You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
postgres/src/pl/plperl/plperl.c

1831 lines
46 KiB

/**********************************************************************
* plperl.c - perl as a procedural language for PostgreSQL
*
* IDENTIFICATION
*
* This software is copyrighted by Mark Hollomon
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
* but is shameless cribbed from pltcl.c by Jan Wieck.
*
* The author hereby grants permission to use, copy, modify,
* distribute, and license this software and its documentation
* for any purpose, provided that existing copyright notices are
* retained in all copies and that this notice is included
* verbatim in any distributions. No written agreement, license,
* or royalty fee is required for any of the authorized uses.
* Modifications to this software may be copyrighted by their
* author and need not follow the licensing terms described
* here, provided that the new terms are clearly indicated on
* the first page of each file where they apply.
*
* IN NO EVENT SHALL THE AUTHOR OR DISTRIBUTORS BE LIABLE TO ANY
* PARTY FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR
* CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OF THIS
* SOFTWARE, ITS DOCUMENTATION, OR ANY DERIVATIVES THEREOF, EVEN
* IF THE AUTHOR HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH
* DAMAGE.
*
* THE AUTHOR AND DISTRIBUTORS SPECIFICALLY DISCLAIM ANY
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
* WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
* PURPOSE, AND NON-INFRINGEMENT. THIS SOFTWARE IS PROVIDED ON
* AN "AS IS" BASIS, AND THE AUTHOR AND DISTRIBUTORS HAVE NO
* OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES,
* ENHANCEMENTS, OR MODIFICATIONS.
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/pl/plperl/plperl.c,v 1.98 2005/12/29 14:28:31 adunstan Exp $
*
**********************************************************************/
#include "postgres.h"
/* Defined by Perl */
#undef _
/* system stuff */
#include <ctype.h>
#include <fcntl.h>
#include <unistd.h>
/* postgreSQL stuff */
#include "commands/trigger.h"
#include "executor/spi.h"
#include "funcapi.h"
#include "utils/lsyscache.h"
#include "utils/memutils.h"
#include "utils/typcache.h"
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
#include "miscadmin.h"
#include "mb/pg_wchar.h"
/* define this before the perl headers get a chance to mangle DLLIMPORT */
extern DLLIMPORT bool check_function_bodies;
/* perl stuff */
#include "EXTERN.h"
#include "perl.h"
#include "XSUB.h"
#include "ppport.h"
#include "spi_internal.h"
/* just in case these symbols aren't provided */
#ifndef pTHX_
#define pTHX_
#define pTHX void
#endif
/**********************************************************************
* The information we cache about loaded procedures
**********************************************************************/
typedef struct plperl_proc_desc
{
char *proname;
TransactionId fn_xmin;
CommandId fn_cmin;
bool fn_readonly;
bool lanpltrusted;
bool fn_retistuple; /* true, if function returns tuple */
bool fn_retisset; /* true, if function returns set */
bool fn_retisarray; /* true if function returns array */
Oid result_oid; /* Oid of result type */
FmgrInfo result_in_func; /* I/O function and arg for result type */
Oid result_typioparam;
int nargs;
FmgrInfo arg_out_func[FUNC_MAX_ARGS];
bool arg_is_rowtype[FUNC_MAX_ARGS];
SV *reference;
} plperl_proc_desc;
/**********************************************************************
* Global data
**********************************************************************/
static int plperl_firstcall = 1;
static bool plperl_safe_init_done = false;
static PerlInterpreter *plperl_interp = NULL;
static HV *plperl_proc_hash = NULL;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
static bool plperl_use_strict = false;
/* these are saved and restored by plperl_call_handler */
static plperl_proc_desc *plperl_current_prodesc = NULL;
static FunctionCallInfo plperl_current_caller_info;
static Tuplestorestate *plperl_current_tuple_store;
static TupleDesc plperl_current_tuple_desc;
/**********************************************************************
* Forward declarations
**********************************************************************/
static void plperl_init_all(void);
static void plperl_init_interp(void);
Datum plperl_call_handler(PG_FUNCTION_ARGS);
Datum plperl_validator(PG_FUNCTION_ARGS);
void plperl_init(void);
static Datum plperl_func_handler(PG_FUNCTION_ARGS);
static Datum plperl_trigger_handler(PG_FUNCTION_ARGS);
static plperl_proc_desc *compile_plperl_function(Oid fn_oid, bool is_trigger);
static SV *plperl_hash_from_tuple(HeapTuple tuple, TupleDesc tupdesc);
static void plperl_init_shared_libs(pTHX);
static HV *plperl_spi_execute_fetch_result(SPITupleTable *, int, int);
/*
* This routine is a crock, and so is everyplace that calls it. The problem
* is that the cached form of plperl functions/queries is allocated permanently
* (mostly via malloc()) and never released until backend exit. Subsidiary
* data structures such as fmgr info records therefore must live forever
* as well. A better implementation would store all this stuff in a per-
* function memory context that could be reclaimed at need. In the meantime,
* fmgr_info_cxt must be called specifying TopMemoryContext so that whatever
* it might allocate, and whatever the eventual function might allocate using
* fn_mcxt, will live forever too.
