Re: pg_upgrade

Brian Hirt wrote:
> It looks like it's related to atol

Yep, I found the use of atol in the pg_upgrade code too.  Working on a
patch now.

---------------------------------------------------------------------------


> $ cat test-atol.c
> #include <stdlib.h>
> #include <stdio.h>
>
> int
> main(int argc, char **argv)
> {
>   unsigned int test1;
>   long test2;
>   long long test3;
>   unsigned int test4;
>
>   test1 = (unsigned int)atol("3000767169");
>   test2 = (long)atol("3000767169");
>   test3 = atoll("3000767169");
>   test4 = (unsigned int)atoll("3000767169");
>
>   fprintf(stderr,"%u %ld %lld %u\n",test1,test2,test3,test4);
> }
>
> $ make test-atol
> cc     test-atol.c   -o test-atol
> $ ./test-atol
> 2147483647 2147483647 3000767169 3000767169
>
>
> I think C90 and C99 specify different behaviors with atol
>
> Is there some standard way postgresql parses integer strings?  Maybe that method should be used instead of
duplicatingthe functionality so at least the two behave consistently. 
>
> --brian
>
> On Sep 28, 2010, at 2:00 PM, Bruce Momjian wrote:
>
> > Brian Hirt wrote:
> >> Looks like pg_upgrade is using 32bit oids.  2147483647 is the max signed 32 bit int, but the oids for my tables
areclearly larger than that.  
> >>
> >> == output from pg_upgrade ==
> >> Database: basement84_dev
> >> relname: mit.company: reloid: 2147483647 reltblspace:
> >> relname: mit.company_history: reloid: 2147483647 reltblspace:
> >>
> >> == output from catalog query ==
> >> basement84_dev=# select c.oid,c.relname from pg_catalog.pg_namespace n, pg_catalog.pg_class c where n.oid =
c.relnamespaceand n.nspname = 'mit'; 
> >>    oid     |      relname
> >> ------------+--------------------
> >> 3000767630 | company
> >> 3000767633 | company_history
> >> (22 rows)
> >>
> >
> > Interesting.  Odd it would report the max 32-bit signed int.  I wonder
> > if it somehow is getting set to -1.  I looked briefly at the pg_upgrade
> > code and it appears to put all oids in unsigned ints.
> >
> > --
> >  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
> >  EnterpriseDB                             http://enterprisedb.com
> >
> >  + It's impossible for everything to be true. +
> >
> > --
> > Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
> > To make changes to your subscription:
> > http://www.postgresql.org/mailpref/pgsql-general

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + It's impossible for everything to be true. +