hw1 - Ç-transitions you don’t need to use Thompson’s...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
HW 1: Lexical Analysis CSC 4351, Spring 2011 Due: 21 February 2011 1. Certain assemblers form their integer literals in the following way. Binary literals consist of one or more binary digits (0, 1) followed by the letter B ; e.g., 10110B . Octal literals consist of one or more octal digits (0 through 7) followed by the letter Q (since O looks too much like 0 ); e.g., 1234567Q . Hexadecimal literals consist of at least one decimal digit (0 through 9), followed by zero or more hexadecimal digits (0 through 9 and A through F) followed by the letter H ; e.g., 0ABCDEFH . Decimal literals consist of at least one decimal digit optionally followed by the letter D ; e.g., 1234 . (a) Draw the state diagram of an NFA (not a DFA) for these literal forms; you may use
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ǫ-transitions, you don’t need to use Thompson’s construction. (b) Give a regular expression for the literals; you may use ǫ . (c) Modula-3 based literals have the form base digits (e.g., 2_11111111 , 8_377 , 16_FF , 255 are equivalent integer literals). Why is the Modula-3 form preferable to the assembler form introduced here, from the standpoint of a compiler (or assembler) writer? 2. (a) Using Thompson’s construction, construct an NFA that recognizes the same language as defined by the following regular expression: (1 * 01 * 0) * 1 * (b) Using the subset construction, convert the NFA into a DFA. Optimize the resulting DFA by merging any equivalent states. 1...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online