*/
static void
perm_fmgr_info(Oid functionId, FmgrInfo *finfo)
{
fmgr_info_cxt(functionId, finfo, TopMemoryContext);
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Perform initialization during postmaster startup. */
void
plperl_init(void)
{
if (!plperl_firstcall)
return;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
DefineCustomBoolVariable(
"plperl.use_strict",
"If true, will compile trusted and untrusted perl code in strict mode",
NULL,
&plperl_use_strict,
PGC_USERSET,
NULL, NULL);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
EmitWarningsOnPlaceholders("plperl");
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
plperl_init_interp();
plperl_firstcall = 0;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Perform initialization during backend startup. */
static void
plperl_init_all(void)
{
if (plperl_firstcall)
plperl_init();
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* We don't need to do anything yet when a new backend starts. */
}
/* Each of these macros must represent a single string literal */
#define PERLBOOT \
"SPI::bootstrap(); use vars qw(%_SHARED);" \
"sub ::plperl_warn { my $msg = shift; " \
" $msg =~ s/\\(eval \\d+\\) //g; &elog(&NOTICE, $msg); } " \
"$SIG{__WARN__} = \\&::plperl_warn; " \
"sub ::plperl_die { my $msg = shift; " \
" $msg =~ s/\\(eval \\d+\\) //g; die $msg; } " \
"$SIG{__DIE__} = \\&::plperl_die; " \
"sub ::mkunsafefunc {" \
" my $ret = eval(qq[ sub { $_[0] $_[1] } ]); " \
" $@ =~ s/\\(eval \\d+\\) //g if $@; return $ret; }" \
"use strict; " \
"sub ::mk_strict_unsafefunc {" \
" my $ret = eval(qq[ sub { use strict; $_[0] $_[1] } ]); " \
" $@ =~ s/\\(eval \\d+\\) //g if $@; return $ret; } " \
"sub ::_plperl_to_pg_array {" \
" my $arg = shift; ref $arg eq 'ARRAY' || return $arg; " \
" my $res = ''; my $first = 1; " \
" foreach my $elem (@$arg) " \
" { " \
" $res .= ', ' unless $first; $first = undef; " \
" if (ref $elem) " \
" { " \
" $res .= _plperl_to_pg_array($elem); " \
" } " \
" elsif (defined($elem)) " \
" { " \
" my $str = qq($elem); " \
" $str =~ s/([\"\\\\])/\\\\$1/g; " \
" $res .= qq(\"$str\"); " \
" } " \
" else " \
" { "\
" $res .= 'NULL' ; " \
" } "\
" } " \
" return qq({$res}); " \
"} "
#define SAFE_MODULE \
"require Safe; $Safe::VERSION"
#define SAFE_OK \
"use vars qw($PLContainer); $PLContainer = new Safe('PLPerl');" \
"$PLContainer->permit_only(':default');" \
"$PLContainer->permit(qw[:base_math !:base_io sort time]);" \
"$PLContainer->share(qw[&elog &spi_exec_query &return_next " \
"&spi_query &spi_fetchrow " \
"&_plperl_to_pg_array " \
"&DEBUG &LOG &INFO &NOTICE &WARNING &ERROR %_SHARED ]);" \
"sub ::mksafefunc {" \
" my $ret = $PLContainer->reval(qq[sub { $_[0] $_[1] }]); " \
" $@ =~ s/\\(eval \\d+\\) //g if $@; return $ret; }" \
"$PLContainer->permit('require'); $PLContainer->reval('use strict;');" \
"$PLContainer->deny('require');" \
"sub ::mk_strict_safefunc {" \
" my $ret = $PLContainer->reval(qq[sub { BEGIN { strict->import(); } $_[0] $_[1] }]); " \
" $@ =~ s/\\(eval \\d+\\) //g if $@; return $ret; }"
#define SAFE_BAD \
"use vars qw($PLContainer); $PLContainer = new Safe('PLPerl');" \
"$PLContainer->permit_only(':default');" \
"$PLContainer->share(qw[&elog &ERROR ]);" \
"sub ::mksafefunc { return $PLContainer->reval(qq[sub { " \
" elog(ERROR,'trusted Perl functions disabled - " \
" please upgrade Perl Safe module to version 2.09 or later');}]); }" \
"sub ::mk_strict_safefunc { return $PLContainer->reval(qq[sub { " \
" elog(ERROR,'trusted Perl functions disabled - " \
" please upgrade Perl Safe module to version 2.09 or later');}]); }"
static void
plperl_init_interp(void)
{
static char *embedding[3] = {
"", "-e", PERLBOOT
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
};
plperl_interp = perl_alloc();
if (!plperl_interp)
elog(ERROR, "could not allocate Perl interpreter");
perl_construct(plperl_interp);
perl_parse(plperl_interp, plperl_init_shared_libs, 3, embedding, NULL);
perl_run(plperl_interp);
plperl_proc_hash = newHV();
}
static void
plperl_safe_init(void)
{
SV *res;
double safe_version;
res = eval_pv(SAFE_MODULE, FALSE); /* TRUE = croak if failure */
safe_version = SvNV(res);
/*
* We actually want to reject safe_version < 2.09, but it's risky to
* assume that floating-point comparisons are exact, so use a slightly
* smaller comparison value.
*/
if (safe_version < 2.0899)
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
{
/* not safe, so disallow all trusted funcs */
eval_pv(SAFE_BAD, FALSE);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
}
else
{
eval_pv(SAFE_OK, FALSE);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
}
plperl_safe_init_done = true;
}
/*
* Perl likes to put a newline after its error messages; clean up such
*/
static char *
strip_trailing_ws(const char *msg)
{
char *res = pstrdup(msg);
int len = strlen(res);
while (len > 0 && isspace((unsigned char) res[len - 1]))
res[--len] = '\0';
return res;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Build a tuple from a hash. */
static HeapTuple
plperl_build_tuple_result(HV *perlhash, AttInMetadata *attinmeta)
{
TupleDesc td = attinmeta->tupdesc;
char **values;
SV *val;
char *key;
I32 klen;
HeapTuple tup;
values = (char **) palloc0(td->natts * sizeof(char *));
hv_iterinit(perlhash);
while ((val = hv_iternextsv(perlhash, &key, &klen)))
{
int attn = SPI_fnumber(td, key);
if (attn <= 0 || td->attrs[attn - 1]->attisdropped)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("Perl hash contains nonexistent column \"%s\"",
key)));
if (SvOK(val) && SvTYPE(val) != SVt_NULL)
values[attn - 1] = SvPV(val, PL_na);
}
hv_iterinit(perlhash);
tup = BuildTupleFromCStrings(attinmeta, values);
pfree(values);
return tup;
}
/*
* convert perl array to postgres string representation
*/
static SV *
