1. http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction1 COMP30330 2009-2010 Compiler Construction Lecturer: Dr. Arthur Cater Teaching Assistant: Santiago Villalba Demonstrator: Zeeshan Ahmed

2. Admin issues 24 lectures, 3 assignments, 11 practical sessions 1 st assignment set on Thursday 21 Jan, due on Friday 12 Feb, worth 10% 2 nd assignment set on Thursday 4 Feb, due on Friday 26 March, worth 20% 3 rd assignment set on Thursday 4 March, due on Friday 23 April, worth 20% 2 hour exam will occur after semester end. Book “Compilers: Principles, Techniques and Tools” by Aho, Lam, Sethi & Ullman: 2 nd edition Each student should register for Monday practicals or Tuesday practicals. Attendance records will be kept. A Module Moodle exists at http://csimoodle.ucd.ie/moodle/course/view.php?id=98 http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction2

3. What does a compiler do? Compilers translate programs written in a “high-level language” into some other form That other form may be  machine code that can be directly executed by computer hardware  relocatable binary, needing more work on address references  assembly code, needing assembling &c  code for a virtual machine, such as the JVM for Java  equivalent code in another HLL, such as C http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction3

4. Front-end and Back-end The job of translating a program from one form to another is often broken down into two major stages: 1)Front-end: Analysing the source program, determining how its characters form words, how its words form statements, procedures, class definitions, etc how its statements etc conform to language rules, such as using only declared variables, using operands of proper type for operators … and reporting statically detectable errors in the program if they exist 1)Back-end: Generating an equivalent program in the target language Multiple implementations of a source language for different computers may share a front end and match it with different back ends. http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction4

5. Compilers vs Interpreters Interpreters do not translate programs, rather they simulate them. http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction5 Compiler (&c) Interpreter source program target program input output

6. Some special varieties of compiler Cross compiler Debugging compiler Optimizing compiler Batch compiler Load-and-go compiler It is quite common for a compiler for a language to be written in that same language. http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction6

7. Phases of a typical compiler http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction7 source program lexical analyzer token stream Machine-Independent Code Optimizer syntax analyzer intermediate representation syntax tree semantic analyzer intermediate code generator Code Generator target-machine code Machine-Dependent Code Optimizer target-machine code symbol table

8. Software relatives Various other software tools perform similar analysis functions to a compiler’s Syntax-directed editors –automatically insert text fragments near reserved words Prettyprinters and colorizers Static checkers –look for e.g.  unreachable code,  undeclared / unused variables,  datatype mismatches html / xml browsers http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction8

9. Metalanguages Chomsky hierarchy of types of language, distinguished by what restrictions may be placed on “productions” in an adequately descriptive “generative grammar”: http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction9 Type 0 (unrestricted) any LHS may be replaced by any RHS aXbYc  Pqr Type 1 (context-sensitive) a single nonterminal in the context of a LHS may be replaced by anything else in the same context aXbYc  aZwbYc Type 2 (context-free) a LHS may mention only a single nonterminal Type 3 (regular) a LHS may mention only a single nonterminal and a RHS may mention at most one terminal followed by at most one nonterminal X  pQr X  pY

10. Relevance of types of language Regular languages are often used to describe the word-level syntax of a HLL rules for valid identifiers, numbers, strings, reserved words, etc finite-state automata can recognise regular languages tools such as ‘lex’, ‘Flex’, ‘ANTLR’, ‘JavaCC’ can build a lexical analyzer program ( lexer, scanner ) when supplied with a regular grammar describing the desired regular language; or hand coding can be used Context-free languages are used to describe the phrase-level syntax of a HLL rules for expressions, statements, compound statements, conditionals, etc push-down automata can recognise context-free languages tools such as ‘Yacc’, ‘Bison’, ‘ANTLR’, ‘JavaCC’ can build a syntax analyzer program ( parser ) when supplied with a context-free grammar describing the desired context-free language; or hand coding can be used http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction10

11. Beyond Recognition A Finite State Automaton can classify character sequences as numbers, ids, etc. A parser can operate simply at the level of token sequences. But mere yes/no judgements are not what is required of lexers, parsers. Associating “semantic actions” with grammar productions allows lexers, parsers to  maintain a symbol table, distinguishing different identifiers  build the values of numeric expressions  generate simple code as a by-product of parsing http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction11

12. Symbol table An unsophisticated symbol table may have the following form: http://csiweb.ucd.ie/staff/acater/comp30330.html Compiler Construction12 tom\\dick harry 0:0 1:4 2:9 3:\\ 4:\\ 5:\\ 6:\\ Semantic actions associated with state transitions in a finite state automaton can accumulate characters in a buffer, then at an accepting state look up in symbol table, and insert new entry if no match is found.