2b8bc6e4f884a37
of 2010-08-21, as reported by Itagaki Takahiro. Also a couple of
minor cosmetic adjustments.
– Minor copy-editing in CREATE TRIGGER reference page. Per
suggestions from Thom Brown and Robert Haas.
– Improve description of inquiry functions that accept regclass. Per
a suggestion from Thom Brown, though this is not his proposed patch.
– Add missing keywords to gram.y’s unreserved_keywords list. We
really need an automated check for this … and did VALIDATE really
need to become a keyword at all, rather than picking some other
syntax using existing keywords?
– Assorted editing for collation documentation. I made a pass over
this to familiarize myself with the feature, and found some things
that could be improved.
– A bit more editing for collation documentation.
– Adjust the permissions required for COMMENT ON ROLE. Formerly, any
member of a role could change the role’s comment, as of course could
superusers; but holders of CREATEROLE privilege could not, unless
they were also members. This led to the odd situation that a
CREATEROLE holder could create a role but then could not comment on
it. It also seems a bit dubious to let an unprivileged user change
his own comment, let alone those of group roles he belongs to. So,
change the rule to be “you must be superuser to comment on a
superuser role, or hold CREATEROLE to comment on non-superuser
roles”. This is the same as the privilege check for
creating/dropping roles, and thus fits much better with the rule for
other object types, namely that only the owner of an object can
comment on it. In passing, clean up the documentation for COMMENT a
little bit. Per complaint from Owen Jacobson and subsequent
– Remove collation information from TypeName, where it does not
belong. The initial collations patch treated a COLLATE spec as part
of a TypeName, following what can only be described as brain fade on
the part of the SQL committee. It’s a lot more reasonable to treat
COLLATE as a syntactically separate object, so that it can be added
in only the productions where it actually belongs, rather than
needing to reject it in a boatload of places where it doesn’t belong
(something the original patch mostly failed to do). In addition
this change lets us meet the spec’s requirement to allow COLLATE
anywhere in the clauses of a ColumnDef, and it avoids unfriendly
behavior for constructs such as “foo::type COLLATE collation”. To
do this, pull collation information out of TypeName and put it in
ColumnDef instead, thus reverting most of the collation-related
changes in parse_type.c’s API. I made one additional structural
change, which was to use a ColumnDef as an intermediate node in
AT_AlterColumnType AlterTableCmd nodes. This provides enough room
to get rid of the “transform” wart in AlterTableCmd too, since the
ColumnDef can carry the USING expression easily enough. Also fix
some other minor bugs that have crept in in the same areas, like
failure to copy recently-added fields of ColumnDef in copyfuncs.c.
While at it, document the formerly secret ability to specify a
collation in ALTER TABLE ALTER COLUMN TYPE, ALTER TYPE ADD
ATTRIBUTE, and ALTER TYPE ALTER ATTRIBUTE TYPE; and correct some
misstatements about what the default collation selection will be
when COLLATE is omitted. BTW, the three-parameter form of
format_type() should go away too, since it just contributes to the
confusion in this area; but I’ll do that in a separate patch.
– Fix some oversights in distprep and maintainer-clean targets. At
least two recent commits have apparently imagined that a comment in
a Makefile stating that something would be included in the
distribution tarball was sufficient to make it so. They hadn’t
bothered to hook into the upper maintainer-clean targets either.
Per bug #5923 from Charles Johnson, in which it emerged that the
9.1alpha4 tarballs are short a few files that should be there.
– replication/repl_gram.h needs to be cleaned too …
– Revert addition of third argument to format_type(). Including
collation in the behavior of that function promotes a world view we
do not want. Moreover, it was producing the wrong behavior for
pg_dump anyway: what we want is to dump a COLLATE clause on
attributes whose attcollation is different from the underlying type,
and likewise for domains, and the function cannot do that for us.
Doing it the hard way in pg_dump is a bit more tedious but produces
more correct output. In passing, fix initdb so that the initial
entry in pg_collation is properly pinned. It was droppable before
– Remove duplicate indexterm to silence openjade wrning.
– Create an explicit concept of collations that work for any encoding.
Use collencoding = -1 to represent such a collation in pg_collation.