plperl_convert_to_pg_array(SV *src)
{
SV *rv;
int count;
dSP;
PUSHMARK(SP);
XPUSHs(src);
PUTBACK;
count = call_pv("::_plperl_to_pg_array", G_SCALAR);
SPAGAIN;
if (count != 1)
elog(ERROR, "unexpected _plperl_to_pg_array failure");
rv = POPs;
PUTBACK;
return rv;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Set up the arguments for a trigger call. */
static SV *
plperl_trigger_build_args(FunctionCallInfo fcinfo)
{
TriggerData *tdata;
TupleDesc tupdesc;
int i;
char *level;
char *event;
char *relid;
char *when;
HV *hv;
hv = newHV();
tdata = (TriggerData *) fcinfo->context;
tupdesc = tdata->tg_relation->rd_att;
relid = DatumGetCString(
DirectFunctionCall1(oidout,
ObjectIdGetDatum(tdata->tg_relation->rd_id)
)
);
hv_store(hv, "name", 4, newSVpv(tdata->tg_trigger->tgname, 0), 0);
hv_store(hv, "relid", 5, newSVpv(relid, 0), 0);
if (TRIGGER_FIRED_BY_INSERT(tdata->tg_event))
{
event = "INSERT";
if (TRIGGER_FIRED_FOR_ROW(tdata->tg_event))
hv_store(hv, "new", 3,
plperl_hash_from_tuple(tdata->tg_trigtuple, tupdesc),
0);
}
else if (TRIGGER_FIRED_BY_DELETE(tdata->tg_event))
{
event = "DELETE";
if (TRIGGER_FIRED_FOR_ROW(tdata->tg_event))
hv_store(hv, "old", 3,
plperl_hash_from_tuple(tdata->tg_trigtuple, tupdesc),
0);
}
else if (TRIGGER_FIRED_BY_UPDATE(tdata->tg_event))
{
event = "UPDATE";
if (TRIGGER_FIRED_FOR_ROW(tdata->tg_event))
{
hv_store(hv, "old", 3,
plperl_hash_from_tuple(tdata->tg_trigtuple, tupdesc),
0);
hv_store(hv, "new", 3,
plperl_hash_from_tuple(tdata->tg_newtuple, tupdesc),
0);
}
}
else
event = "UNKNOWN";
hv_store(hv, "event", 5, newSVpv(event, 0), 0);
hv_store(hv, "argc", 4, newSViv(tdata->tg_trigger->tgnargs), 0);
if (tdata->tg_trigger->tgnargs > 0)
{
AV *av = newAV();
for (i = 0; i < tdata->tg_trigger->tgnargs; i++)
av_push(av, newSVpv(tdata->tg_trigger->tgargs[i], 0));
hv_store(hv, "args", 4, newRV_noinc((SV *) av), 0);
}
hv_store(hv, "relname", 7,
newSVpv(SPI_getrelname(tdata->tg_relation), 0), 0);
if (TRIGGER_FIRED_BEFORE(tdata->tg_event))
when = "BEFORE";
else if (TRIGGER_FIRED_AFTER(tdata->tg_event))
when = "AFTER";
else
when = "UNKNOWN";
hv_store(hv, "when", 4, newSVpv(when, 0), 0);
if (TRIGGER_FIRED_FOR_ROW(tdata->tg_event))
level = "ROW";
else if (TRIGGER_FIRED_FOR_STATEMENT(tdata->tg_event))
level = "STATEMENT";
else
level = "UNKNOWN";
hv_store(hv, "level", 5, newSVpv(level, 0), 0);
return newRV_noinc((SV *) hv);
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Set up the new tuple returned from a trigger. */
static HeapTuple
plperl_modify_tuple(HV *hvTD, TriggerData *tdata, HeapTuple otup)
{
SV **svp;
HV *hvNew;
HeapTuple rtup;
SV *val;
char *key;
I32 klen;
int slotsused;
int *modattrs;
Datum *modvalues;
char *modnulls;
TupleDesc tupdesc;
tupdesc = tdata->tg_relation->rd_att;
svp = hv_fetch(hvTD, "new", 3, FALSE);
if (!svp)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("$_TD->{new} does not exist")));
if (!SvOK(*svp) || SvTYPE(*svp) != SVt_RV || SvTYPE(SvRV(*svp)) != SVt_PVHV)
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
errmsg("$_TD->{new} is not a hash reference")));
hvNew = (HV *) SvRV(*svp);
modattrs = palloc(tupdesc->natts * sizeof(int));
modvalues = palloc(tupdesc->natts * sizeof(Datum));
modnulls = palloc(tupdesc->natts * sizeof(char));
slotsused = 0;
hv_iterinit(hvNew);
while ((val = hv_iternextsv(hvNew, &key, &klen)))
{
int attn = SPI_fnumber(tupdesc, key);
if (attn <= 0 || tupdesc->attrs[attn - 1]->attisdropped)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_COLUMN),
errmsg("Perl hash contains nonexistent column \"%s\"",
key)));
if (SvOK(val) && SvTYPE(val) != SVt_NULL)
{
Oid typinput;
Oid typioparam;
FmgrInfo finfo;
/* XXX would be better to cache these lookups */
getTypeInputInfo(tupdesc->attrs[attn - 1]->atttypid,
&typinput, &typioparam);
fmgr_info(typinput, &finfo);
modvalues[slotsused] = FunctionCall3(&finfo,
CStringGetDatum(SvPV(val, PL_na)),
ObjectIdGetDatum(typioparam),
Int32GetDatum(tupdesc->attrs[attn - 1]->atttypmod));
modnulls[slotsused] = ' ';
}
else
{
modvalues[slotsused] = (Datum) 0;
modnulls[slotsused] = 'n';
}
modattrs[slotsused] = attn;
slotsused++;
}
hv_iterinit(hvNew);
rtup = SPI_modifytuple(tdata->tg_relation, otup, slotsused,
modattrs, modvalues, modnulls);
pfree(modattrs);
pfree(modvalues);
pfree(modnulls);
if (rtup == NULL)
elog(ERROR, "SPI_modifytuple failed: %s",
SPI_result_code_string(SPI_result));
return rtup;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/*
* This is the only externally-visible part of the plperl call interface.
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
* The Postgres function and trigger managers call it to execute a
* perl function.
*/
PG_FUNCTION_INFO_V1(plperl_call_handler);
Datum
plperl_call_handler(PG_FUNCTION_ARGS)
{
Datum retval;
plperl_proc_desc *save_prodesc;
FunctionCallInfo save_caller_info;
Tuplestorestate *save_tuple_store;
TupleDesc save_tuple_desc;
plperl_init_all();
save_prodesc = plperl_current_prodesc;
save_caller_info = plperl_current_caller_info;
save_tuple_store = plperl_current_tuple_store;
save_tuple_desc = plperl_current_tuple_desc;
PG_TRY();
{
if (CALLED_AS_TRIGGER(fcinfo))
retval = PointerGetDatum(plperl_trigger_handler(fcinfo));
else
retval = plperl_func_handler(fcinfo);
}
PG_CATCH();
{
plperl_current_prodesc = save_prodesc;
plperl_current_caller_info = save_caller_info;
plperl_current_tuple_store = save_tuple_store;
plperl_current_tuple_desc = save_tuple_desc;
PG_RE_THROW();
}
PG_END_TRY();
plperl_current_prodesc = save_prodesc;
plperl_current_caller_info = save_caller_info;
plperl_current_tuple_store = save_tuple_store;
plperl_current_tuple_desc = save_tuple_desc;
return retval;
}
/*
* This is the other externally visible function - it is called when CREATE
* FUNCTION is issued to validate the function being created/replaced.
