The folly of the previous arrangement was just demonstrated: there's no
convenient way to add fields to ExplainState without breaking ABI, even
if callers have no need to touch those fields.  Since we might well need
to do that again someday in back branches, let's change things so that
only explain.c has to have sizeof(ExplainState) compiled into it.  This
costs one extra palloc() per EXPLAIN operation, which is surely pretty
negligible.

Previously, read() might have returned a length equal to the buffer
length, and then the subsequent store to buf[len] would write a
zero-byte one byte past the end.  This doesn't seem likely to be
a security issue, but there's some chance it could result in
pg_standby misbehaving.

Spotted by Coverity; patch by Michael Paquier, reviewed by me.

Spotted by Coverity.  This isn't likely to matter in practice, but
there's no harm in fixing it.

Michael Paquier

Similar warnings from backend were silenced earlier by commit c8315930,
but there were a few more contrib/hstore.

Michael Paquier

Report by Tatsuo Ishii, Coverity

Backpatch certain files through 9.0

Previously the computation of the total test duration, measured in
microseconds, accidentally overflowed due to accidentally using signed
32bit arithmetic.  As the only consequence is that pg_test_timing
invocations with such, overly large, durations never finished the
practical consequences of this bug are minor.

Pointed out by Coverity.

Backpatch to 9.2 where pg_test_timing was added.

In the unlikely case of stdin (fd 0) being closed, the off-by-one
would lead to pg_xlogdump failing to open files.

Spotted by Coverity.

Backpatch to 9.3 where pg_xlogdump was introduced.

Pointed out by Coverity.

Backpatch to all supported branches, the code has been that way for a
long while.

This makes the functions much nicer to read and edit, and also makes
debugging easier.

Previously, if you wanted anything besides C-string hash keys, you had to
specify a custom hashing function to hash_create().  Nearly all such
callers were specifying tag_hash or oid_hash; which is tedious, and rather
error-prone, since a caller could easily miss the opportunity to optimize
by using hash_uint32 when appropriate.  Replace this with a design whereby
callers using simple binary-data keys just specify HASH_BLOBS and don't
need to mess with specific support functions.  hash_create() itself will
take care of optimizing when the key size is four bytes.

This nets out saving a few hundred bytes of code space, and offers
a measurable performance improvement in tidbitmap.c (which was not
exploiting the opportunity to use hash_uint32 for its 4-byte keys).
There might be some wins elsewhere too, I didn't analyze closely.

In future we could look into offering a similar optimized hashing function
for 8-byte keys.  Under this design that could be done in a centralized
and machine-independent fashion, whereas getting it right for keys of
platform-dependent sizes would've been notationally painful before.

For the moment, the old way still works fine, so as not to break source
code compatibility for loadable modules.  Eventually we might want to
remove tag_hash and friends from the exported API altogether, since there's
no real need for them to be explicitly referenced from outside dynahash.c.

Teodor Sigaev and Tom Lane

Use SSPI authentication to allow connections exclusively from the OS
user that launched the test suite.  This closes on Windows the
vulnerability that commit be76a6d3
closed on other platforms.  Users of "make installcheck" or custom test
harnesses can run "pg_regress --config-auth=DATADIR" to activate the
same authentication configuration that "make check" would use.
Back-patch to 9.0 (all supported versions).

Security: CVE-2014-0067

As with NOT NULL constraints, we consider that such constraints are merely
reports of constraints that are being enforced by the remote server (or
other underlying storage mechanism).  Their only real use is to allow
planner optimizations, for example in constraint-exclusion checks.  Thus,
the code changes here amount to little more than removal of the error that
was formerly thrown for applying CHECK to a foreign table.

(In passing, do a bit of cleanup of the ALTER FOREIGN TABLE reference page,
which had accumulated some weird decisions about ordering etc.)

Shigeru Hanada and Etsuro Fujita, reviewed by Kyotaro Horiguchi and
Ashutosh Bapat.

Add Windows versions of generated scripts, and make sure we only
ignore the scripts int he root directory.

Michael Paquier

Code added in 9.4 would attempt to divide by zero in such cases.
Noted while testing fix for missing-pclose problem.

If the called command fails to return data, runShellCommand forgot to
pclose() the pipe before returning.  This is fairly harmless in the current
code, because pgbench would then abandon further processing of that client
thread; so no more than nclients descriptors could be leaked this way.  But
it's not hard to imagine future improvements whereby that wouldn't be true.
In any case, it's sloppy coding, so patch all branches.  Found by Coverity.

In commit 462bd957, I changed postgres_fdw
to rely on get_plan_rowmark() instead of get_parse_rowmark().  I still
think that's a good idea in the long run, but as Etsuro Fujita pointed out,
it doesn't work today because planner.c forces PlanRowMarks to have
markType = ROW_MARK_COPY for all foreign tables.  There's no urgent reason
to change this in the back branches, so let's just revert that part of
yesterday's commit rather than trying to design a better solution under
time pressure.

