On 07/30/2018 11:57 AM, Emre Hasegeli wrote:
OK, thanks for confirming. I'll get it committed and we'll see what the
animals think soon.
Thank you for fixing this. I wanted to preserve this code but wasn't
sure about the correct place or whether it is still necessary.
There are more places we produce -0. The regression tests have
alternative results to cover them. I have the "float-zero" patch for
this. Although I am not sure if it is a correct fix. I think we
should find the correct fix, and apply it globally to floating point
operations. This can be only enabled for platforms which produce -0,
so the others don't have to pay the price.
Hmmm. It'll be difficult to review such patch without access to a
platform exhibiting such behavior ... IIRC IBM offers free access to
open-source devs, I wonder if that would be a way.
regards
--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
