A macro is composed of two parts: a macro-name and a definition. A macro name is an alphanumeric character string starting with an alphabetic character and terminating with a colon. A macro name must not exceed 20 characters in length and may only contain alphanumeric characters (a-z, 0-9).

A macro definition has a wide variety of forms. It may be 0 to 500 characters in length (this may be increased via the RFPP macro maxbuf:), and may represent a static character string or an expression that changes with each invocation.

Each macro name in the input source code is replaced by its definition in the Fortran output. The macro facility enables complex 'definitions' to be inserted into the Fortran output via the 'name' invocation. To facilitate the definition of macros, and other associated macro functions, the preprocessor provides a variety of 'inbuilt' macro facilities.

Before describing these in detail it is necessary to define the general syntactical form of a macro as it is applied. It is:

		macro-name(arg1,arg2,...argN)

A macro may have from zero arguments (generally referred to as a 'static' macro) up to a maximum of nine.

Before a macro can be applied, it must be defined. This is done using the inbuilt RFPP macro macro:. In the macro definition, macro arguments appear as the digraphs $1,$2,...$9. When a macro in the ratmac code is processed, its arguments argN replace the digraphs in the stored definition. This occurs in 'dynamic' forms of the macro.

Before proceeding with explicit definitions, here are some simple examples of macros.

The static macro cname: is defined as 'BILL' by the command,

			macro:(cname:,BILL)

Thereafter any occurrence of cname: in the ratmac code will be replaced by BILL in the Fortran output.

Another static macro resolution: defines the resolution parameter as the value

			macro:(resolution:,0.00005)

Let us now consider a dynamic macro. The macro swap: is defined as

			macro:(swap:,Q=QX($1);QX($1)=QX($2);QX($2)=Q)

The occurence of swap:(J,K+1) in the ratmac code causes RFPP to output

			Q=QX(J)
			QX(J)=QX(K+1)
			QX(K+1)=Q

into the Fortran code. In another part of the ratmac code this macro may be invoked as swap:(KEY+5,INDEX-1) and this causes RFPP to generate the Fortran code

			Q=QX(KEY+5)
			QX(KEY+5)=QX(INDEX-1)
			QX(INDEX-1)=Q