From: Sandro Santilli <strk@keybit.net>
Date: Wed, 11 Jan 2012 11:17:03 +0000 (+0000)
Subject: Unlease higher dimensions in ST_AsBinary, using SQL/MM (#288)
X-Git-Tag: 2.0.0alpha1~111
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=d2b342f3a61dad1cda37176c4f6b12445e024bbd;p=postgis

Unlease higher dimensions in ST_AsBinary, using SQL/MM (#288)

git-svn-id: http://svn.osgeo.org/postgis/trunk@8769 b70326c6-7e19-0410-871a-916f4a2858ee
---

diff --git a/NEWS b/NEWS
index 9782e6b3a..c36d1b76d 100644
--- a/NEWS
+++ b/NEWS
@@ -44,6 +44,8 @@ PostGIS 2.0.0
   - #817 Renaming old 3D functions to the convention ST_3D (Nicklas Avén) 
   - #548 (sorta), ST_NumGeometries,ST_GeometryN now returns 1 (or the geometry) 
     instead of null for single geometries (Sandro Santilli, Maxime van Noppen)
+  - #287, #288 ST_AsText and ST_AsBinary don't force 2d anymore, using SQL/MM
+    notation for higher dimensions
 
   * New Features *
   
@@ -103,7 +105,6 @@ PostGIS 2.0.0
     (Regina Obe, Leo Hsu / Paragon Corporation / 
     funding provided by Hunter Systems Group)
   - Overall Documentation proofreading and corrections. (Kasif Rasul)
-  - ST_AsText support for higher dimensions (#287)
   
  * Bug Fixes *
   - #1335 ST_AddPoint returns incorrect result on Linux (Even Rouault)
diff --git a/postgis/lwgeom_ogc.c b/postgis/lwgeom_ogc.c
index 2fee53616..9fec072b8 100644
--- a/postgis/lwgeom_ogc.c
+++ b/postgis/lwgeom_ogc.c
@@ -850,21 +850,12 @@ Datum LWGEOM_asBinary(PG_FUNCTION_ARGS)
 	uint8_t *wkb;
 	size_t wkb_size;
 	bytea *result;
-	/* By default we are currently emitting OGC WKB (2D) only */
-	uint8_t variant = WKB_SFSQL;
+	uint8_t variant = WKB_ISO;
 
 	/* Get a 2D version of the geometry */
 	geom = (GSERIALIZED*)PG_DETOAST_DATUM(PG_GETARG_DATUM(0));
 	lwgeom = lwgeom_from_gserialized(geom);
 
-	/* Get a 2D version of the geometry if necessary */
-	if ( lwgeom_ndims(lwgeom) > 2 )
-	{
-		LWGEOM *lwgeom2d = lwgeom_force_2d(lwgeom);
-		lwgeom_free(lwgeom);
-		lwgeom = lwgeom2d;
-	}
-
 	/* If user specified endianness, respect it */
 	if ( (PG_NARGS()>1) && (!PG_ARGISNULL(1)) )
 	{