Also, add a regression test case showing what postgres_fdw does with FOR
UPDATE/SHARE.  I'd blithely assumed there was one already, else I'd have
realized yesterday that this code didn't work.

Ordinarily we can omit checking of a WHERE condition that matches a partial
index's condition, when we are using an indexscan on that partial index.
However, in SELECT FOR UPDATE we must include the "redundant" filter
condition in the plan so that it gets checked properly in an EvalPlanQual
recheck.  The planner got this mostly right, but improperly omitted the
filter condition if the index in question was on an inheritance child
table.  In READ COMMITTED mode, this could result in incorrectly returning
just-updated rows that no longer satisfy the filter condition.

The cause of the error is using get_parse_rowmark() when get_plan_rowmark()
is what should be used during planning.  In 9.3 and up, also fix the same
mistake in contrib/postgres_fdw.  It's currently harmless there (for lack
of inheritance support) but wrong is wrong, and the incorrect code might
get copied to someplace where it's more significant.

Report and fix by Kyotaro Horiguchi.  Back-patch to all supported branches.

Author: Michael Paquier

The old formula was completely bogus with the new WAL record format.

dblink and postgres_fdw use SHLIB_PREREQS = submake-libpq to build libpq
first.  This doesn't work in a PGXS build, because there is no libpq to
build.  So just omit setting SHLIB_PREREQS in this case.

Note that PGXS users can still use SHLIB_PREREQS (although it is not
documented).  The problem here is only that contrib modules can be built
in-tree or using PGXS, and the prerequisite is only applicable in the
former case.

Commit 6697aa2b previously attempted to
address this by creating a somewhat fake submake-libpq target in
Makefile.global.  That was not the right fix, and it was also done in a
nonportable way, so revert that.

Transactions can now set their commit timestamp directly as they commit,
or an external transaction commit timestamp can be fed from an outside
system using the new function TransactionTreeSetCommitTsData().  This
data is crash-safe, and truncated at Xid freeze point, same as pg_clog.

This module is disabled by default because it causes a performance hit,
but can be enabled in postgresql.conf requiring only a server restart.

A new test in src/test/modules is included.

Catalog version bumped due to the new subdirectory within PGDATA and a
couple of new SQL functions.

Authors: Álvaro Herrera and Petr Jelínek

Reviewed to varying degrees by Michael Paquier, Andres Freund, Robert
Haas, Amit Kapila, Fujii Masao, Jaime Casanova, Simon Riggs, Steven
Singer, Peter Eisentraut

The logical decoding patchset introduced PROC_IN_LOGICAL_DECODING flag
PGXACT flag, that allows such backends to be skipped when computing
the xmin horizon/snapshots. That's fine and sensible for walsenders
streaming out logical changes, but not at all fine for SQL backends
doing logical decoding. If the latter set that flag any change they
have performed outside of logical decoding will not be regarded as
visible - which e.g. can lead to that change being vacuumed away.

Note that not setting the flag for SQL backends isn't particularly
bothersome - the SQL backend doesn't do streaming, so it only runs for
a limited amount of time.

Per buildfarm member 'tick' and Alvaro.

Backpatch to 9.4, where logical decoding was introduced.

Michael Paquier

Double-dash additions suggested by Peter Geoghegan

We expose a function IsValidJsonNumber that internally calls the lexer
for json numbers. That allows us to use the same test everywhere,
instead of inventing a broken test for hstore conversions. The new
function is also used in datum_to_json, replacing the code that is now
moved to the new function.

Backpatch to 9.3 where hstore_to_json_loose was introduced.

This is advance preparation for introducing even more test modules; the
easy solution is to add them to contrib, but that's bloated enough that
it seems a good time to think of something different.

Moved modules are dummy_seclabel, test_shm_mq, test_parser and
worker_spi.

(test_decoding was also a candidate, but there was too much opposition
to moving that one.  We can always reconsider later.)

This patch adds a way of iterating through the members of a bitmapset
nondestructively, unlike the old way with bms_first_member().  While
bms_next_member() is very slightly slower than bms_first_member()
(at least for typical-size bitmapsets), eliminating the need to palloc
and pfree a temporary copy of the target bitmapset is a significant win.
So this method should be preferred in all cases where a temporary copy
would be necessary.

Tom Lane, with suggestions from Dean Rasheed and David Rowley

Mark Simonetti reported that libxslt sometimes crashes for him, and that
swapping xslt_process's object-freeing calls around to do them in reverse
order of creation seemed to fix it.  I've not reproduced the crash, but
valgrind clearly shows a reference to already-freed memory, which is
consistent with the idea that shutdown of the xsltTransformContext is
trying to reference the already-freed stylesheet or input document.
With this patch, valgrind is no longer unhappy.

I have an inquiry in to see if this is a libxslt bug or if we're just
abusing the library; but even if it's a library bug, we'd want to adjust
our code so it doesn't fail with unpatched libraries.

Back-patch to all supported branches, because we've been doing this in
the wrong(?) order for a long time.

Code that check the flag no longer need #ifdef's, which is more convenient.
In particular, makes it easier to write extensions that depend on it.