*/
PG_FUNCTION_INFO_V1(plperl_validator);
Datum
plperl_validator(PG_FUNCTION_ARGS)
{
Oid funcoid = PG_GETARG_OID(0);
HeapTuple tuple;
Form_pg_proc proc;
char functyptype;
int numargs;
Oid *argtypes;
char **argnames;
char *argmodes;
bool istrigger = false;
int i;
/* Get the new function's pg_proc entry */
tuple = SearchSysCache(PROCOID,
ObjectIdGetDatum(funcoid),
0, 0, 0);
if (!HeapTupleIsValid(tuple))
elog(ERROR, "cache lookup failed for function %u", funcoid);
proc = (Form_pg_proc) GETSTRUCT(tuple);
functyptype = get_typtype(proc->prorettype);
/* Disallow pseudotype result */
/* except for TRIGGER, RECORD, or VOID */
if (functyptype == 'p')
{
/* we assume OPAQUE with no arguments means a trigger */
if (proc->prorettype == TRIGGEROID ||
(proc->prorettype == OPAQUEOID && proc->pronargs == 0))
istrigger = true;
else if (proc->prorettype != RECORDOID &&
proc->prorettype != VOIDOID)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("plperl functions cannot return type %s",
format_type_be(proc->prorettype))));
}
/* Disallow pseudotypes in arguments (either IN or OUT) */
numargs = get_func_arg_info(tuple,
&argtypes, &argnames, &argmodes);
for (i = 0; i < numargs; i++)
{
if (get_typtype(argtypes[i]) == 'p')
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("plperl functions cannot take type %s",
format_type_be(argtypes[i]))));
}
ReleaseSysCache(tuple);
/* Postpone body checks if !check_function_bodies */
if (check_function_bodies)
{
plperl_proc_desc *prodesc;
plperl_init_all();
prodesc = compile_plperl_function(funcoid, istrigger);
}
/* the result of a validator is ignored */
PG_RETURN_VOID();
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Uses mksafefunc/mkunsafefunc to create an anonymous sub whose text is
* supplied in s, and returns a reference to the closure. */
22 years ago
static SV *
plperl_create_sub(char *s, bool trusted)
{
dSP;
SV *subref;
int count;
char *compile_sub;
if (trusted && !plperl_safe_init_done)
{
plperl_safe_init();
SPAGAIN;
}
ENTER;
SAVETMPS;
PUSHMARK(SP);
XPUSHs(sv_2mortal(newSVpv("my $_TD=$_[0]; shift;", 0)));
XPUSHs(sv_2mortal(newSVpv(s, 0)));
PUTBACK;
22 years ago
/*
* G_KEEPERR seems to be needed here, else we don't recognize compile
* errors properly. Perhaps it's because there's another level of eval
* inside mksafefunc?
*/
if (trusted && plperl_use_strict)
compile_sub = "::mk_strict_safefunc";
else if (plperl_use_strict)
compile_sub = "::mk_strict_unsafefunc";
else if (trusted)
compile_sub = "::mksafefunc";
else
compile_sub = "::mkunsafefunc";
count = perl_call_pv(compile_sub, G_SCALAR | G_EVAL | G_KEEPERR);
SPAGAIN;
if (count != 1)
{
PUTBACK;
FREETMPS;
LEAVE;
elog(ERROR, "didn't get a return item from mksafefunc");
}
if (SvTRUE(ERRSV))
{
(void) POPs;
PUTBACK;
FREETMPS;
LEAVE;
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("creation of Perl function failed: %s",
strip_trailing_ws(SvPV(ERRSV, PL_na)))));
}
/*
* need to make a deep copy of the return. it comes off the stack as a
* temporary.
*/
subref = newSVsv(POPs);
if (!SvROK(subref) || SvTYPE(SvRV(subref)) != SVt_PVCV)
{
PUTBACK;
FREETMPS;
LEAVE;
/*
* subref is our responsibility because it is not mortal
*/
SvREFCNT_dec(subref);
elog(ERROR, "didn't get a code ref");
}
PUTBACK;
FREETMPS;
LEAVE;
return subref;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/**********************************************************************
* plperl_init_shared_libs() -
*
* We cannot use the DynaLoader directly to get at the Opcode
* module (used by Safe.pm). So, we link Opcode into ourselves
* and do the initialization behind perl's back.
*
**********************************************************************/
EXTERN_C void boot_DynaLoader(pTHX_ CV *cv);
EXTERN_C void boot_SPI(pTHX_ CV *cv);
static void
plperl_init_shared_libs(pTHX)
{
char *file = __FILE__;
newXS("DynaLoader::boot_DynaLoader", boot_DynaLoader, file);
newXS("SPI::bootstrap", boot_SPI, file);
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
22 years ago
static SV *
plperl_call_perl_func(plperl_proc_desc *desc, FunctionCallInfo fcinfo)
{
dSP;
SV *retval;
int i;
int count;
SV *sv;
ENTER;
SAVETMPS;
PUSHMARK(SP);
XPUSHs(&PL_sv_undef); /* no trigger data */
for (i = 0; i < desc->nargs; i++)
{
if (fcinfo->argnull[i])
XPUSHs(&PL_sv_undef);
else if (desc->arg_is_rowtype[i])
{
HeapTupleHeader td;
Oid tupType;
int32 tupTypmod;
TupleDesc tupdesc;
HeapTupleData tmptup;
SV *hashref;
td = DatumGetHeapTupleHeader(fcinfo->arg[i]);
/* Extract rowtype info and find a tupdesc */
tupType = HeapTupleHeaderGetTypeId(td);
tupTypmod = HeapTupleHeaderGetTypMod(td);
tupdesc = lookup_rowtype_tupdesc(tupType, tupTypmod);
/* Build a temporary HeapTuple control structure */
tmptup.t_len = HeapTupleHeaderGetDatumLength(td);
tmptup.t_data = td;
hashref = plperl_hash_from_tuple(&tmptup, tupdesc);
XPUSHs(sv_2mortal(hashref));
}
else
{
char *tmp;
tmp = DatumGetCString(FunctionCall1(&(desc->arg_out_func[i]),
fcinfo->arg[i]));
sv = newSVpv(tmp, 0);
#if PERL_BCDVERSION >= 0x5006000L
if (GetDatabaseEncoding() == PG_UTF8)
SvUTF8_on(sv);
#endif
XPUSHs(sv_2mortal(sv));
pfree(tmp);
}
}
PUTBACK;
/* Do NOT use G_KEEPERR here */
count = perl_call_sv(desc->reference, G_SCALAR | G_EVAL);
SPAGAIN;
if (count != 1)
{
PUTBACK;
FREETMPS;
LEAVE;
elog(ERROR, "didn't get a return item from function");
}
if (SvTRUE(ERRSV))
{
(void) POPs;
PUTBACK;
FREETMPS;
LEAVE;
/* XXX need to find a way to assign an errcode here */
ereport(ERROR,
(errmsg("error from Perl function: %s",
strip_trailing_ws(SvPV(ERRSV, PL_na)))));
}
retval = newSVsv(POPs);
PUTBACK;
FREETMPS;
LEAVE;
return retval;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
static SV *
plperl_call_perl_trigger_func(plperl_proc_desc *desc, FunctionCallInfo fcinfo,
SV *td)
{
dSP;
SV *retval;
Trigger *tg_trigger;
int i;
int count;
ENTER;
SAVETMPS;
PUSHMARK(sp);
XPUSHs(td);
tg_trigger = ((TriggerData *) fcinfo->context)->tg_trigger;
for (i = 0; i < tg_trigger->tgnargs; i++)
XPUSHs(sv_2mortal(newSVpv(tg_trigger->tgargs[i], 0)));
PUTBACK;
/* Do NOT use G_KEEPERR here */
count = perl_call_sv(desc->reference, G_SCALAR | G_EVAL);
SPAGAIN;
if (count != 1)
{
PUTBACK;
FREETMPS;
LEAVE;
elog(ERROR, "didn't get a return item from trigger function");
}
if (SvTRUE(ERRSV))
{
(void) POPs;
PUTBACK;
FREETMPS;
LEAVE;
/* XXX need to find a way to assign an errcode here */
ereport(ERROR,
(errmsg("error from Perl trigger function: %s",
strip_trailing_ws(SvPV(ERRSV, PL_na)))));
}
retval = newSVsv(POPs);
PUTBACK;
FREETMPS;
LEAVE;
return retval;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
static Datum
plperl_func_handler(PG_FUNCTION_ARGS)
{
plperl_proc_desc *prodesc;
SV *perlret;
Datum retval;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
ReturnSetInfo *rsi;
SV *array_ret = NULL;
if (SPI_connect() != SPI_OK_CONNECT)
elog(ERROR, "could not connect to SPI manager");
prodesc = compile_plperl_function(fcinfo->flinfo->fn_oid, false);
plperl_current_prodesc = prodesc;
plperl_current_caller_info = fcinfo;
plperl_current_tuple_store = 0;
plperl_current_tuple_desc = 0;
rsi = (ReturnSetInfo *) fcinfo->resultinfo;
if (prodesc->fn_retisset)
{
/* Check context before allowing the call to go through */
if (!rsi || !IsA(rsi, ReturnSetInfo) ||
(rsi->allowedModes & SFRM_Materialize) == 0 ||
rsi->expectedDesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("set-valued function called in context that "
"cannot accept a set")));
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
perlret = plperl_call_perl_func(prodesc, fcinfo);
/************************************************************
* Disconnect from SPI manager and then create the return
* values datum (if the input function does a palloc for it
* this must not be allocated in the SPI memory context
* because SPI_finish would free it).
