[Benchmarking] split contour files are different

[Jeff McKenna]

Iván,

In order to display the 8 contour shapefiles efficiently in MapServer, a 
tile index file must be generated.  MapServer requires the files in a 
tile index to have the exact same attribute definitions and they must me 
in the same order.

In our full vector dataset this is not the case: contour-5, contour-6, 
contour-7 files contain different fields.  Here is a sample:

***
contour-4.dbf



ID: Real (31.15)
ELEMENTO: Real (31.15)
ID_SIMPLE: Real (31.15)
MODIFICADO: Real (31.15)
ID_HOJA: Real (31.15)
ID_CODE: String (6.0)
METADATO: Real (31.15)
F_ALTA: String (10.0)
F_BAJA: String (10.0)
ETIQUETA: String (254.0)
ETIQUETAS: Real (31.15)
ETIQUETA1: String (10.0)
SIMPLE: String (250.0)
ETIQUETA0: Real (31.15)
PROYECTO: String (3.0)
LONGITUD: Real (31.15)
PERIMETRO: Real (31.15)
SUPERFICIE: Real (31.15)
TIPO_0201: String (6.0)
CATEG_0201: String (6.0)
ORIGE_0201: String (6.0)
COTA_0201: Real (31.15)
NNECCCPP: String (8.0)
ORIGEN_Z: Real (31.15)

contour-5.dbf

ID: Real (31.15)
ID_SIMPLE: Real (31.15)
MODIFICADO: Real (31.15)
ID_HOJA: Real (11.0)
ID_CODE: String (6.0)
ID_MET: Real (31.15)
F_ALTA: String (10.0)
F_BAJA: String (10.0)
ETIQUETA: String (254.0)
ID_COD_ETI: String (10.0)
PROYECTO: String (3.0)
TIPO_0201: String (2.0)
CATEG_0201: String (2.0)
ORIGE_0201: String (2.0)
COTA_0201: Real (31.15)
NNECCCPP: String (8.0)
ORIGEN_Z: Real (31.15)


The contour files must be re-processed so that they contain the same 
attributes, and then provided to all benchmarking teams.

-jeff