Phase Project 1
CS 441 Fall 1992 Craig A. Rich
A Pure Lisp interpreter will be constructed of several packages in two phases. In order to insure synchronization of the phases, the constants and types which are used throughout the interpreter will be specified in a separate LANGUAGE package to be used by most others. The visible entities of each package will also be specified, controlling the interaction between packages. This can be implemented using module extensions available in Vax Pascal, standard packages available in Ada, or classes available in C++. Language and Lexical Analyzer Notes 1. I have compiled the source files for the LANGUAGE and LEXICAL_ANALYZER packages and the compilation products are available for linking, so you should not create these source files. The object files and compilation unit files produced from the Ada source files reside in the library [cs.carich.adalib]. The object files and environment files produced from the Pascal source files are: [cs.carich]language.obj [cs.carich]language.pen [cs.carich]lexical_analyzer.obj [cs.carich]lexical_analyzer.pen
Lexical Analyzer Notes 1. The lexical analyzer reads source code from the standard input file, which must be defined to a physical file before you run your interpreter. Invocations of NEXT_TOKEN will return successive tokens recognized in the source code. 2. If an identifier lexeme consisting of a letter followed by zero or more letters or digits is seen, NEXT_TOKEN will return a token whose SYMBOL is id; the identifiers T, NIL, QUOTE, CAR, CDR, CONS, EQ, ATOM, COND, LABEL, and LAMBDA are reserved and NEXT_TOKEN will return a token whose SYMBOL is T, NIL, QUOTE, CAR, CDR, CONS, EQ, ATOM_t, LABEL, or LAMBDA, respectively. In the Pascal implementation, identifiers longer than MAX_STRING_LENGTH will be truncated. Letters occurring in identifiers may be upper or lower case and will be converted to upper case. If the lexemes ( or ) are seen, NEXT_TOKEN will return a token whose SYMBOL is '(' or ')', respectively. If a comment lexeme beginning with { and ending with } is seen, NEXT_TOKEN will ignore and bypass it. If no lexemes remain, NEXT_TOKEN will return a token whose SYMBOL is '$'. Lexemes may be separated by any number of blanks, horizontal tabs, or carriage Implementation returns.
Stack Notes 1. Implement the specified stack operations, using an implementation which does not bound the number of elements on the stack. Each element will contain a TOKEN and a SET_OF_ITEMS--representing grammar symbols in a parse tree which is being constructed "bottom-up" and the current state of a DFA which is recognizing viable prefixes of right sentential forms. Syntax Analyzer Implementation Notes 1. Complete the implementation of an SLR syntax analyzer which accepts Pure Lisp programs generated by the CFG in homeworks 1 and 2. 2. The productions P and SLR(1) parse table M constructed in homework 2 should be declared globally in the syntax analyzer. Their values may be initialized by declaration or by loading from external files. Their values should not be initialized by successive assignment statements. Appropriate external files are: [cs.carich]ada_productions.dat [cs.carich]ada_parsetable.dat [cs.carich]pascal_productions.dat [cs.carich]pascal_parsetable.dat
3. The syntax analyzer should output "shift" followed by the symbol in TERMINAL which is being shifted whenever a shift occurs. It should output "reduce" followed by the production (not its number) which is being used whenever a reduce occurs. Note that the sequence of reduce productions is the reverse sequence of produ...