************************************************************/
if (SPI_finish() != SPI_OK_FINISH)
elog(ERROR, "SPI_finish() failed");
if (prodesc->fn_retisset)
{
/*
* If the Perl function returned an arrayref, we pretend that it
* called return_next() for each element of the array, to handle old
* SRFs that didn't know about return_next(). Any other sort of return
* value is an error.
*/
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
if (SvTYPE(perlret) == SVt_RV &&
SvTYPE(SvRV(perlret)) == SVt_PVAV)
{
int i = 0;
SV **svp = 0;
AV *rav = (AV *) SvRV(perlret);
while ((svp = av_fetch(rav, i, FALSE)) != NULL)
{
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
plperl_return_next(*svp);
i++;
}
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
else if (SvTYPE(perlret) != SVt_NULL)
{
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
errmsg("set-returning Perl function must return "
"reference to array or use return_next")));
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
rsi->returnMode = SFRM_Materialize;
if (plperl_current_tuple_store)
{
rsi->setResult = plperl_current_tuple_store;
rsi->setDesc = plperl_current_tuple_desc;
}
retval = (Datum) 0;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
}
else if (SvTYPE(perlret) == SVt_NULL)
{
/* Return NULL if Perl code returned undef */
if (rsi && IsA(rsi, ReturnSetInfo))
rsi->isDone = ExprEndResult;
fcinfo->isnull = true;
retval = (Datum) 0;
}
else if (prodesc->fn_retistuple)
{
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Return a perl hash converted to a Datum */
TupleDesc td;
AttInMetadata *attinmeta;
HeapTuple tup;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
if (!SvOK(perlret) || SvTYPE(perlret) != SVt_RV ||
SvTYPE(SvRV(perlret)) != SVt_PVHV)
{
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
errmsg("composite-returning Perl function "
"must return reference to hash")));
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* XXX should cache the attinmeta data instead of recomputing */
if (get_call_result_type(fcinfo, NULL, &td) != TYPEFUNC_COMPOSITE)
{
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("function returning record called in context "
"that cannot accept type record")));
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
attinmeta = TupleDescGetAttInMetadata(td);
tup = plperl_build_tuple_result((HV *) SvRV(perlret), attinmeta);
retval = HeapTupleGetDatum(tup);
}
else
{
/* Return a perl string converted to a Datum */
char *val;
if (prodesc->fn_retisarray && SvROK(perlret) &&
SvTYPE(SvRV(perlret)) == SVt_PVAV)
{
array_ret = plperl_convert_to_pg_array(perlret);
SvREFCNT_dec(perlret);
perlret = array_ret;
}
val = SvPV(perlret, PL_na);
retval = FunctionCall3(&prodesc->result_in_func,
CStringGetDatum(val),
ObjectIdGetDatum(prodesc->result_typioparam),
Int32GetDatum(-1));
}
if (array_ret == NULL)
SvREFCNT_dec(perlret);
return retval;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
static Datum
plperl_trigger_handler(PG_FUNCTION_ARGS)
{
plperl_proc_desc *prodesc;
SV *perlret;
Datum retval;
SV *svTD;
HV *hvTD;
/* Connect to SPI manager */
if (SPI_connect() != SPI_OK_CONNECT)
elog(ERROR, "could not connect to SPI manager");
/* Find or compile the function */
prodesc = compile_plperl_function(fcinfo->flinfo->fn_oid, true);
plperl_current_prodesc = prodesc;
svTD = plperl_trigger_build_args(fcinfo);
perlret = plperl_call_perl_trigger_func(prodesc, fcinfo, svTD);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
hvTD = (HV *) SvRV(svTD);
/************************************************************
* Disconnect from SPI manager and then create the return
* values datum (if the input function does a palloc for it
* this must not be allocated in the SPI memory context
* because SPI_finish would free it).
