CPL_IS_LSB not found in raw/btdataset.cpp under MSYS

[Roger.Bivand@…]

The recent cvs build OK on my prime platform (RH9, i386). I'm trying to find
ways of building the rgdal R package for Windows using MSYS (Win XP, i386), and
until CPL_IS_LSB was introduced to cpl_port.h, everything built. When CPL_IS_LSB
arrived, the code in the single place it is used in raw/btdataset.cpp claims not
to find it. When the defines are commented out in cpl_port.h and added reversed:

#if defined(CPL_LSB)
#  define CPL_IS_LSB 1
#else
#  define CPL_IS_LSB 0
#endif

in raw/btdataset.cpp just above its use, it is found, and GDAL compiles to
completion. I don't know why the MSYS/MinGW build train should fall over on
this. Otherwise current GDAL and PROJ.4 seem to behave well compiled statically
under MSYS, and are linked into the R package DLLs as needed.

[warmerdam]

Roger, 

I can't imagine how this could be necessary, but it is harmless so I have
applied it.  Surely there must have been some issue with cpl_port.h not having
been updated from CVS on your first attempt or something. 

[Roger.Bivand@…]

Runs to completion now with

#ifdef CPL_LSB
#define CPL_IS_LSB 1
#else
#define CPL_IS_LSB 0
#endif

in cpl_port.h.

I don't understand it either - perhaps some other whitespace character? I ran
g++ -E to look for it, but couldn't see what was going on. I'll keep an eye on
it with subsequent updates.

Roger