Re: ecpg-timestamp- interval problem V8.2.x - Mailing list pgsql-general

From Tom Lane
Subject Re: ecpg-timestamp- interval problem V8.2.x
Date
Msg-id 1258.1207666923@sss.pgh.pa.us
Whole thread Raw
In response to ecpg-timestamp- interval problem V8.2.x  (paul tilles <Paul.Tilles@noaa.gov>)
List pgsql-general
paul tilles <Paul.Tilles@noaa.gov> writes:
> {
>  timestamp dt1st, dt2nd;
>  interval *int_day, *int_pt;

>  char timebeg[40];
>  char timeend[40];
>  int rc;

>  strcpy(timebeg,"2008-04-01");
>  strcpy(timeend,"2008-04-05");

>  printf("timebeg: %s\n", timebeg);
>  printf("timeend: %s\n", timeend);

>  dt1st=PGTYPEStimestamp_from_asc(timebeg,NULL);
>  dt2nd=PGTYPEStimestamp_from_asc(timeend,NULL);
>  int_day=PGTYPESinterval_from_asc("02 01:02:03", NULL);
>  printf("int_day : %s\n", PGTYPESinterval_to_asc(int_day));

>  rc=PGTYPEStimestamp_sub(&dt2nd, &dt1st, int_pt);
>  printf("rc, date2 - date1 = %d, %s\n", rc, PGTYPESinterval_to_asc(int_pt));
>  exit(0);
> }

Surely this code is using "int_pt" without having first assigned to it?
IOW the interval result is being stored at some unpredictable place that
could get clobbered during the printf.

            regards, tom lane

pgsql-general by date:

Previous
From: paul tilles
Date:
Subject: ecpg-timestamp- interval problem V8.2.x
Next
From: Simone Brunozzi
Date:
Subject: Re: Postgresql Conferences/events in Europe?