VICAR: support reading more data types and pixel organization (refs O…

…SGeo#1855)

- Support FORMAT=HALF, DOUB and COMP
- Support big-endian order for integer & floating point values
- Support BIP and BIL organizations
- Ignore binary label records (NBL), and properly skip binary prefixes (NBB)
- Add tests

autotest/cpp/test_gdal.cpp

@@ -1278,7 +1278,7 @@ namespace tut
             ensure_equals( static_cast<int>(sLayout.eInterleaving),
                            static_cast<int>(GDALDataset::RawBinaryLayout::Interleaving::UNKNOWN) );
             ensure_equals( sLayout.eDataType, GDT_Byte );
-            ensure( !sLayout.bLittleEndianOrder );
+            ensure( sLayout.bLittleEndianOrder );
             ensure_equals( sLayout.nImageOffset, 9680U );
             ensure_equals( sLayout.nPixelOffset, 1 );
             ensure_equals( sLayout.nLineOffset, 400 );

autotest/gdrivers/vicar.py

@@ -73,6 +73,26 @@ def test_vicar_1():
     for key in expected_md:
         assert md[key] == expected_md[key]
 
-
 
+read_datatypes_lists = [
+    ('vicar_byte', gdal.GDT_Byte, 129),
+    ('vicar_int16', gdal.GDT_Int16, 129),
+    ('vicar_int32', gdal.GDT_Int32, 129),
+    ('vicar_float32_bsq', gdal.GDT_Float32, 123),
+    ('vicar_float32_bil', gdal.GDT_Float32, 123),
+    ('vicar_float32_bip', gdal.GDT_Float32, 123),
+    ('vicar_float64', gdal.GDT_Float64, 129),
+    ('vicar_cfloat32', gdal.GDT_CFloat32, 148),
+]
 
+@pytest.mark.parametrize(
+    'filename,dt,checksum',
+    read_datatypes_lists,
+    ids=[tup[0] for tup in read_datatypes_lists],
+)
+def test_vicar_read_datatypes(filename, dt, checksum):
+
+    ds = gdal.Open('data/%s.vic' % filename)
+    b = ds.GetRasterBand(1)
+    assert b.DataType == dt
+    assert b.Checksum() == checksum