> * Make error messages more consistent [error]

width types and varlena types, since with the introduction of CSTRING as
a more-or-less-real type, these concepts aren't identical.  I've tried to
use varlena consistently to denote datatypes with typlen = -1, ie, they
have a length word and are potentially TOASTable; while the term variable
width covers both varlena and cstring (and, perhaps, someday other types
with other rules for computing the actual width).  No code changes in this
commit except for renaming a couple macros.

value '-2' is used to indicate a variable-width type whose width is
computed as strlen(datum)+1.  Everything that looks at typlen is updated
except for array support, which Joe Conway is working on; at the moment
it wouldn't work to try to create an array of cstring.

> * -SELECT cash_out(2) crashes because of opaque

the server.  Previously we allocated a new String object for the entire final
query we were sending to the database.  If you had a big query, or especially
if you had large bind values you ended up with essentially two copies in memory.
This change will reuse the existing objects and therefore should take 1/2 the
memory it does today for a given query.  This restructuring will also allow
in the future the ability to stream bytea data to the server instead of the current approach of pulling it all into memory.
I also fixed a test that was failing on a 7.2 database.
Also renamed some internal variables and some minor cleanup.

 Modified Files:
 	jdbc/org/postgresql/core/QueryExecutor.java
 	jdbc/org/postgresql/jdbc1/AbstractJdbc1Connection.java
 	jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java
 	jdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSet.java
 	jdbc/org/postgresql/jdbc2/AbstractJdbc2Statement.java
 	jdbc/org/postgresql/test/jdbc2/DatabaseMetaDataTest.java

the declarations of some index access method support functions.  Support
SQL functions returning VOID.

instance in which the docs mention that 'NAMEDATALEN == 32".

Neil Conway

> * Allow xlog directory location to be specified during initdb, perhaps
>   using symlinks

Teodor Sigaev

settings of NAMEDATALEN. I looked through the docs for other
references to NAMEDATALEN, but this is the only one I could find.

Neil Conway

> * Add GUC parameter to print queries that generate errors

docs and find the other obsolete statements that no doubt lurk.

for a good long while, I see no reason to maintain this original copy.

Remove all traces of the ODBC driver, which is now on GBorg as the psqlodbc
project ...

addition of a column list clause to the COPY command.  Spotted by
Martin Renters.

tighten foreign-key check (a self-reference should not prevent TRUNCATE),
improve error message, cause a relation's TOAST table to be truncated
along with the relation.

>>" It's also possible to select no escape character by writing ESCAPE ''.
>>In this case there is no way to turn off the special meaning of
>>underscore and percent signs in the pattern."

Joe Conway

for repeat(). Again, somewhat off-the-cuff, so I might have missed
something...

test=# select lpad('xxxxx',1431655765,'yyyyyyyyyyyyyyyy');
ERROR:  Requested length too large
test=# select rpad('xxxxx',1431655765,'yyyyyyyyyyyyyyyy');
ERROR:  Requested length too large

(That's on a Unicode DB, haven't tested other encodings but AFAICT
this fix should still work.)

Neil Conway

> Neil Conway <neilc@samurai.com> writes:
> > +   /* Check for integer overflow */
> > +   if (tlen / slen != count)
> > +           elog(ERROR, "Requested buffer is too large.");
>
> What about slen == 0?

Good point -- that wouldn't cause incorrect results or a security
problem, but it would reject input that we should really accept.

Revised patch is attached.

Neil Conway

> > is a patch to add some cross-referencing.

Oliver Elphick

constraints


The issue with finding and removing foreign key constraints is no longer
an issue, so please apply the attached.

It does NOT check for rules or on delete triggers (old style foreign
keys) as those are difficult to deal with (remove, truncate, re-add).

Rod Taylor

> * Allow bytea to handle LIKE with non-TEXT patterns

bytealike to TEXT.

This leaves like_escape_bytea() without anything to do, but I left it in
place in anticipation of the eventual bytea pattern selectivity
functions. If there is agreement that this would be the best long term
solution, I'll take it as a TODO for 7.4.

Joe Conway

replace(string, from, to)
   -- replaces all occurrences of "from" in "string" to "to"
split(string, fldsep, column)
   -- splits "string" on "fldsep" and returns "column" number piece
to_hex(int32_num) & to_hex(int64_num)
   -- takes integer number and returns as hex string

Joe Conway