The data type of a parameter is declared in the routine (see Defining a parameter). However, a variable passed to a routine parameter may have been declared with a different data type in the calling routine. In this case, the data type actually used in the routine may not be what you would expect.
Keep the following rules in mind regarding what happens to argument data types when passing arguments to routines:
- If the parameter in the routine is defined as numeric (n), the data type of the passed-in variable will be used in the routine. If the parameter is passed to another routine, the caller’s data type is passed to the called routine.
For example, if a decimal variable is passed to a routine where the parameter is defined as numeric, the parameter will be treated as a decimal within the routine. If an implied-decimal (d.) variable is passed, it will be treated as implied-decimal. We recommend that you define parameters as n rather than d, d., or i to ensure that the passed-in data retains its data type. You can use the numeric parameter compiler option to convert all decimal-type arguments to numeric. See Compiler options.
- In traditional Synergy without strong prototyping, if the parameter in the routine is explicitly typed (a, d, d., i, p, p.), that type will be used within the routine, regardless of the data type in the calling routine. If this parameter is passed as an argument to another subroutine, it is passed as the defined type, not the type passed by the caller.
For example, if a decimal variable of four decimal digits is passed to a routine where the parameter is defined as integer, the argument will be treated as an i4 variable by interpreting the passed variable as binary data—with no conversion within the routine.
- If the parameter in the routine is defined as implied-numeric (n.), the compiler will ensure that the intermediate results of any expression that uses this variable will result in an implied result. This also applies to a d variable passed to a parameter defined as d. and a p variable passed to a parameter defined as p..
|
If the passed-in argument is implied-decimal and the parameter in the routine is defined as decimal or integer, the whole number part is used after the fractional part is rounded. You can change the program default to truncate in traditional Synergy by setting system option #11. Specifying the ROUND or TRUNCATE option on a MAIN, FUNCTION, or SUBROUTINE statement overrides the default rounding behavior for that routine (including if system option #11 is set). See Preprocessor and Compiler Directives for information on compiler directives. (Neither TRUNCATE nor system option #11 are supported in Synergy .NET.) |
If a subroutine passes the data to a second subroutine, the rules defined above still apply. For example, say you have two subroutines, sub1 and sub2. In sub1, arg1 is declared as alpha; in sub2, arg1 is declared as numeric. If var is defined as a decimal in the calling program, and then passed to sub1 as arg1, it will be treated as an alpha within sub1. If sub1 then passes var to sub2 as its arg1, it will be treated as an alpha within sub2.
|
|
If your routine has multiple optional arguments and you want to make use of strong prototyping, you can specify the VARARGS modifier in your routine definition to enable you to pass more arguments than are declared without causing problems. (Note that if you use VARARGS, you will not be able to run your code on versions of Synergy prior to 9.1.1.) |