************************************************************/
if (SPI_finish() != SPI_OK_FINISH)
elog(ERROR, "SPI_finish() failed");
if (!(perlret && SvOK(perlret) && SvTYPE(perlret) != SVt_NULL))
{
/* undef result means go ahead with original tuple */
TriggerData *trigdata = ((TriggerData *) fcinfo->context);
if (TRIGGER_FIRED_BY_INSERT(trigdata->tg_event))
retval = (Datum) trigdata->tg_trigtuple;
else if (TRIGGER_FIRED_BY_UPDATE(trigdata->tg_event))
retval = (Datum) trigdata->tg_newtuple;
else if (TRIGGER_FIRED_BY_DELETE(trigdata->tg_event))
retval = (Datum) trigdata->tg_trigtuple;
else
retval = (Datum) 0; /* can this happen? */
}
else
{
HeapTuple trv;
char *tmp;
tmp = SvPV(perlret, PL_na);
if (pg_strcasecmp(tmp, "SKIP") == 0)
trv = NULL;
else if (pg_strcasecmp(tmp, "MODIFY") == 0)
{
TriggerData *trigdata = (TriggerData *) fcinfo->context;
if (TRIGGER_FIRED_BY_INSERT(trigdata->tg_event))
trv = plperl_modify_tuple(hvTD, trigdata,
trigdata->tg_trigtuple);
else if (TRIGGER_FIRED_BY_UPDATE(trigdata->tg_event))
trv = plperl_modify_tuple(hvTD, trigdata,
trigdata->tg_newtuple);
else
{
ereport(WARNING,
(errcode(ERRCODE_E_R_I_E_TRIGGER_PROTOCOL_VIOLATED),
errmsg("ignoring modified tuple in DELETE trigger")));
trv = NULL;
}
}
else
{
ereport(ERROR,
(errcode(ERRCODE_E_R_I_E_TRIGGER_PROTOCOL_VIOLATED),
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
errmsg("result of Perl trigger function must be undef, "
"\"SKIP\" or \"MODIFY\"")));
trv = NULL;
}
retval = PointerGetDatum(trv);
}
SvREFCNT_dec(svTD);
if (perlret)
SvREFCNT_dec(perlret);
return retval;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
static plperl_proc_desc *
compile_plperl_function(Oid fn_oid, bool is_trigger)
{
HeapTuple procTup;
Form_pg_proc procStruct;
char internal_proname[64];
int proname_len;
plperl_proc_desc *prodesc = NULL;
int i;
SV **svp;
/* We'll need the pg_proc tuple in any case... */
procTup = SearchSysCache(PROCOID,
ObjectIdGetDatum(fn_oid),
0, 0, 0);
if (!HeapTupleIsValid(procTup))
elog(ERROR, "cache lookup failed for function %u", fn_oid);
procStruct = (Form_pg_proc) GETSTRUCT(procTup);
/************************************************************
* Build our internal proc name from the functions Oid
************************************************************/
if (!is_trigger)
sprintf(internal_proname, "__PLPerl_proc_%u", fn_oid);
else
sprintf(internal_proname, "__PLPerl_proc_%u_trigger", fn_oid);
proname_len = strlen(internal_proname);
/************************************************************
* Lookup the internal proc name in the hashtable
************************************************************/
svp = hv_fetch(plperl_proc_hash, internal_proname, proname_len, FALSE);
if (svp)
{
bool uptodate;
prodesc = (plperl_proc_desc *) SvIV(*svp);
/************************************************************
* If it's present, must check whether it's still up to date.
* This is needed because CREATE OR REPLACE FUNCTION can modify the
* function's pg_proc entry without changing its OID.
************************************************************/
uptodate = (prodesc->fn_xmin == HeapTupleHeaderGetXmin(procTup->t_data) &&
prodesc->fn_cmin == HeapTupleHeaderGetCmin(procTup->t_data));
if (!uptodate)
{
/* need we delete old entry? */
prodesc = NULL;
}
}
/************************************************************
* If we haven't found it in the hashtable, we analyze
* the functions arguments and returntype and store
* the in-/out-functions in the prodesc block and create
* a new hashtable entry for it.
*
* Then we load the procedure into the Perl interpreter.
************************************************************/
if (prodesc == NULL)
{
HeapTuple langTup;
HeapTuple typeTup;
Form_pg_language langStruct;
Form_pg_type typeStruct;
Datum prosrcdatum;
bool isnull;
char *proc_source;
/************************************************************
* Allocate a new procedure description block
************************************************************/
prodesc = (plperl_proc_desc *) malloc(sizeof(plperl_proc_desc));
if (prodesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("out of memory")));
MemSet(prodesc, 0, sizeof(plperl_proc_desc));
prodesc->proname = strdup(internal_proname);
prodesc->fn_xmin = HeapTupleHeaderGetXmin(procTup->t_data);
prodesc->fn_cmin = HeapTupleHeaderGetCmin(procTup->t_data);
/* Remember if function is STABLE/IMMUTABLE */
prodesc->fn_readonly =
(procStruct->provolatile != PROVOLATILE_VOLATILE);
/************************************************************
* Lookup the pg_language tuple by Oid
************************************************************/
langTup = SearchSysCache(LANGOID,
ObjectIdGetDatum(procStruct->prolang),
0, 0, 0);
if (!HeapTupleIsValid(langTup))
{
free(prodesc->proname);
free(prodesc);
elog(ERROR, "cache lookup failed for language %u",
procStruct->prolang);
}
langStruct = (Form_pg_language) GETSTRUCT(langTup);
prodesc->lanpltrusted = langStruct->lanpltrusted;
ReleaseSysCache(langTup);
/************************************************************
* Get the required information for input conversion of the
* return value.
************************************************************/
if (!is_trigger)
{
typeTup = SearchSysCache(TYPEOID,
ObjectIdGetDatum(procStruct->prorettype),
0, 0, 0);
if (!HeapTupleIsValid(typeTup))
{
free(prodesc->proname);
free(prodesc);
elog(ERROR, "cache lookup failed for type %u",
procStruct->prorettype);
}
typeStruct = (Form_pg_type) GETSTRUCT(typeTup);
/* Disallow pseudotype result, except VOID or RECORD */
if (typeStruct->typtype == 'p')
{
if (procStruct->prorettype == VOIDOID ||
procStruct->prorettype == RECORDOID)
23 years ago
/* okay */ ;
else if (procStruct->prorettype == TRIGGEROID)
{
free(prodesc->proname);
free(prodesc);
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
errmsg("trigger functions may only be called "
"as triggers")));
}
else
{
free(prodesc->proname);
free(prodesc);
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("plperl functions cannot return type %s",
format_type_be(procStruct->prorettype))));
}
}
prodesc->result_oid = procStruct->prorettype;
prodesc->fn_retisset = procStruct->proretset;
prodesc->fn_retistuple = (typeStruct->typtype == 'c' ||
procStruct->prorettype == RECORDOID);
prodesc->fn_retisarray =
(typeStruct->typlen == -1 && typeStruct->typelem);
perm_fmgr_info(typeStruct->typinput, &(prodesc->result_in_func));
prodesc->result_typioparam = getTypeIOParam(typeTup);
ReleaseSysCache(typeTup);
}
/************************************************************
* Get the required information for output conversion
* of all procedure arguments
************************************************************/
if (!is_trigger)
{
prodesc->nargs = procStruct->pronargs;
for (i = 0; i < prodesc->nargs; i++)
{
typeTup = SearchSysCache(TYPEOID,
ObjectIdGetDatum(procStruct->proargtypes.values[i]),
0, 0, 0);
if (!HeapTupleIsValid(typeTup))
{
free(prodesc->proname);
free(prodesc);
elog(ERROR, "cache lookup failed for type %u",
procStruct->proargtypes.values[i]);
}
typeStruct = (Form_pg_type) GETSTRUCT(typeTup);
/* Disallow pseudotype argument */
if (typeStruct->typtype == 'p')
{
free(prodesc->proname);
free(prodesc);
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("plperl functions cannot take type %s",
format_type_be(procStruct->proargtypes.values[i]))));
}
if (typeStruct->typtype == 'c')
prodesc->arg_is_rowtype[i] = true;
else
{
prodesc->arg_is_rowtype[i] = false;
perm_fmgr_info(typeStruct->typoutput,
&(prodesc->arg_out_func[i]));
}
ReleaseSysCache(typeTup);
}
}
/************************************************************
* create the text of the anonymous subroutine.
