CP2 - Parser

Get input character stream from tokenizer, then build the parse tree. (PT)

• Parent node to create an “empty” child node

• Pass the row number corresponding to the newly created node as argument

• Example: ParseIf() --
  • Call getToken(), check it is an 'if'
    • If not, print error and stop
• Example: ParseStmt() --
  • Call getToken, return a number, and apply one-token look ahead
  • Call corresponding alternative, e.g. ParseIf()
    • If the getToken returns an identifier, call parseAssign();
• Example: ParseExp() -- "9-5+4"

1. Check if current node is a <fac>
2. PT[n,0] = type_fac
3. PT[n,2] = nextRow++; // next free row in PT[]
4. Call ParseFac(PT[n,2]), create 1st child node
5. Call getToken() and check its type
   1. if it is a '+' -- PT[n,1] = 2
   2. if it is a '-' -- PT[n,1] = 3
   3. Otherwise -- PT[n,1] = 1, return
6. if PT[n,1] is 2/3, skipToken()
7. PT[n,3] = nextRow++; // next free row in PT[]
8. Call ParseExp(PT[n,3])
9. return;

• Example: ParseFac()

1. Check if current node is an <op>
2. PT[n,0] = type_op
3. Call ParseOp()
4. Call getToken()
   1. if the returned value is a '*',  -- PT[n,1] = 2, skipToken()
   2. if the returned value is not a '*',  -- PT[n,1] = 1, return
5. Call parseFac() if it is the 2nd alternative, then return

• ParseAssign() -- 

  • Check if the node is an identifier
    
    PT[n,0] = 7; // this is an <assign> node, construct the row and skip that token (for now)
    

  • PT[n,2] = nextRow++; // next free row in PT[]
    

  • Call ParseID(PT[n,2]) -- the identifier to the left of the '='
  • Now curser at '=', and check if it is actually a '=', then skipToken()

  • PT[n,3] = nextRow++; // next free row in PT[]
    

  • Now parse the expression on the right of the '=' -- parseExp()

  • if (s != ‘‘;’’} {**Quit!!**} else {T.skipToken();};
    return; }
    

        ParseID()

• Call getToken() and check if it is an ID
• Call IdName of the identifier
• Check the global variable to know if we are parsing <decl seq> or <stmt seq>
• Get the string, and look up that string in the IdName table to see if it is there, and make decision based on that 
• ParseID() construct the ID node and the put the id in the IdName table
• Skip pass the identifier -- skipToken()