JC> Just to add to this: most common C++ compilers also have front ends that
JC> handle C, but in every case of which I'm aware, the same code generator
JC> is used for both.  Some versions of GCC also handle Objective-C input,
JC> again using the same back-end for all three languages.
    If I recall correctly, one of the c++ compilers I've had the misfortune
    to use just uses a preproccessor to convert c++ to c to compile it.
    My guess is that is just does the syntax/symantec error checking unique
    to c++ and then uses the c compiler to check the rest (very slow).
JC> Likewise, at least at one time, MS' C, Pascal and FORTRAN compilers all
JC> used the same back-ends.  At roughly the same time, MS shipped the
JC> compilers for both DOS and Xenix.  In this case, you could have the same
JC> front-end if you were compiling C for either system, but a one code
JC> generator for DOS, and a different one for Xenix.  (Actually, even the
JC> code generator wasn't a lot different, though the standard library was.)
    The scanner/parser/intermidiate code generator  can be used for any
    machine: no machine dependant code. Then there was YACC (yet another
    compiler compiler). It (given a syntax) would auto-generate a compiler
    to produce a compiler. The scanner portion uses regular expressions to
    define variable, keyword, constants, ect.
    EG:
         letter [a-zA-Z]
         char   [a-zA-Z_0-9]
    
         variable (letter)(char)*
    a letter is a - z lower or capitalized.
    a character is a letter or an under score or a number
    a variable must start with a letter and can be followed by
    0 or more instances of a character. example:
    b
    b_
    b1_a
    baaaa2222_222934AAA
    are all legal variables.
--- GEcho 1.00
---------------