offset past the last-used-item-plus-one, since that would result in
leaving uninitialized holes in the item pointer array.  AFAICT the only
place that was depending on this was btree index build, which was being
cavalier about when to fill in the P_HIKEY pointer; easily fixed.
Also a small performance improvement: shuffle itemid's by means of
memmove, not a one-at-a-time loop.

pg_convert3 does not belong in a backend/catalog file at all, IMHO;
it should be in utils/adt.

Teodor Sigaev

rangefunc regression test for the new behavior.

Joe Conway

>> alias in this case.  What do you think?
>
> I guess that would make sense. I'll make a separate patch just for
that
> change if that's OK.
>

Simple change -- patch attached.

test=# select * from myfoo1() as z;
  z
----
   1
   2
   3
(3 rows)

Joe Conway

don't return type RECORD. It also catches a core dump condition when a
function returning RECORD had an alias list instead of a coldeflist.

Now both conditions throw an ERROR.

Joe Conway

START TRANSACTION command.

Neil Conway

executed in an implicitely aborted transaction (e.g. after an occur
occurs), we return an error (and not just a warning). For example:

nconway=# begin;
BEGIN
nconway=# insert; -- syntax error
ERROR:  parser: parse error at or near ";"
nconway=# select * from a;
ERROR:  current transaction is aborted, queries ignored until end of
transaction block

The old behavior was:

nconway=# begin;
BEGIN
nconway=# insert;
ERROR:  parser: parse error at or near ";"
nconway=# select * from a;
WARNING:  current transaction is aborted, queries ignored until end
of transaction block
*ABORT STATE*

Which can be confusing: if the client isn't paying careful attention,
they will conclude that the query has executed (because no error is
returned).

Neil Conway

The local buffer manager is no longer used for newly-created relations
(unless they are TEMP); a new non-TEMP relation goes through the shared
bufmgr and thus will participate normally in checkpoints.  But TEMP relations
use the local buffer manager throughout their lifespan.  Also, operations
in TEMP relations are not logged in WAL, thus improving performance.
Since it's no longer necessary to fsync relations as they move out of the
local buffers into shared buffers, quite a lot of smgr.c/md.c/fd.c code
is no longer needed and has been removed: there's no concept of a dirty
relation anymore in md.c/fd.c, and we never fsync anything but WAL.
Still TODO: improve local buffer management algorithms so that it would
be reasonable to increase NLocBuffer.

Still needs to be filled with more information, but it gives us a
framework to have a User's Guide with complete coverage of the basic
SQL operations.  Move arrays into data type chapter, inheritance into
DDL chapter (for now).

Make <comment>s show up in the output while the version number ends in
"devel".

Allow cross-book references with entities &cite-user; etc.

hardwired lists of index names for each catalog, use the relcache's
mechanism for caching lists of OIDs of indexes of any table.  This
reduces the common case of updating system catalog indexes to a single
line, makes it much easier to add a new system index (in fact, you
can now do so on-the-fly if you want to), and as a nice side benefit
improves performance a little.  Per recent pghackers discussion.

of functions returning domain types, update documentation for typtype,
move get_typtype to lsyscache.c (actually, resurrect the old version),
add defense against creating pseudo-typed table columns, fix some
bogus list-parsing in grammar.  Issues remain with respect to alias
handling and type checking; Joe is on those.

 location feature.

in checkretval about which paths are for base or complex return type.

types for Table Functions, as previously proposed on HACKERS. Here is a
brief explanation:

1. Creates a new pg_type typtype: 'p' for pseudo type (currently either
     'b' for base or 'c' for catalog, i.e. a class).

2. Creates new builtin type of typtype='p' named RECORD. This is the
     first of potentially several pseudo types.

3. Modify FROM clause grammer to accept:
     SELECT * FROM my_func() AS m(colname1 type1, colname2 type1, ...)
     where m is the table alias, colname1, etc are the column names, and
     type1, etc are the column types.

4. When typtype == 'p' and the function return type is RECORD, a list
     of column defs is required, and when typtype != 'p', it is
disallowed.

5. A check was added to ensure that the tupdesc provide via the parser
     and the actual return tupdesc match in number and type of
attributes.

When creating a function you can do:
     CREATE FUNCTION foo(text) RETURNS setof RECORD ...

When using it you can do:
     SELECT * from foo(sqlstmt) AS (f1 int, f2 text, f3 timestamp)
       or
     SELECT * from foo(sqlstmt) AS f(f1 int, f2 text, f3 timestamp)
       or
     SELECT * from foo(sqlstmt) f(f1 int, f2 text, f3 timestamp)

Included in the patches are adjustments to the regression test sql and
expected files, and documentation.

p.s.
     This potentially solves (or at least improves) the issue of builtin
     Table Functions. They can be bootstrapped as returning RECORD, and
     we can wrap system views around them with properly specified column
     defs. For example:

     CREATE VIEW pg_settings AS
       SELECT s.name, s.setting
       FROM show_all_settings()AS s(name text, setting text);

     Then we can also add the UPDATE RULE that I previously posted to
     pg_settings, and have pg_settings act like a virtual table, allowing
     settings to be queried and set.


Joe Conway

 Can now do queries of the form: SELECT value IS OF (integer, float8);
Define macros for handling typmod manipulation for date/time types.
 Should be more robust than all of that brute-force inline code.
Rename macros for masking and typmod manipulation to put TIMESTAMP_
 or INTERVAL_ in front of the macro name, to reduce the possibility
 of name space collisions.
Allow bit string constants without fully-specified length declaration.
Try implementing CREATE TABLE/OF as a mapping to inheritance.
 May be appropriate, or may be replace later with something more exactly
 like one might expect from databases without the feature.

 select value IS OF (integer, float8);

 strings. Should go back in and look at doing this a bit more elegantly
 and (hopefully) cheaper. Probably not too bad anyway, but it seems a
 shame to scan the strings twice: once for length for this buffer overrun
 protection, and once to parse the line.
Remove use of pow() in date/time handling; was already gone from everything
 *but* the time data types.
Define macros for handling typmod manipulation for date/time types.
 Should be more robust than all of that brute-force inline code.
Rename macros for masking and typmod manipulation to put TIMESTAMP_
 or INTERVAL_ in front of the macro name, to reduce the possibility
 of name space collisions.

 Should be more robust than all of that brute-force inline code.
Rename macros for masking and typmod manipulation to put TIMESTAMP_
 or INTERVAL_ in front of the macro name, to reduce the possibility
 of name space collisions.

 May not be the long-term solution (some continuing discussion with
 Peter E.) but better than the current mapping of a conversion to integer
 which I'd put in years ago before we had any bit string types at all.
This is already supported in the bit string implementation elsewhere.

Implement conversion between 8-byte integers and bit strings.
 Similar to what is done for 4-byte integers.

Which was true when he submitted it, but is so no longer.