Thread: Horology test failure while compiling PostgreSQL 8.3.3 on Solaris 10 x64

Horology test failure while compiling PostgreSQL 8.3.3 on Solaris 10 x64

From
"Dot Yet"
Date:
Hi Eveyone,

I am trying to compile PostgreSQL 8.3.3 on Solaris 10 x64. The tools involved were:

OpenSSL 0.9.8h 28 May 2008
PostgreSQL 8.3.3
CC: Sun C++ 5.8 2005/10/13

Configure command:
./configure --prefix=/opt/usr/local/amd64 --with-openssl --with-libraries=/opt/usr/local/amd64/lib --with-includes=/opt/usr/local/amd64/include --libdir=/opt/usr/local/amd64/lib --includedir=/opt/usr/local/amd64/include CFLAGS="-R/opt/usr/local/amd64/lib -xmodel=medium -xtarget=generic64 -fast -xarch=amd64"

The error log snippet is:

#################################################

parallel group (5 tests):  type_sanity horology geometry oidjoins opr_sanity
     geometry             ... ok
     horology             ... FAILED
     oidjoins             ... ok
     type_sanity          ... ok
     opr_sanity           ... ok

============== shutting down postmaster               ==============
server stopped

========================
 1 of 114 tests failed.
========================

The differences that caused some tests to fail can be viewed in the
file "./regression.diffs".  A copy of the test summary that you see
above is saved in the file "./regression.out".

make[2]: *** [check] Error 1
make[2]: Leaving directory `/export/home/pgadmin/postgresql-8.3.3/src/test/regress'
make[1]: *** [check] Error 2
make[1]: Leaving directory `/export/home/pgadmin/postgresql-8.3.3/src/test'
make: *** [check] Error 2

cat regression.diff:

*** ./expected/horology.out     Wed Jul 25 13:22:36 2007
--- ./results/horology.out      Fri Aug 15 10:49:39 2008
***************
*** 2135,2141 ****
  SELECT '' AS ten, f1 AS interval, reltime(f1) AS reltime
    FROM INTERVAL_TBL;
   ten |           interval            |            reltime
! -----+-------------------------------+-------------------------------
       | @ 1 min                       | @ 1 min
       | @ 5 hours                     | @ 5 hours
       | @ 10 days                     | @ 10 days
--- 2135,2141 ----
  SELECT '' AS ten, f1 AS interval, reltime(f1) AS reltime
    FROM INTERVAL_TBL;
   ten |           interval            |             reltime
! -----+-------------------------------+----------------------------------
       | @ 1 min                       | @ 1 min
       | @ 5 hours                     | @ 5 hours
       | @ 10 days                     | @ 10 days
***************
*** 2143,2149 ****
       | @ 3 mons                      | @ 3 mons
       | @ 14 secs ago                 | @ 14 secs ago
       | @ 1 day 2 hours 3 mins 4 secs | @ 1 day 2 hours 3 mins 4 secs
!      | @ 6 years                     | @ 6 years
       | @ 5 mons                      | @ 5 mons
       | @ 5 mons 12 hours             | @ 5 mons 12 hours
  (10 rows)
--- 2143,2149 ----
       | @ 3 mons                      | @ 3 mons
       | @ 14 secs ago                 | @ 14 secs ago
       | @ 1 day 2 hours 3 mins 4 secs | @ 1 day 2 hours 3 mins 4 secs
!      | @ 6 years                     | @ 5 years 12 mons 5 days 6 hours
       | @ 5 mons                      | @ 5 mons
       | @ 5 mons 12 hours             | @ 5 mons 12 hours
  (10 rows)

======================================================================




###############################################

Any idea what maybe causing this issue?

Thanks in advance,
dotyet.
"Dot Yet" <dot.yet@gmail.com> writes:
> Configure command:
> ./configure --prefix=/opt/usr/local/amd64 --with-openssl
> --with-libraries=/opt/usr/local/amd64/lib
> --with-includes=/opt/usr/local/amd64/include
> --libdir=/opt/usr/local/amd64/lib --includedir=/opt/usr/local/amd64/include
> CFLAGS="-R/opt/usr/local/amd64/lib -xmodel=medium -xtarget=generic64 -fast
> -xarch=amd64"

What does "-fast" do?  If it involves any reinterpretation of IEEE float
arithmetic accuracy requirements, drop it.

            regards, tom lane

Re: Horology test failure while compiling PostgreSQL 8.3.3 on Solaris 10 x64

From
"F. Jovan Jester"
Date:
>> CFLAGS="-R/opt/usr/local/amd64/lib -xmodel=medium -
>> xtarget=generic64 -fast
>> -xarch=amd64"
>
> What does "-fast" do?  If it involves any reinterpretation of IEEE
> float
> arithmetic accuracy requirements, drop it.
>

The -fast option using sun studio cc expands to include the flag -
fsimple=2 which does affect floating point calculations. If you remove
the -fast flag it should pass the tests.

any special reason why you are using xtarget=generic64 instead of
native64?

-jovan