==== Compilers Exam, 2015-12-18 ====
=== Theory ===

  - What is the input and output of the //lexical analyser (lexer)//?
  - Which of the above inputs match the ''[ab]*(ab)+'' regular expression?
    * ''a'', ''b'', ''ab'', ''ba'', ''aab'', ''bab'', ''baba'', ''ababaab''
  - Given the following token definitions, how will the lexer tokenize the ''if else ifelse apple'' text?<code>
if      cout << "IF";
else    cout << "ELSE";
[a-z]+  cout << "IDENTIFIER";
" "+    cout << "SPACES"
</code>
  - Write down the steps of the derivation of ''aaabbb'' (//S => ... => aaabbb//) using the grammar below!
    * //S -> ab | aSb//
  - What is the difference between top-down and bottom-up parsing?
  - Write down the function of a recursive descent parser corresponding to the ''A -> aB | c'' rules!
  - What are the two operations that an LR parser can perform?
  - What is a symbol table used for?
  - Write a 32-bit NASM assembly program that swaps the contents of the ''dword [a]'' and ''dword [b]'' memory locations!
  - Write a code generation rule for an assignment operation of the form //assignment -> variable := expression//!

=== Practice ===

Extend the //While language// with a new type, called //time//. An example:
<code>
program exam
  time t;
  integer i;
begin
  t := 09:52;
  i := hour(t) + minute(10:02);
  write(i);
end
</code>

  * Literals of type time should be of the form //HH:MM//, where the two digits of the hour and minute are arbitrary decimal digits. (It is OK to accept impossible time literals like ''99:99'', but you are free to create smart regular expressions that allow literals only from the ''00:00'' - ''23:59'' range.)
  * It should be possible to declare variables of type //time//.
  * Assignement operations should also work for time-typed variables.
  * //hour// and //minute// should get a time parameter and return integers, the hour and minute respectively.
  * It is not possible to input or output time values!

Tasks:
  * Download and see the [[http://deva.web.elte.hu/compilers/compilersExam1.zip|test files]].
  * Grade 2: Extend the lexer with the necessary regular expressions!
  * Grade 3: Extend the parser with the new rules needed!
  * Grade 4: Extend the type checking with //time//-related functionality!
  * Grade 5: Extend the code generation to support the new type!
    * Hint: Store type values as integers, using the encoding //100*HH + MM//, or store the hours and minutes in two different bytes.