pg.ddx.io pgsql-sql@postgresql.org mailing list archive
help / color / mirror / Atom feedreplicable problem with PL/Perl
5+ messages / 2 participants
[nested] [flat]
* replicable problem with PL/Perl
@ 2003-04-19 06:46 Rajesh Kumar Mallah <mallah@trade-india.com>
2003-04-19 16:34 ` Re: replicable problem with PL/Perl Tom Lane <tgl@sss.pgh.pa.us>
2003-04-20 22:29 ` Re: replicable problem with PL/Perl Tom Lane <tgl@sss.pgh.pa.us>
0 siblings, 2 replies; 5+ messages in thread
From: Rajesh Kumar Mallah @ 2003-04-19 06:46 UTC (permalink / raw)
To: pgsql-sql
Hi this strage phenomenon is happening with my PL/Perl
if one pl/perl function raises an ERROR even other pl/perl
functions stop working and keep rasing the previous error
Regds
Mallah.
tradein_clients=# UPDATE companies set xml = utils.print_file('/usr/local/perlapache/htdocs/templates/hp/data/company_' || company_id || '.xml');
UPDATE 713
tradein_clients=#
tradein_clients=# UPDATE companies set xml = utils.print_file('/usr/local/perlapache/htdocs/templates/hp/data/company_' || company_id || '.xml');
UPDATE 713
tradein_clients=# SELECT company_id, utils.pgxml_xpath_pl('/Company@CompanyId' , xml) from companies where xml is not null and company_id=619;
INFO: pgxml_xpath_pl called
ERROR: There was an error:
mismatched tag at line 45, column 867, byte 2849 at /opt/perl/lib/site_perl/5.8.0/i686-linux/XML/Parser.pm line 185
tradein_clients=# UPDATE companies set xml = utils.print_file('/usr/local/perlapache/htdocs/templates/hp/data/company_' || company_id || '.xml');
ERROR: plperl: error from function:
mismatched tag at line 45, column 867, byte 2849 at /opt/perl/lib/site_perl/5.8.0/i686-linux/XML/Parser.pm line 185
This query was working previously and utils.print_file is independent of utils.pgxml_xpath_pl
tradein_clients=#
--
Rajesh Kumar Mallah,
Project Manager (Development)
Infocom Network Limited, New Delhi
phone: +91(11)6152172 (221) (L) ,9811255597 (M)
Visit http://www.trade-india.com ,
India's Leading B2B eMarketplace.
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: replicable problem with PL/Perl
2003-04-19 06:46 replicable problem with PL/Perl Rajesh Kumar Mallah <mallah@trade-india.com>
@ 2003-04-19 16:34 ` Tom Lane <tgl@sss.pgh.pa.us>
2003-04-19 16:42 ` Re: replicable problem with PL/Perl Rajesh Kumar Mallah <mallah@trade-india.com>
1 sibling, 1 reply; 5+ messages in thread
From: Tom Lane @ 2003-04-19 16:34 UTC (permalink / raw)
To: Rajesh Kumar Mallah <mallah@trade-india.com>; +Cc: pgsql-sql
Rajesh Kumar Mallah <mallah@trade-india.com> writes:
> if one pl/perl function raises an ERROR even other pl/perl
> functions stop working and keep rasing the previous error
This was reported before:
http://archives.postgresql.org/pgsql-hackers/2002-10/msg00170.php
I'm not convinced that John's proposed fix is correct though
(see followup discussion). Someone should bring up the issue
with some Perl gurus and see whether clearing $@ is really an
appropriate thing to do.
regards, tom lane
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: replicable problem with PL/Perl
2003-04-19 06:46 replicable problem with PL/Perl Rajesh Kumar Mallah <mallah@trade-india.com>
2003-04-19 16:34 ` Re: replicable problem with PL/Perl Tom Lane <tgl@sss.pgh.pa.us>
@ 2003-04-19 16:42 ` Rajesh Kumar Mallah <mallah@trade-india.com>
0 siblings, 0 replies; 5+ messages in thread
From: Rajesh Kumar Mallah @ 2003-04-19 16:42 UTC (permalink / raw)
To: Tom Lane <tgl@sss.pgh.pa.us>; +Cc: pgsql-sql
Thanks Tom,
and sorry for the repost :)
regds
mallah.
