[gegl] vector-fill: use RaGaBaA float pixel format

[Øyvind "pippin" Kolås <ok src gnome org>]

commit 0220bbeceb48aa5d7677e54d4178b6e7908b7f85
Author: Øyvind Kolås <pippin gimp org>
Date:   Thu Jul 1 05:10:40 2021 +0200

    vector-fill: use RaGaBaA float pixel format

 operations/external/vector-fill.c     |   5 ++---
 tests/compositions/reference/gegl.png | Bin 105361 -> 110907 bytes
 2 files changed, 2 insertions(+), 3 deletions(-)
---
diff --git a/operations/external/vector-fill.c b/operations/external/vector-fill.c
index d094f86d8..5560afb99 100644
--- a/operations/external/vector-fill.c
+++ b/operations/external/vector-fill.c
@@ -146,7 +146,7 @@ process (GeglOperation       *operation,
   }
   else
   {
-    formats[0]=babl_format ("RGBA float");
+    formats[0]=babl_format ("RaGaBaA float");
   }
 
   if (input)
@@ -311,8 +311,7 @@ gegl_op_class_init (GeglOpClass *klass)
     "name",        "gegl:fill-path",
     "title",       _("Fill Path"),
     "categories",  "render:vector",
-    "reference-hash", "b4bd682b327caa2444c6aa93cd009396",
-    "reference-hashB", "0ab76036217007d53232b39de4bd5cf2",
+    "reference-hash", "92f09df038c01fed0460305d33f4250b",
     "description", _("Renders a filled region"),
     "reference-composition", composition,
     NULL);
diff --git a/tests/compositions/reference/gegl.png b/tests/compositions/reference/gegl.png
index 9f1b14142..366f89ad9 100644
Binary files a/tests/compositions/reference/gegl.png and b/tests/compositions/reference/gegl.png differ