[postgis-tickets] [PostGIS] #4138: PostGIS 2.5.0beta1 fails tests on MIPS64EL

[PostGIS]

#4138: PostGIS 2.5.0beta1 fails tests on MIPS64EL
----------------------+---------------------------
  Reporter:  komzpa   |      Owner:  pramsey
      Type:  defect   |     Status:  new
  Priority:  medium   |  Milestone:  PostGIS 2.5.0
 Component:  postgis  |    Version:  trunk
Resolution:           |   Keywords:
----------------------+---------------------------

Comment (by pramsey):

 So, looks like MIPS is generating what we'd consider a valid POINT EMPTY
 (it's a WKB point with NaN in the coordinates) but unfortunately it is
 possible to represent NAN a bunch of ways, so we get this issue. I'm
 pretty sure isnan on any platform would agree that is a NAN. Shame it's a
 different bit pattern.

 Fixing this will, I think, involve using the nan() function to try and be
 specific about what gets put into the non-specified parts of the nan, I
 need to experiment a bit and see how much control we really have. And then
 of course, platform to platform how that function actually behaves (or
 exists (or is in what header)) could be different. We've had a lot of
 platform issues with nan/inf, it's the most common place we get platform-
 to-platform inconsistency, I feel.

-- 
Ticket URL: <https://trac.osgeo.org/postgis/ticket/4138#comment:2>
PostGIS <http://trac.osgeo.org/postgis/>
The PostGIS Trac is used for bug, enhancement & task tracking, a user and developer wiki, and a view into the subversion code repository of PostGIS project.