In the passing, modify sslinfo's ssl_is_used function to check ssl_in_use
instead of the OpenSSL specific 'ssl' pointer. It doesn't make any
difference currently, as sslinfo is only compiled when built with OpenSSL,
but seems cleaner anyway.

This is further infrastructure for parallelism.

Amit Khandekar, Noah Misch, Robert Haas

postgres_fdw would send query conditions involving system columns to the
remote server, even though it makes no effort to ensure that system
columns other than CTID match what the remote side thinks.  tableoid,
in particular, probably won't match and might have some use in queries.
Hence, prevent sending conditions that include non-CTID system columns.

Also, create_foreignscan_plan neglected to check local restriction
conditions while determining whether to set fsSystemCol for a foreign
scan plan node.  This again would bollix the results for queries that
test a foreign table's tableoid.

Back-patch the first fix to 9.3 where postgres_fdw was introduced.
Back-patch the second to 9.2.  The code is probably broken in 9.1 as
well, but the patch doesn't apply cleanly there; given the weak state
of support for FDWs in 9.1, it doesn't seem worth fixing.

Etsuro Fujita, reviewed by Ashutosh Bapat, and somewhat modified by me

Patch by me, Peter Geoghegan and Michael Paquier, reviewed by Amit Kapila.

Each WAL record now carries information about the modified relation and
block(s) in a standardized format. That makes it easier to write tools that
need that information, like pg_rewind, prefetching the blocks to speed up
recovery, etc.

There's a whole new API for building WAL records, replacing the XLogRecData
chains used previously. The new API consists of XLogRegister* functions,
which are called for each buffer and chunk of data that is added to the
record. The new API also gives more control over when a full-page image is
written, by passing flags to the XLogRegisterBuffer function.

This also simplifies the XLogReadBufferForRedo() calls. The function can dig
the relation and block number from the WAL record, so they no longer need to
be passed as arguments.

For the convenience of redo routines, XLogReader now disects each WAL record
after reading it, copying the main data part and the per-block data into
MAXALIGNed buffers. The data chunks are not aligned within the WAL record,
but the redo routines can assume that the pointers returned by XLogRecGet*
functions are. Redo routines are now passed the XLogReaderState, which
contains the record in the already-disected format, instead of the plain
XLogRecord.

The new record format also makes the fixed size XLogRecord header smaller,
by removing the xl_len field. The length of the "main data" portion is now
stored at the end of the WAL record, and there's a separate header after
XLogRecord for it. The alignment padding at the end of XLogRecord is also
removed. This compansates for the fact that the new format would otherwise
be more bulky than the old format.

Reviewed by Andres Freund, Amit Kapila, Michael Paquier, Alvaro Herrera,
Fujii Masao.

This can cause problems on Windows, where files that are still open
can't be unlinked.

Jeff Janes

There are basically three situations in which logical decoding needs
to perform cache invalidation. During/After replaying a transaction
with catalog changes, when skipping a uninteresting transaction that
performed catalog changes and when erroring out while replaying a
transaction. Unfortunately these three cases were all done slightly
differently - partially because 8de3e410, which greatly simplifies
matters, got committed in the midst of the development of logical
decoding.

The actually problematic case was when logical decoding skipped
transaction commits (and thus processed invalidations). When used via
the SQL interface cache invalidation could access the catalog - bad,
because we didn't set up enough state to allow that correctly. It'd
not be hard to setup sufficient state, but the simpler solution is to
always perform cache invalidation outside a valid transaction.

Also make the different cache invalidation cases look as similar as
possible, to ease code review.

This fixes the assertion failure reported by Antonin Houska in
53EE02D9.7040702@gmail.com. The presented testcase has been expanded
into a regression test.

Backpatch to 9.4, where logical decoding was introduced.

The hope is that we can use this to produce better diagnostics in
some cases.

Peter Geoghegan, reviewed by Michael Paquier, with some further
changes by me.

Heikki noticed in 544E23C0.8090605@vmware.com that slot.c and
snapbuild.c were missing the FIN_CRC32 call when computing/checking
checksums of on disk files. That doesn't lower the the error detection
capabilities of the checksum, but is inconsistent with other usages.

In a followup mail Heikki also noticed that, contrary to a comment,
the 'version' and 'length' struct fields of replication slot's on disk
data where not covered by the checksum. That's not likely to lead to
actually missed corruption as those fields are cross checked with the
expected version and the actual file length. But it's wrong
nonetheless.

As fixing these issues makes existing on disk files unreadable, bump
the expected versions of on disk files for both slots and logical
decoding historic catalog snapshots.  This means that loading old
files will fail with
ERROR: "replication slot file ... has unsupported version 1"
and
ERROR: "snapbuild state file ... has unsupported version 1 instead of
2" respectively. Given the low likelihood of anybody already using
these new features in a production setup that seems acceptable.

Fixing these issues made me notice that there's no regression test
covering the loading of historic snapshot from disk - so add one.

Backpatch to 9.4 where these features were introduced.