On Saturday 19 Apr 2003 10:04 pm, Tom Lane wrote:
> Rajesh Kumar Mallah <mallah@trade-india.com> writes:
> > if one pl/perl function raises an ERROR even other pl/perl
> > functions stop working and keep rasing the previous error
>
> This was reported before:
> http://archives.postgresql.org/pgsql-hackers/2002-10/msg00170.php
>
> I'm not convinced that John's proposed fix is correct though
> (see followup discussion). Someone should bring up the issue
> with some Perl gurus and see whether clearing $@ is really an
> appropriate thing to do.
>
> regards, tom lane
--
Rajesh Kumar Mallah,
Project Manager (Development)
Infocom Network Limited, New Delhi
phone: +91(11)6152172 (221) (L) ,9811255597 (M)
Visit http://www.trade-india.com ,
India's Leading B2B eMarketplace.
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: replicable problem with PL/Perl
2003-04-19 06:46 replicable problem with PL/Perl Rajesh Kumar Mallah <mallah@trade-india.com>
@ 2003-04-20 22:29 ` Tom Lane <tgl@sss.pgh.pa.us>
1 sibling, 0 replies; 5+ messages in thread
From: Tom Lane @ 2003-04-20 22:29 UTC (permalink / raw)
To: Rajesh Kumar Mallah <mallah@trade-india.com>; +Cc: John Worsley <lx@openvein.com>; pgsql-sql
Rajesh Kumar Mallah <mallah@trade-india.com> writes:
> if one pl/perl function raises an ERROR even other pl/perl
> functions stop working and keep rasing the previous error
> [ similar report last fall from John Worsley ]
I have applied the attached patch to fix these problems. (It's against
7.3 or current, but should be adaptable to 7.2 if you need to.) The
key point is to get rid of G_KEEPERR in the perl_call_sv function;
but it turns out that the setjmp/longjmp logic is broken too. Since the
latter wasn't doing anything particularly useful, I just diked it out.
regards, tom lane
inline patch 1: src/pl/plperl/plperl.c (+32/-59 lines) download inline-1.patch
*** src/pl/plperl/plperl.c.orig Sat Sep 21 14:39:26 2002
--- src/pl/plperl/plperl.c Sun Apr 20 17:15:34 2003
***************
*** 92,99 ****
* Global data
**********************************************************************/
static int plperl_firstcall = 1;
- static int plperl_call_level = 0;
- static int plperl_restart_in_progress = 0;
static PerlInterpreter *plperl_interp = NULL;
static HV *plperl_proc_hash = NULL;
--- 92,97 ----
***************
*** 143,148 ****
--- 141,155 ----
if (!plperl_firstcall)
return;
+ /************************************************************
+ * Free the proc hash table
+ ************************************************************/
+ if (plperl_proc_hash != NULL)
+ {
+ hv_undef(plperl_proc_hash);
+ SvREFCNT_dec((SV *) plperl_proc_hash);
+ plperl_proc_hash = NULL;
+ }
/************************************************************
* Destroy the existing Perl interpreter
***************
*** 155,170 ****
}
/************************************************************
- * Free the proc hash table
- ************************************************************/
- if (plperl_proc_hash != NULL)
- {
- hv_undef(plperl_proc_hash);
- SvREFCNT_dec((SV *) plperl_proc_hash);
- plperl_proc_hash = NULL;
- }
-
- /************************************************************
* Now recreate a new Perl interpreter
************************************************************/
plperl_init_interp();
--- 162,167 ----
***************
*** 202,209 ****
perl_parse(plperl_interp, plperl_init_shared_libs, 3, embedding, NULL);
perl_run(plperl_interp);
-
-
/************************************************************
* Initialize the proc and query hash tables
************************************************************/
--- 199,204 ----
***************
*** 212,218 ****
}
-
/**********************************************************************
* plperl_call_handler - This is the only visible function
* of the PL interpreter. The PostgreSQL
--- 207,212 ----
***************
*** 229,235 ****
Datum retval;
/************************************************************
! * Initialize interpreters on first call
************************************************************/
if (plperl_firstcall)
plperl_init_all();
--- 223,229 ----
Datum retval;
/************************************************************
! * Initialize interpreter on first call
************************************************************/
if (plperl_firstcall)
plperl_init_all();
***************
*** 239,248 ****
************************************************************/
if (SPI_connect() != SPI_OK_CONNECT)
elog(ERROR, "plperl: cannot connect to SPI manager");
- /************************************************************
- * Keep track about the nesting of Perl-SPI-Perl-... calls
- ************************************************************/
- plperl_call_level++;
/************************************************************
* Determine if called as function or trigger and
--- 233,238 ----
***************
*** 261,268 ****
else
retval = plperl_func_handler(fcinfo);
- plperl_call_level--;
-
return retval;
}
--- 251,256 ----
***************
*** 272,284 ****
* create the anonymous subroutine whose text is in the SV.