We need this to make the “default” entry work sanely, and a later
patch will fix the C/POSIX entries to be represented this way
instead of duplicating them across all encodings. All lookup
operations now search first for an entry that’s
database-encoding-specific, and then for the same name with
collencoding = -1. Also some incidental code cleanup in
collationcmds.c and pg_collation.c.
– Split CollateClause into separate raw and analyzed node types.
CollateClause is now used only in raw grammar output, and
CollateExpr after parse analysis. This is for clarity and to avoid
carrying collation names in post-analysis parse trees: that’s both
wasteful and possibly misleading, since the collation’s name could
be changed while the parsetree still exists. Also, clean up
assorted infelicities and omissions in processing of the node type.
– Put in some more safeguards against executing a division-by-zero.
Add dummy returns before every potential division-by-zero in int8.c,
because apparently further “improvements” in gcc’s optimizer have
enabled it to break functions that weren’t broken before. Aurelien
Jarno, via Martin Pitt
– On further reflection, we’d better do the same in int.c. We
previously heard of the same problem in int24div(), so there’s not a
good reason to suppose the problem is confined to cases involving
– Make all comparisons done for/with statistics use the default
collation. While this will give wrong answers when estimating
selectivity for a comparison operator that’s using a non-default
collation, the estimation error probably won’t be large; and anyway
the former approach created estimation errors of its own by trying
to use a histogram that might have been computed with some other
collation. So we’ll adopt this simplified approach for now and
perhaps improve it sometime in the future. This patch incorporates
changes from Andres Freund to make sure that selfuncs.c passes a
valid collation OID to any datatype-specific function it calls, in
case that function wants collation information. Said OID will now
always be DEFAULT_COLLATION_OID, but at least we won’t get errors.
– Simplify list traversal logic in add_path(). Its mechanism for
recovering after deleting the current list cell was a bit klugy.
Borrow the technique used in other places.
Robert Haas pushed:
– Synchronous replication doc corrections. Thom Brown
– Reword alpha release note item on SSI. Per Josh Berkus; some
additional explanatory text by me.
– Create “replication and recovery” section in alpha release notes.
– Update alpha release notes for latest commits.
– Make alpha release notes more consistent as regards periods.
– Replication README updates. Fujii Masao
– Add missing index terms for recovery control functions. Fujii Masao
– Emit a LOG message when pausing at the recovery target. Fujii Masao
– Minor sync rep corrections. Fujii Masao, with a bit of additional
wordsmithing by me.
– Remove obsolete comment. In earlier versions of the sync rep patch,
waiters removed themselves from the queue, but now walsender removes
them before doing the wakeup. Report by Fujii Masao.
– More synchronous replication tweaks. SyncRepRequested() must check
not only the value of the synchronous_replication GUC but also
whether max_wal_senders > 0. Otherwise, we might end up waiting for
sync rep even when there’s no possibility of a standby ever managing
to connect. There are some existing cross-checks to prevent this,
but they’re not quite sufficient: the user can start the server with
max_wal_senders=0, synchronous_standby_names=”, and
synchronous_replication=off and then subsequent make
synchronous_standby_names not empty using pg_ctl reload, and then
SET synchronous_standby=on, leading to an indefinite hang. Along
the way, rename the global variable for the synchronous_replication
GUC to match the name of the GUC itself, for clarity. Report by
Fujii Masao, though I didn’t use his patch.
– More synchronous replication typo fixes. Fujii Masao
– Make error handling of synchronous_standby_names consistent. It’s
not a good idea to kill the postmaster just because someone muffs
this, and it’s not consistent with what we do for other, similar
GUCs. Fujii Masao, with a bit more hacking by me
Peter Eisentraut pushed:
– Fix behavior when raising plpy.Fatal(). It should cause a
elog(FATAL) error, and it fact it was simply causing a elog(ERROR).
– Report Python errors from iterators with PLy_elog. This improves
reporting, as the error string now includes the actual Python
exception. As a side effect, this no longer sets the errcode to
ERRCODE_DATA_EXCEPTION, which might be considered a feature, as it’s
not documented and not clear why iterator errors should be treated
differently. Jan Urbański.
– Ignore files built by coverage builds.
