Thread

1. Re: Call for platforms (HP-UX)

   Giles Lean <giles@nemeton.com.au> — 2001-04-07T05:24:05Z

   > Okay, here are my results:
   > 
   > Box 1: C180 (2.0 PA8000), HPUX 10.20
   > 
   > Compile with gcc: all tests pass
   > Compile with cc: two lines of diffs in geometry (attached)
   > 
   > Box 2: 715/75 (1.1 PA7100LC), HPUX 10.20
   > 
   > Compile with gcc: all tests pass
   > Compile with cc: all tests pass
   
   I haven't had time to look at this further yet, except to build 7.1RC3
   a couple of times with the HP ANSI C compiler today:
   
   PA-RISC 1.1 code (-Ae +O2 +DAportable):   all tests pass
   PA-RISC 2.0 code (-Ae +O2 +DA2.0 +DS2.0): geometry failures
   
   I'm not sure how interesting these differences are anymore -- is there
   anyone familiar enough with floating point to determine if the results
   are acceptable (although currently unexpected :-) or not?
   
   Regards,
   
   Giles
   
   
   
   
   
   
   

2. Re: Call for platforms (HP-UX)

   Tom Lane <tgl@sss.pgh.pa.us> — 2001-04-07T05:32:15Z

   Giles Lean <giles@nemeton.com.au> writes:
   > I'm not sure how interesting these differences are anymore -- is there
   > anyone familiar enough with floating point to determine if the results
   > are acceptable (although currently unexpected :-) or not?
   
   Differences in the last couple of decimal places in the geometry test
   are definitely not a cause for worry.  Although we've tried to create
   exact-match reference files for the most popular platforms, I think
   that's largely an exercise in time-wasting.  Eventually we will figure
   out a way to make the geometry output round off a few digits, and then
   the cross-platform differences should mostly vanish.
   
   			regards, tom lane
   
   

3. Call for platforms (Solaris)

   Mathijs Brands <mathijs@ilse.nl> — 2001-04-07T05:54:16Z

   Hi
   
   I've been running RC3 regression tests, starting with a FreeBSD 4.2-STABLE
   and a Solaris 7 Sparc box. Both tests ran without any problems. I tried
   Solaris 8 Sparc next: it still suffered from the same unix socket problems.
   I had a look at the code and it seems to me that the use of unix sockets
   in RC3 is still enabled, even though it (appearently) doesn't work reliably
   on Solaris.
   
   Since it was rather strange that RC3 did work correctly on Solaris 7 but
   not 8, I also ran regression tests on another Solaris 7 and another 8 box,
   with the same results. Since I still didn't trust it, I also ran RC1 again
   on both Solaris 7 and 8; same result. And now things start getting weird.
   A little more than a week ago the RC1 regression tests ran with on average
   10-15 tests randomly failing. Now, however, I can run the regression several
   times without any test failing. But if I run the regression test enough
   times (4-6 times), I do have tests that fail (about 2-5). The configuration
   of these servers hasn't changed in the last months and I used the same RC1
   source and binaries.
   
   Can somebody confirm whether pgsql Solaris does or does not work correctly
   out-of-the-box? Disabling unix sockets will probably fix all these problems,
   so I'm naturally wondering whether unix socket will or will not be disabled
   in pgsql 7.1...
   
   Regards,
   
   Mathijs
   
   Ps. Vince, could you remove test results 46 and 47? I don't trust them
   anymore.
   -- 
   "A book is a fragile creature.  It suffers the wear of time,
    it fears rodents, the elements, clumsy hands." 
           Umberto Eco