[postgis-devel] closest_segment_2d corner case (32bit failure)

Sandro Santilli strk at kbt.io
Wed Sep 29 13:54:57 PDT 2021


On Wed, Sep 29, 2021 at 10:46:58PM +0200, Sandro Santilli wrote:
> On Wed, Sep 29, 2021 at 06:29:50PM +0300, Darafei "Komяpa" Praliaskouski wrote:
> > Hi,
> > 
> > Can you build it with -fexcess-precision=standard? If it helps, we need to
> > add it to the 32-bit configure.
> 
> Yes, that flag seems to fix that issue. Exposes others but I'm about
> to fix them.

Well, I've fixed one issue, but this one i'm not sure how we want to
fix. Was the goal of "printing" unit test to show we would control
exact printing of numbers ? Because there are these two failures
which only occur with -fexcess-precision=standard switch:

liblwgeom/cunit# ./cu_tester printing

  1. cu_print.c:456  - ASSERT_STRING_EQUAL
        * Expected: 1e-8
        * Obtained: 0.00000001
  2. cu_print.c:457  - ASSERT_STRING_EQUAL
        * Expected: -1e-8
        * Obtained: -0.00000001

--strk;


More information about the postgis-devel mailing list