[Boost-users] [PREPROCESSOR] Working around the 256 element limit for sequences

Subject: [Boost-users] [PREPROCESSOR] Working around the 256 element limit for sequences
From: Rush Manbert (rush_at_[hidden])
Date: 2009-10-12 20:12:37

We define exception error codes and other useful information like this:

#define EXCEPTION_TABLE \
( (enumName0, "description string", "another sting", "yet another  
string") ) \
( (enumName1, "description string", "another sting", "yet another  
string") ) \
.
.
.
// End EXCEPTION_TABLE def

Then we write a macro that creates an enum like this:

#define ENUM_ERROR_CODES(EXCEPTION_TABLE_MACRO_NAME,  
firstErrorCodeValue)\
public:																		typedef enum	\
{ \																		\
     BOOST_PP_SEQ_FOR_EACH_I(MAKE_EXCEPTION_CODE_ENUM, ~,  
EXCEPTION_TABLE_MACRO_NAME) \
} ErrorCodesT;		\
// End macro definition

and we invoke ENUM_ERROR_CODES as:

ENUM_EROR_CODEAS(EXCEPTION_TABLE, 1)

This has worked just beautifully until recently. Now we're trying to  
use this code to define an enum (and other things) based on an  
EXCEPTION_TABLE that carries error code for a hardware device. There  
are about 500 of them.

I ran into the 256 element limitation for sequences. I have tried  
rewriting the enum macro like this:
#define ENUM_ERROR_CODES(EXCEPTION_TABLE_MACRO_NAME,  
firstErrorCodeValue)\
public:																		typedef enum	\
{ \																		\
     BOOST_PP_SEQ_FOR_EACH_I(MAKE_EXCEPTION_CODE_ENUM, ~,  
BOOST_PP_SEQ_SUBSEQ(EXCEPTION_TABLE_MACRO_NAME,   0, 250))\
     BOOST_PP_SEQ_FOR_EACH_I(MAKE_EXCEPTION_CODE_ENUM, ~,  
BOOST_PP_SEQ_SUBSEQ(EXCEPTION_TABLE_MACRO_NAME, 250, 250))\
     BOOST_PP_SEQ_FOR_EACH_I(MAKE_EXCEPTION_CODE_ENUM, ~,  
BOOST_PP_SEQ_SUBSEQ(EXCEPTION_TABLE_MACRO_NAME, 500, 250))\
     BOOST_PP_SEQ_FOR_EACH_I(MAKE_EXCEPTION_CODE_ENUM, ~,  
BOOST_PP_SEQ_SUBSEQ(EXCEPTION_TABLE_MACRO_NAME, 750, 250))\

} ErrorCodesT;		\
// End macro definition

thinking that I could just peel off groups of 250, but that doesn't  
work. I can peel off the first N, where N is <= the size of the  
sequence, but I certainly can't do what I wrote.

Each of our exception classes is defined this way, and the number of  
elements in each EXCEPTION_TABLE is different for each.

Is there some standard method for working around this problem?

Thanks,
Rush