|From:||Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>|
|To:||Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>|
|Cc:||emre(at)hasegeli(dot)com, Kyotaro HORIGUCHI <horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp>, Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>|
|Subject:||Re: [PATCH] Improve geometric types|
|Views:||Raw Message | Whole Thread | Download mbox|
On 09/26/2018 06:45 PM, Tom Lane wrote:
> Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com> writes:
>> Pushed. Now let's wait for the buildfarm to complain ...
> gaur's not happy, but rather surprisingly, it looks like we're
> mostly OK elsewhere. Do you need me to trace down exactly what's
> going wrong on gaur?
Hmmm, interesting. It seems both failures happen in the chunk that
multiplies paths with points, i.e. essentially point_mul_point. So it
seems most platforms end up with
(0,0) * (-3,4) = (-0, 0)
while gaur apparently thinks it's (0,0). And indeed, that's what the
attached trivial program does - I'd bet if you run it on gaur, it'll
print 0.000000, not -0.000000.
Or you could just try doing
select '(0,0)'::point * '(-3,4)'::point;
If this is what's going on, I'd say the best solution is to make it
produce (0,0) everywhere, so that we don't expect -0.0 anywhere.
We could do that either by adding the == 0.0 check to yet another place,
or to point_construct() directly. Adding it to point_construct() means
we'll pay the price always, but I guess there are few paths where we
know we don't need it. And if we add it to many places it's likely about
as expensive as adding it to point_construct.
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
|Next Message||Tom Lane||2018-09-26 20:21:19||Re: Allowing printf("%m") only where it actually works|
|Previous Message||Andres Freund||2018-09-26 19:48:43||Re: pgbench's expression parsing & negative numbers|