CSC 533: Programming Languages
Spring 2021

HW4: Implementing Subroutines

The SILLY interpreter you wrote for HW3 extends the original version to allow for local variables and nested scopes. For this final SILLY assignment, you are to extend the language further by implementing simple subroutines (without return values).

In order to implement simple subroutines, you will need to define two new classes, a subDecl class for declaring a subroutine and a subCall class for calling a subroutine. The grammar rules for these new statements are:

<subdecl> --> 'sub' <id> '(' [ <p-list> ] ')' <compound> <p-list> --> <type> <id> { ',' <type> <id> } <subcall> --> 'call' <id> '(' [ <e-list> ] ')' ';' <e-list> --> <expr> { ',' <expr> }

Note that a subroutine declaration specifies the name of the subroutine, an arbitrary number of parameters (type/variable pairs in parentheses, separated by commas), and a compound statement specifying the body of the subroutine. A subroutine call specifies the subroutine name and the sequence of expressions (also in parentheses, separated by commas) that correspond to the parameters. A subroutine call should result in a run-time error is there is no declared subroutine with that name, the number of expressions in a subroutine call does not match the number of parameters in its declaration, or if there are type mismatches between the expressions and parameters. Note: every subroutine declaration is assumed to be in the global scope, regardless of where that declaration occurs.

When a subroutine is called, it defines a new, independent scope. Parameters are treated as local variables in that scope, initialized based on the expressions provided in the subroutine call. In addition, there can be nested scopes within a subroutine via compound statements. Because a subroutine's scope is independent, statements within a subroutine cannot access variables outside that subroutine. As a result, it is legal to declare a subroutine parameter or local variable with whose name is already declared in an external scope.

For example:

SAMPLE CODE (output in red) 
  >>> sub foo ( ) {
          output "foo" ;
      }
  >>> call foo ( ) ;
  foo
  >>> sub avg ( int num1 , int num2 ) {
          int sum = num1 + num2 ;
          int val = sum / 2 ;
          output "average" , "of" , num1 ,  
                 "and" , num2 , "=" , val ;
      }
  >>> call avg ( 2 , 4 ) ;
  average of 2 and 4 = 3  
  >>> int n = 20 ;
  >>> call avg ( n , n + 5 ) ;
  average of 20 and 25 = 22  
  >>> sub sumTo ( int num ) {
          int sum = 0 ;
          while num > 0 {
              sum = sum + num ;
              num = num - 1 ;
          }
          output sum ;
      }
  >>> call sumTo ( 10 ) ;
  55 
  >>> call sumTo ( 100 ) ;
  5050 


  >>> sub countdown ( int n ) {
          if n == 0 {
              output "BLASTOFF" ;
          }
          else {
              output n ;
              call countdown ( n - 1 ) ;
          }
      }
  >>> call countdown ( 5 ) ;
  5
  4
  3
  2
  1
  BLASTOFF 
  >>> sub stamp ( str word , int times ) {
          str final = "" ;
          repeat times {
              final = final + word ;
          }
          output final ;
      }
  >>> call stamp ( "foo" , 3 ) ;
  foofoofoo 
  >>> call stamp ( "X" , 10 ) ; 
  XXXXXXXXXX

Note that executing a subroutine declaration does NOT result in its code being executed. Declaring a subroutine simply stores the code associated with that subroutine in memory so that it can be called later. You will need to add a third field to the MemorySpace class, corresponding to the code segment. When a subroutine declaration is executed, the parameters and code for that subroutine should be stored in the code segment. Subsequently, when a subroutine call is executed, the corresponding parameters and code must be accessed from the code segment so that it can be executed.