* we do not use a named subroutine so that we can call directly
* through the reference.
************************************************************/
prosrcdatum = SysCacheGetAttr(PROCOID, procTup,
Anum_pg_proc_prosrc, &isnull);
if (isnull)
elog(ERROR, "null prosrc");
proc_source = DatumGetCString(DirectFunctionCall1(textout,
prosrcdatum));
/************************************************************
* Create the procedure in the interpreter
************************************************************/
prodesc->reference = plperl_create_sub(proc_source, prodesc->lanpltrusted);
pfree(proc_source);
if (!prodesc->reference) /* can this happen? */
{
free(prodesc->proname);
free(prodesc);
elog(ERROR, "could not create internal procedure \"%s\"",
internal_proname);
}
hv_store(plperl_proc_hash, internal_proname, proname_len,
newSViv((IV) prodesc), 0);
}
ReleaseSysCache(procTup);
return prodesc;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/* Build a hash from all attributes of a given tuple. */
22 years ago
static SV *
plperl_hash_from_tuple(HeapTuple tuple, TupleDesc tupdesc)
{
HV *hv;
int i;
hv = newHV();
for (i = 0; i < tupdesc->natts; i++)
{
Datum attr;
bool isnull;
char *attname;
char *outputstr;
Oid typoutput;
bool typisvarlena;
int namelen;
SV *sv;
if (tupdesc->attrs[i]->attisdropped)
continue;
attname = NameStr(tupdesc->attrs[i]->attname);
namelen = strlen(attname);
attr = heap_getattr(tuple, i + 1, tupdesc, &isnull);
if (isnull)
{
/* Store (attname => undef) and move on. */
hv_store(hv, attname, namelen, newSV(0), 0);
continue;
}
/* XXX should have a way to cache these lookups */
getTypeOutputInfo(tupdesc->attrs[i]->atttypid,
&typoutput, &typisvarlena);
outputstr = DatumGetCString(OidFunctionCall1(typoutput, attr));
sv = newSVpv(outputstr, 0);
#if PERL_BCDVERSION >= 0x5006000L
if (GetDatabaseEncoding() == PG_UTF8)
SvUTF8_on(sv);
#endif
hv_store(hv, attname, namelen, sv, 0);
pfree(outputstr);
}
return newRV_noinc((SV *) hv);
}
HV *
plperl_spi_exec(char *query, int limit)
{
HV *ret_hv;
/*
* Execute the query inside a sub-transaction, so we can cope with errors
* sanely
*/
MemoryContext oldcontext = CurrentMemoryContext;
ResourceOwner oldowner = CurrentResourceOwner;
BeginInternalSubTransaction(NULL);
/* Want to run inside function's memory context */
MemoryContextSwitchTo(oldcontext);
PG_TRY();
{
int spi_rv;
spi_rv = SPI_execute(query, plperl_current_prodesc->fn_readonly,
limit);
ret_hv = plperl_spi_execute_fetch_result(SPI_tuptable, SPI_processed,
spi_rv);
/* Commit the inner transaction, return to outer xact context */
ReleaseCurrentSubTransaction();
MemoryContextSwitchTo(oldcontext);
CurrentResourceOwner = oldowner;
/*
* AtEOSubXact_SPI() should not have popped any SPI context, but just
* in case it did, make sure we remain connected.
*/
SPI_restore_connection();
}
PG_CATCH();
{
ErrorData *edata;
/* Save error info */
MemoryContextSwitchTo(oldcontext);
edata = CopyErrorData();
FlushErrorState();
/* Abort the inner transaction */
RollbackAndReleaseCurrentSubTransaction();
MemoryContextSwitchTo(oldcontext);
CurrentResourceOwner = oldowner;
/*
* If AtEOSubXact_SPI() popped any SPI context of the subxact, it will
* have left us in a disconnected state. We need this hack to return
* to connected state.
*/
SPI_restore_connection();
/* Punt the error to Perl */
croak("%s", edata->message);
/* Can't get here, but keep compiler quiet */
return NULL;
}
PG_END_TRY();
return ret_hv;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
static HV *
plperl_spi_execute_fetch_result(SPITupleTable *tuptable, int processed,
int status)
{
HV *result;
result = newHV();
hv_store(result, "status", strlen("status"),
newSVpv((char *) SPI_result_code_string(status), 0), 0);
hv_store(result, "processed", strlen("processed"),
newSViv(processed), 0);
if (status == SPI_OK_SELECT)
{
AV *rows;
SV *row;
int i;
rows = newAV();
for (i = 0; i < processed; i++)
{
row = plperl_hash_from_tuple(tuptable->vals[i], tuptable->tupdesc);
av_push(rows, row);
}
hv_store(result, "rows", strlen("rows"),
newRV_noinc((SV *) rows), 0);
}
SPI_freetuptable(tuptable);
return result;
}
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
/*
* Note: plperl_return_next is called both in Postgres and Perl contexts.
* We report any errors in Postgres fashion (via ereport). If called in
* Perl context, it is SPI.xs's responsibility to catch the error and
* convert to a Perl error. We assume (perhaps without adequate justification)
* that we need not abort the current transaction if the Perl code traps the
* error.
