Hudson regress failing on sql-mm-circularstring test

[robe]

I'm sure others have noticed:

Looks like might be a rounding issue possibly caused by our gserialized bbox to double maybe?

--- sql-mm-circularstring_expected	2011-06-13 14:31:49.000000000 -0700
+++ /var/home/buildbot/opt/hudson/jobs/PostGIS-trunk-regression/workspace/results_8.4.5_3.2.2_4.7.0/test_37_out	2011-06-15 17:10:38.000000000 -0700
@@ -108,5 +108,5 @@
 public.circularstring.the_geom_3dz effectively removed.
 public.circularstring.the_geom_3dm effectively removed.
 public.circularstring.the_geom_2d effectively removed.
-POLYGON((220187.382089758 150406.434743975,220187.382089758 150506.717070259,220288.81590723 150506.717070259,220288.81590723 150406.434743975,220187.382089758 150406.434743975))
+POLYGON((220187.382089667 150406.434743975,220187.382089667 150506.71707029,220288.815907205 150506.71707029,220288.815907205 150406.434743975,220187.382089667 150406.434743975))
 npoints_is_null|