gmake check error on HPUX
(Sorry if this is a duplicate.)
I'm installing 7.3.1 on an HPUX 11.0 system. configure and gmake run
fine, but 'gmake SHELL=/bin/ksh check' fails on the random test:
parallel group (16 tests): select_distinct_on portals select_distinct
transactions select_having
select_into random arrays subselect union select_implicit hash_index
case aggregates btree_index join
select_into ... ok
select_distinct ... ok
select_distinct_on ... ok
select_implicit ... ok
select_having ... ok
subselect ... ok
union ... ok
case ... ok
join ... ok
aggregates ... ok
transactions ... ok
random ... failed (ignored)
portals ... ok
arrays ... ok
btree_index ... ok
hash_index ... ok
test privileges ... ok
test misc ... ok
The regression.diffs contains the following:
*** ./expected/random.out Thu Jan 6 00:40:54 2000
--- ./results/random.out Wed Jan 15 09:06:36 2003
***************
*** 31,35 ****
WHERE random NOT BETWEEN 80 AND 120;
random
--------
! (0 rows)--- 31,36 ----
WHERE random NOT BETWEEN 80 AND 120;
random
--------
! 131
! (1 row)Can anyone tell me what this means? How do I fix it? Do I need to fix
it since it says it was ignored? I should mention, perhaps, that this
is the second HP box I'm installing pg onto. On the previous box, gmake
check ran without errors (once Tom and Giles helped me fix some things;
thanks guys!)
Any help would be apprciated.
Petre Scheie
petre.scheie@nextelpartners.com
On Wed, 2003-01-15 at 13:38, Petre Scheie wrote:
I'm installing 7.3.1 on an HPUX 11.0 system. configure and gmake run
fine, but 'gmake SHELL=/bin/ksh check' fails on the random test:
Can anyone tell me what this means?
The "random" test fails randomly (heh, as one might expect). If you see
intermittent failures, it's not something to worry about. If you run the
regression tests again, does "random" still fail?
Cheers,
Neil
--
Neil Conway <neilc@samurai.com> || PGP Key ID: DB3C29FC
Okay, I ran 'gmake SHELL=/bin/ksh check' ten times, and nine of those
times there were no errors; on the one that failed, it was on random
again.The one time it failed was in the middle of the ten times; that
is, it ran with no failures five times, failed on the sixth, and then
the remaining four times had no failures. So...what, if anything, does
this mean? Do I have a problem?
Neil Conway wrote:
Show quoted text
On Wed, 2003-01-15 at 15:16, Petre Scheie wrote:
Yes, I ran the tests again and got the same thing.
Ok, if you run the tests repeatedly (say, 10 times), do they continue to
fail every time? If so, what unexpected output is produced? (i.e. what's
in regression.diffs)The reason I'm skeptical of there being a bug here is that the test is
supposed to take a 1000 tuple table and produce "about" every tenth row
(the function that decides when to include a tuple in the result set
uses a PSRNG). In your case, it's producing about 13% of the table
(rather than between 8% and 12%, as the regression tests expect).Cheers,
Neil
Petre Scheie <petre.scheie@nextelpartners.com> writes:
Okay, I ran 'gmake SHELL=/bin/ksh check' ten times, and nine of those
times there were no errors; on the one that failed, it was on random
again.
This is expected behavior. See
http://www.ca.postgresql.org/users-lounge/docs/7.3/postgres/regress-evaluation.html#AEN23423
regards, tom lane