When we need to insert a new entry and the queue is full, compact the
entire queue in the hopes of making room for the new entry.  Doing this
on every insertion might worsen contention on BgWriterCommLock, but
when the queue it's full, it's far better than allowing the backend to
perform its own fsync, per testing by Greg Smith as reported in
http://archives.postgresql.org/pgsql-hackers/2011-01/msg02665.php

Original idea from Greg Smith.  Patch by me.  Review by Chris Browne
and Greg Smith

Although this improves the style, an ulterior motive is to keep the two
table links from breaking across lines in PDF output, per complaint from
Josh Kupershmidt.

The link to the CREATE CONVERSION manual page was split across a page
boundary in the PDF output, leading to "\pdfendlink ended up in different
nesting level than \pdfstartlink" error while building PDFs.

It wouldn't be worth changing text that's undergoing active editing to
avoid this, since other editing might result in moving the link away from
the page end anyway.  But this paragraph has been static for a long time,
so might as well fix it to prevent it from being an issue in future.

Security: CVE-2010-4015

contrib/intarray's gettoken() uses a fixed-size buffer to collect an
integer's digits, and did not guard against overrunning the buffer.
This is at least a backend crash risk, and in principle might allow
arbitrary code execution.  The code didn't check for overflow of the
integer value either, which while not presenting a crash risk was still
bad.

Thanks to Apple Inc's security team for reporting this issue and supplying
the fix.

Security: CVE-2010-4015

We only need that header when compiling with icc, since the gcc variant of
ia64_get_bsp() uses in-line assembly code.  Per report from Frank Brendel,
the header doesn't exist on all IA64 platforms; so don't include it unless
we need it.

transaction loss for a _database_ crash.

This reverts commit a06e41de of 2011-01-26.
Per discussion, this behavior is not wanted, as it would need to change if
we ever made composite types support DEFAULT.

paragraphs, per suggestion from Dan Birken.

In the case where the initial call of systable_getnext_ordered() returned
NULL, this function would nonetheless call it again.  That's undefined
behavior that only by chance failed to not give visibly incorrect results.
Put an if-test around the final loop to prevent that, and in passing
improve some comments.  No back-patch since there's no actual failure.

Per report from YAMAMOTO Takashi.

Synchronize pg_config.h.in with configure.in (someone must have
forgotten to run autoheader or autoreconf), and clean up some spurious
change in configure introduced by the last commit there.

It is redundant, given the error context.

Jan Urbański

Use the built-in TypeError, not SPIError, for errors having to do with
argument counts or types.  Use SPIError, not simply plpy.Error, for
errors in PLy_spi_execute_plan.  Finally, do not set a Python
exception if PyArg_ParseTuple failed, as it already sets the correct
exception.

Jan Urbański

The temporarily broken plpython_unicode test shows a case where this
is used.

Do remaining fix-ups on the expected files at the same time.

Older versions of GCC appear to report these with the current standard
option set, newer versions need -Wformat-security.

sections, rather than just calling it "/contrib/module_name".

Also update pg_test_fsync build instructions now that it is in /contrib.

non-durable items, per Robert Haas.

Noah Misch, slightly revised.

The previous coding prevented ALTER TABLE .. ADD COLUMN from being used
with a non-NULL default in situations where the table's rowtype was being
used elsewhere.  But this is a completely arbitrary restriction since
you could do the same operation in multiple steps (add the column, add
the default, update the table).

Inspired by a patch from Noah Misch, though I didn't use his code.

always 8k writes, per suggestion from Tom.  Also adjust open_sync output
layout.

Robvert Haas.

The latter is the correct name of the operation to change the data type
of a column.

Noah Misch

We used to do that on pg_listener, but pg_listener is no more.

Also add a bit more documentation for ShareRowExclusive mode.

Noah Misch

There isn't any need to track this state on a table-wide basis, and trying
to do so introduces undesirable semantic fuzziness.  Move the flag to
pg_index, where it clearly describes just a single index and can be
immutable after index creation.

This feature allows a unique or pkey constraint to be created using an
already-existing unique index.  While the constraint isn't very
functionally different from the bare index, it's nice to be able to do that
for documentation purposes.  The main advantage over just issuing a plain
ALTER TABLE ADD UNIQUE/PRIMARY KEY is that the index can be created with
CREATE INDEX CONCURRENTLY, so that there is not a long interval where the
table is locked against updates.

On the way, refactor some of the code in DefineIndex() and index_create()
so that we don't have to pass through those functions in order to create
the index constraint's catalog entries.  Also, in parse_utilcmd.c, pass
around the ParseState pointer in struct CreateStmtContext to save on
notation, and add error location pointers to some error reports that didn't
have one before.

Gurjeet Singh, reviewed by Steve Singer and Tom Lane

Fujii Masao

it is 8k as expected.

This option was recently introduced, but the documentation in help.c
was not updated.

This way errors from fetching tuples are correctly reported as errors
in the SPI call.  While at it, avoid palloc(0).

Jan Urbański

Instead of checking whether the arglist is NULL and then if its length
is 0, do it in one step, and outside of the try/catch block.

Jan Urbański