* Returns the SV containing the RV to the closure.
**********************************************************************/
! static
! SV *
plperl_create_sub(char *s, bool trusted)
{
dSP;
!
! SV *subref = NULL;
int count;
ENTER;
--- 260,270 ----
* create the anonymous subroutine whose text is in the SV.
* Returns the SV containing the RV to the closure.
**********************************************************************/
! static SV *
plperl_create_sub(char *s, bool trusted)
{
dSP;
! SV *subref;
int count;
ENTER;
***************
*** 286,295 ****
--- 272,294 ----
PUSHMARK(SP);
XPUSHs(sv_2mortal(newSVpv(s, 0)));
PUTBACK;
+ /*
+ * 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?
+ */
count = perl_call_pv((trusted ? "mksafefunc" : "mkunsafefunc"),
G_SCALAR | G_EVAL | G_KEEPERR);
SPAGAIN;
+ if (count != 1)
+ {
+ PUTBACK;
+ FREETMPS;
+ LEAVE;
+ elog(ERROR, "plperl: didn't get a return item from mksafefunc");
+ }
+
if (SvTRUE(ERRSV))
{
POPs;
***************
*** 299,307 ****
elog(ERROR, "creation of function failed: %s", SvPV(ERRSV, PL_na));
}
- if (count != 1)
- elog(ERROR, "creation of function failed - no return from mksafefunc");
-
/*
* need to make a deep copy of the return. it comes off the stack as a
* temporary.
--- 298,303 ----
***************
*** 324,329 ****
--- 320,326 ----
PUTBACK;
FREETMPS;
LEAVE;
+
return subref;
}
***************
*** 352,372 ****
* plperl_call_perl_func() - calls a perl function through the RV
* stored in the prodesc structure. massages the input parms properly
**********************************************************************/
! static
! SV *
plperl_call_perl_func(plperl_proc_desc * desc, FunctionCallInfo fcinfo)
{
dSP;
-
SV *retval;
int i;
int count;
-
ENTER;
SAVETMPS;
! PUSHMARK(sp);
for (i = 0; i < desc->nargs; i++)
{
if (desc->arg_is_rel[i])
--- 349,366 ----
* plperl_call_perl_func() - calls a perl function through the RV
* stored in the prodesc structure. massages the input parms properly
**********************************************************************/
! static SV *
plperl_call_perl_func(plperl_proc_desc * desc, FunctionCallInfo fcinfo)
{
dSP;
SV *retval;
int i;
int count;
ENTER;
SAVETMPS;
! PUSHMARK(SP);
for (i = 0; i < desc->nargs; i++)
{
if (desc->arg_is_rel[i])
***************
*** 401,407 ****
}
}
PUTBACK;
! count = perl_call_sv(desc->reference, G_SCALAR | G_EVAL | G_KEEPERR);
SPAGAIN;
--- 395,403 ----
}
}
PUTBACK;
!
