Patch a case where nodes with very close to zero values created an in…

…finite alloc loop. (#8388)

The situation arose when the condition

if ((dz = z[this_side] - z[this_side+1]) == 0.0f) continue;

was never met by a tiny bit. Comparing to 0.0 is never a good idea. So I made it

if (fabs((dz = z[this_side] - z[this_side+1])) < 1e-10) continue;

I get 6 grdcontour test failures but they don't seem related to this change (more due to the endless headache of slight different grids).

Fixes #8387

src/gmt_support.c

@@ -2646,11 +2646,10 @@ GMT_LOCAL uint64_t gmtsupport_trace_contour (struct GMT_CTRL *GMT, struct GMT_GR
  n_nan++;
  continue;
  }
-
  /* Skip if no zero-crossing on this edge */
 
  if (z[this_side+1] * z[this_side] > 0.0f) continue;
- if ((dz = z[this_side] - z[this_side+1]) == 0.0f) continue;
+ if (fabs((dz = z[this_side] - z[this_side+1])) < 1e-10) continue;
 
  /* Save normalized distance along edge from corner this_side to crossing of edge this_side */