*/
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
void
plperl_return_next(SV *sv)
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
{
plperl_proc_desc *prodesc = plperl_current_prodesc;
FunctionCallInfo fcinfo = plperl_current_caller_info;
ReturnSetInfo *rsi = (ReturnSetInfo *) fcinfo->resultinfo;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
MemoryContext cxt;
HeapTuple tuple;
TupleDesc tupdesc;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
if (!sv)
return;
if (!prodesc->fn_retisset)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("cannot use return_next in a non-SETOF function")));
if (prodesc->fn_retistuple &&
!(SvOK(sv) && SvTYPE(sv) == SVt_RV && SvTYPE(SvRV(sv)) == SVt_PVHV))
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
errmsg("setof-composite-returning Perl function "
"must call return_next with reference to hash")));
cxt = MemoryContextSwitchTo(rsi->econtext->ecxt_per_query_memory);
if (!plperl_current_tuple_store)
plperl_current_tuple_store =
tuplestore_begin_heap(true, false, work_mem);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
if (prodesc->fn_retistuple)
{
TypeFuncClass rettype;
AttInMetadata *attinmeta;
rettype = get_call_result_type(fcinfo, NULL, &tupdesc);
tupdesc = CreateTupleDescCopy(tupdesc);
attinmeta = TupleDescGetAttInMetadata(tupdesc);
tuple = plperl_build_tuple_result((HV *) SvRV(sv), attinmeta);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
}
else
{
Datum ret;
bool isNull;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
tupdesc = CreateTupleDescCopy(rsi->expectedDesc);
if (SvOK(sv) && SvTYPE(sv) != SVt_NULL)
{
char *val = SvPV(sv, PL_na);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
ret = FunctionCall3(&prodesc->result_in_func,
PointerGetDatum(val),
ObjectIdGetDatum(prodesc->result_typioparam),
Int32GetDatum(-1));
isNull = false;
}
else
{
ret = (Datum) 0;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
isNull = true;
}
tuple = heap_form_tuple(tupdesc, &ret, &isNull);
}
if (!plperl_current_tuple_desc)
plperl_current_tuple_desc = tupdesc;
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
tuplestore_puttuple(plperl_current_tuple_store, tuple);
At 2005-05-21 20:18:50 +0530, ams@oryx.com wrote: > > > The second issue is where plperl returns a large result set. I have attached the following seven patches to address this problem: 1. Trivial. Replaces some errant spaces with tabs. 2. Trivial. Fixes the spelling of Jan's name, and gets rid of many inane, useless, annoying, and often misleading comments. Here's a sample: "plperl_init_all() - Initialize all". (I have tried to add some useful comments here and there, and will continue to do so now and again.) 3. Trivial. Splits up some long lines. 4. Converts SRFs in PL/Perl to use a Tuplestore and SFRM_Materialize to return the result set, based on the PL/PgSQL model. There are two major consequences: result sets will spill to disk when they can no longer fit in work_mem; and "select foo_srf()" no longer works. (I didn't lose sleep over the latter, since that form is not valid in PL/PgSQL, and it's not documented in PL/Perl.) 5. Trivial, but important. Fixes use of "undef" instead of undef. This would cause empty functions to fail in bizarre ways. I suspect that there's still another (old) bug here. I'll investigate further. 6. Moves the majority of (4) out into a new plperl_return_next() function, to make it possible to expose the functionality to Perl; cleans up some of the code besides. 7. Add an spi_return_next function for use in Perl code. If you want to apply the patches and try them out, 8-composite.diff is what you should use. (Note: my patches depend upon Andrew's use-strict and %_SHARED patches being applied.) Here's something to try: create or replace function foo() returns setof record as $$ $i = 0; for ("World", "PostgreSQL", "PL/Perl") { spi_return_next({f1=>++$i, f2=>'Hello', f3=>$_}); } return; $$ language plperl; select * from foo() as (f1 integer, f2 text, f3 text); (Many thanks to Andrews Dunstan and Supernews for their help.) Abhijit Menon-Sen
20 years ago
heap_freetuple(tuple);
MemoryContextSwitchTo(cxt);
}
SV *
plperl_spi_query(char *query)
{
SV *cursor;
/*
* Execute the query inside a sub-transaction, so we can cope with errors
* sanely
*/
MemoryContext oldcontext = CurrentMemoryContext;
ResourceOwner oldowner = CurrentResourceOwner;
BeginInternalSubTransaction(NULL);
/* Want to run inside function's memory context */
MemoryContextSwitchTo(oldcontext);
PG_TRY();
{
void *plan;
Portal portal = NULL;
/* Create a cursor for the query */
plan = SPI_prepare(query, 0, NULL);
if (plan)
portal = SPI_cursor_open(NULL, plan, NULL, NULL, false);
if (portal)
cursor = newSVpv(portal->name, 0);
else
cursor = newSV(0);
/* Commit the inner transaction, return to outer xact context */
ReleaseCurrentSubTransaction();
MemoryContextSwitchTo(oldcontext);
CurrentResourceOwner = oldowner;
/*
* AtEOSubXact_SPI() should not have popped any SPI context, but just
* in case it did, make sure we remain connected.
*/
SPI_restore_connection();
}
PG_CATCH();
{
ErrorData *edata;
/* Save error info */
MemoryContextSwitchTo(oldcontext);
edata = CopyErrorData();
FlushErrorState();
/* Abort the inner transaction */
RollbackAndReleaseCurrentSubTransaction();
MemoryContextSwitchTo(oldcontext);
CurrentResourceOwner = oldowner;
/*
* If AtEOSubXact_SPI() popped any SPI context of the subxact, it will
* have left us in a disconnected state. We need this hack to return
* to connected state.
*/
SPI_restore_connection();
/* Punt the error to Perl */
croak("%s", edata->message);
/* Can't get here, but keep compiler quiet */
return NULL;
}
PG_END_TRY();
return cursor;
}
SV *
plperl_spi_fetchrow(char *cursor)
{
SV *row;
/*
* Execute the FETCH inside a sub-transaction, so we can cope with errors
* sanely
*/
MemoryContext oldcontext = CurrentMemoryContext;
ResourceOwner oldowner = CurrentResourceOwner;
BeginInternalSubTransaction(NULL);
/* Want to run inside function's memory context */
MemoryContextSwitchTo(oldcontext);
PG_TRY();
{
Portal p = SPI_cursor_find(cursor);
if (!p)
row = newSV(0);
else
{
SPI_cursor_fetch(p, true, 1);
if (SPI_processed == 0)
{
SPI_cursor_close(p);
row = newSV(0);
}
else
{
row = plperl_hash_from_tuple(SPI_tuptable->vals[0],
SPI_tuptable->tupdesc);
}
SPI_freetuptable(SPI_tuptable);
}
/* Commit the inner transaction, return to outer xact context */
ReleaseCurrentSubTransaction();
MemoryContextSwitchTo(oldcontext);
CurrentResourceOwner = oldowner;
/*
* AtEOSubXact_SPI() should not have popped any SPI context, but just
* in case it did, make sure we remain connected.
*/
SPI_restore_connection();
}
PG_CATCH();
{
ErrorData *edata;
/* Save error info */
MemoryContextSwitchTo(oldcontext);
edata = CopyErrorData();
FlushErrorState();
/* Abort the inner transaction */
RollbackAndReleaseCurrentSubTransaction();
MemoryContextSwitchTo(oldcontext);
CurrentResourceOwner = oldowner;
/*
* If AtEOSubXact_SPI() popped any SPI context of the subxact, it will
* have left us in a disconnected state. We need this hack to return
* to connected state.
*/
SPI_restore_connection();
/* Punt the error to Perl */
croak("%s", edata->message);
/* Can't get here, but keep compiler quiet */
return NULL;
}
PG_END_TRY();
return row;
}