Just a data point on the geometry test under NetBSD/i386 issue:
/etc/ld.so.conf by default now contains:
libm.so.0 machdep.fpu_present 1:libm387.so.0,libm.so.0
which means that if the sysctl machdep.fpu_present returns 1, load the
shared library libm387 to make use of the fpu.
If you remove /etc/ld.so.conf, so that ldd `which psql` does not show
-lm.0 => /usr/lib/libm387.so.0 -lm.0 => /usr/lib/libm.so.0
but only the libm.so.0 line
======================All 76 tests passed.
======================
If you replace the /etc/ld.so.conf file and have an fpu, then the geometry
test will fail with slightly different rounding.
Do we want a specific geometry-netbsd-i386-with-fpu.out where you must
also test
% sysctl machdep.fpu_present
machdep.fpu_present = 1
?
Cheers,
Patrick
PS: AFAIK geometry-positive-zeros-bsd works for all NetBSD platforms - the
above difference is only for i386 + fpu.
On Thu, Mar 22, 2001 at 07:12:39PM +0200, Marko Kreen wrote:
> On Thu, Mar 22, 2001 at 05:25:01PM +0100, Peter Eisentraut wrote:
> > Marko Kreen writes:
> > >
> > > OK?: NetBSD 1.5 i586 / egcs 2.91.66 / (netbsd-1-5 from Jan)
> > >
> > > netbsd FAILED the geometry test, diff attached, dunno if its
> > > critical or not.
> >
> > Can you check whether it matches any of the other possible geometry
> > results? See
>
> Yes, it matches geometry-positive-zeros-bsd.out. There is
> another report about NetBSD 1.5/i386 which has comment:
>
> > one spurious floating point test failure
> > (mail sent to postgresql-bugs with details)
>
> But I could not find it in archive page. (reporter Giles Lean
> <giles@nemeton.com.au>) Perhaps same thing?
>
> --
> marko
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 2: you can get off all lists at once with the unregister command
> (send "unregister YourEmailAddressHere" to majordomo@postgresql.org)