– Fix parallel make when running make install before make all. In
addition to the all-foo-recurse: all-bar-recurse dependencies that
constraint the order of the rule execution, we need
install-foo-recurse: install-bar-recurse dependencies in case one
runs make install without a make all first, as some people
– Add test case for collation mismatch in recursive query. This isn’t
very important by itself, but was left on my list of things without
test coverage for the collation feature.
Michael Meskes pushed:
– Added new version of ecpg’s parser test script which was written by
Bruce Momjian pushed:
– Document that char() ignores spaces in non-pattern comparisons, not
in pattern comparisons such as LIKE and regex.
– Adjust CHAR() doc mention of pattern matching issues for trailing
– Fix file descriptor leaks in pg_upgrade in failure code paths.
– Improve wording of initdb and pg_controldata manual pages.
– Remove ‘=’ from initdb switch syntax.
– Document that server single-user mode does not do checkpoints or
other background processing.
– Mention gcc version in C comment.
– Use ‘=’ when documenting long options.
– Add proper git-external-diff script to src/tools.
– Clarify what commands are not allowed in SQL functions.
– Improve SQL function disallowed command wording.
– More wording improvements for disallowed SQL function commands.
– Reference doc “examples” section for pg_dump options that often need
complex quoting, e.g. -t and -n.
– Clarify C comment that O_SYNC/O_FSYNC are really the same settting,
as opposed to O_DSYNC.
– Modify pg_test_fsync to match the behavior of git head in regards to
– When a smart pg_ctl shutdown fails, mention -m fast as a tip.
– Add C comment that new new pg_dump -X options are to be created.
– Remove pg_dump -X options that are only in 9.1 and not needed for
– Change “TIP” to “HINT” to match backend message style, in pg_ctl -m
– Add comment to tools/git-external-diff.
– Document that the parenthesized VACUUM syntax is deprecated, not the
– Document that libpq’s PQgetResult() should be called after a fatal
error to fully process errors.
– Update C comment about O_DIRECT and fsync().
– Document that pg_dump –clean might generate some harmless errors on
– Update kernel docs for Solaris 10. Josh Berkus
– Fix “unparenthesized” mention in vacuum docs.
– Improve extract(day) documentation with interval values.
– In plpsql docs, use RAISE rather than undefined log function.
– Adds index entries for session_user and pg_describe_object. Removes
extraneous closing parenthesis from pg_describe_object. Puts
pg_describe_object and has_sequence_privilege in correct
alphabetical position in function listing. Thom Brown
– Document how listen_addresses can do only IPv4 or IPv6.
– Document that to_char(‘FM’) removes only trailing zeros.
– Update documentation on FreeBSD write cache control.
– In docs, rename “backwards compatibility” to “backward
compatibility” for consistency.
– Use macros for time-based constants, rather than constants.
– Use “backend process” rather than “backend server”, where
ITAGAKI Takahiro pushed:
– synchronous_standby_names is a string parameter.
– Remove ‘s’ from recovery_target_timeline’s’ from the release note.
– Cleanup copyright years and file names in the header comments of
== Rejected Patches (for now) ==
No one was disappointed this week
== Pending Patches ==
Jan Urbanski sent in another flock of patches to fix PL/PythonU.
Fujii Masao sent in another revision of the replication server timeout
Thom Brown sent in two revisions of doc patches for syncrep.
Fujii Masao sent in four revisions of a patch to add a sample
recovery.conf which illustrates the use of recovery control functions.
Gurjeet Singh sent in another revision of the patch to allow psql to
use relative paths for includes, this time with a long option
Noah Misch sent in a patch to add test for FOR KEY LOCK.
Mark Kirkwood sent in another revision of the patch to constrain
backend temporary file space.
Kevin Grittner sent in a patch to reformat the SSI files per pgindent.
Bruce Momjian and Christopher Browne traded patches to make constants
for magic time values.
Robert Haas sent in a patch to enable ALTER TABLE … ALTER CONSTRAINT
… VALID, which makes the system check whether the constraint is true
if not already enforced.
Noah Misch sent in a patch to fix an issue with on-the-fly index tuple
deletion and hot standby.
Bruce Momjian sent in a patch to add a comment to the template0
Robert Haas sent in a patch to add the new keywords 9.1 features have
Noah Misch sent in a patch to fix some memory-related bugs he turned
up while testing an instrumented version of PostgreSQL.
Comments are closed.