! /* Do NOT use G_KEEPERR here */
! count = perl_call_sv(desc->reference, G_SCALAR | G_EVAL);
SPAGAIN;
***************
*** 424,439 ****
retval = newSVsv(POPs);
-
PUTBACK;
FREETMPS;
LEAVE;
return retval;
-
-
}
/**********************************************************************
* plperl_func_handler() - Handler for regular function calls
**********************************************************************/
--- 420,433 ----
retval = newSVsv(POPs);
PUTBACK;
FREETMPS;
LEAVE;
return retval;
}
+
/**********************************************************************
* plperl_func_handler() - Handler for regular function calls
**********************************************************************/
***************
*** 443,465 ****
plperl_proc_desc *prodesc;
SV *perlret;
Datum retval;
- sigjmp_buf save_restart;
/* Find or compile the function */
prodesc = compile_plperl_function(fcinfo->flinfo->fn_oid, false);
- /* Set up error handling */
- memcpy(&save_restart, &Warn_restart, sizeof(save_restart));
-
- if (sigsetjmp(Warn_restart, 1) != 0)
- {
- memcpy(&Warn_restart, &save_restart, sizeof(Warn_restart));
- plperl_restart_in_progress = 1;
- if (--plperl_call_level == 0)
- plperl_restart_in_progress = 0;
- siglongjmp(Warn_restart, 1);
- }
-
/************************************************************
* Call the Perl function
************************************************************/
--- 437,446 ----
***************
*** 490,503 ****
SvREFCNT_dec(perlret);
- memcpy(&Warn_restart, &save_restart, sizeof(Warn_restart));
- if (plperl_restart_in_progress)
- {
- if (--plperl_call_level == 0)
- plperl_restart_in_progress = 0;
- siglongjmp(Warn_restart, 1);
- }
-
return retval;
}
--- 471,476 ----
* [REPOST] replicable problem with PL/Perl
@ 2003-04-19 16:06 Rajesh Kumar Mallah <mallah@trade-india.com>
0 siblings, 0 replies; 5+ messages in thread
From: Rajesh Kumar Mallah @ 2003-04-19 16:06 UTC (permalink / raw)
To: pgsql-sql
Hi this strage phenomenon is happening with my PL/Perl
if one pl/perl function raises an ERROR even other pl/perl
functions stop working and keep rasing the previous error
Regds
Mallah.
tradein_clients=# UPDATE companies set xml = utils.print_file('/usr/local/perlapache/htdocs/templates/hp/data/company_' || company_id || '.xml');
UPDATE 713
tradein_clients=#
tradein_clients=# UPDATE companies set xml = utils.print_file('/usr/local/perlapache/htdocs/templates/hp/data/company_' || company_id || '.xml');
UPDATE 713
tradein_clients=# SELECT company_id, utils.pgxml_xpath_pl('/Company@CompanyId' , xml) from companies where xml is not null and company_id=619;
INFO: pgxml_xpath_pl called
ERROR: There was an error:
mismatched tag at line 45, column 867, byte 2849 at /opt/perl/lib/site_perl/5.8.0/i686-linux/XML/Parser.pm line 185
tradein_clients=# UPDATE companies set xml = utils.print_file('/usr/local/perlapache/htdocs/templates/hp/data/company_' || company_id || '.xml');
ERROR: plperl: error from function:
mismatched tag at line 45, column 867, byte 2849 at /opt/perl/lib/site_perl/5.8.0/i686-linux/XML/Parser.pm line 185
This query was working previously and utils.print_file is independent of utils.pgxml_xpath_pl
tradein_clients=#
--
Rajesh Kumar Mallah,
Project Manager (Development)
Infocom Network Limited, New Delhi
phone: +91(11)6152172 (221) (L) ,9811255597 (M)
Visit http://www.trade-india.com ,
India's Leading B2B eMarketplace.
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
^ permalink raw reply [nested|flat] 5+ messages in thread
end of thread, other threads:[~2003-04-20 22:29 UTC | newest]
Thread overview: 5+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2003-04-19 06:46 replicable problem with PL/Perl Rajesh Kumar Mallah <mallah@trade-india.com>
2003-04-19 16:34 ` Tom Lane <tgl@sss.pgh.pa.us>
2003-04-19 16:42 ` Rajesh Kumar Mallah <mallah@trade-india.com>
2003-04-20 22:29 ` Tom Lane <tgl@sss.pgh.pa.us>
2003-04-19 16:06 [REPOST] replicable problem with PL/Perl Rajesh Kumar Mallah <mallah@trade-india.com>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox