Compiler Chap 6' Symbol Table

1
CompilerChap 6. Symbol Table

• Internet Management Technology Lab.

2
Contents

• Introduction
• Various implementation techniques
• Representing scope information
• Forward reference

3
Introduction

• Symbol Table
• A data structure that stores information about
  various source language constructs
• A compiler uses symbol table to keep track of
  scope and binding information about names
• Entries
• An entry is for the declaration of a name
• The format depends on the usage of the name (e.g.
  records)
• It could have multiple entries for the same name
  in some cases
• e.g. int x vs. struct x float y, z
• ? variable name vs. tag name
• The content of entries
• attribute names, scope identifier, type, size,
  etc.

4
Introduction

• Basic symbol table interface
• Insert function
• insert a new entry into the table
• Lookup function
• find an existing entry (even reserved words can
  be handled in the same manner)

5
Various implementation techniques (1/4)

• Unordered list
• Simplest possible storage mechanism
• But impractically slow for large table
• Ordered list
• The entries are kept ordered
• Lookup is fairly efficient (O(logn), but
  insertion operation is relatively expensive
• It is useful for small sized tables
• Binary or 2-3 search trees
• Designed to combine the size flexibility and
  insertion efficiency of a linked data structure
• O(logn) time for average case, but not guaranteed
  because the variable uses are not random

6
Various implementation techniques (2/4)

• Hash tables
• The most common means of implementing the symbol
  tables in production compilers
• It needs a good hash function and collision
  handing techniques
• O(c) time complexity
• hash function
• Map each of possible names into one of a fixed
  number of positions
• The properties are
• hash function h(n) is function of n
• h can be computed quickly
• h is uniform and randomizing in order to minimize
  collision

7
Various implementation techniques (3/4)

• Hash tables
• Simple techniques to find h(n)
• Add up the integer values of the characters in a
  string
• hi  hi-1 ci where h1 0
• hi  ahi-1 xor ci
• Resolving collisions when h(n) is occupied
• linear resolution
• try (h(n) 1) mod n, (h(n) 2) mod n
• should be careful about the long chain
• add-the-hash rehash
• try (2 h(n)) mod m, (3 h(n)) mod m
• Prevent long chain and m must be prime to try
  every slots

8
Various implementation techniques (4/4)

• Hash tables
• Resolving collisions when h(n) is occupied
  (contd)
• Quadratic rehash
• Try (h(n) 12) mod m, (h(n) 22)
• Chaining (Dynamic hash)
• Records that hash to a given value are chained
  together on a linked list

9
Representing scope information

• Scope rules of the source language determine
  which declaration is appropriate
• Example
• main()
•    
•        int i
•            i 5
•                   
• int i
•                    i 7 print i
•            
•    
• Simple approach is to keep separate tables for
  each scope

10
Forward reference

• Problem
• Reference to a name appears before definition of
  the language
• e.g.
• S1
• TEST
• JUMP  place
• S2      
• place
• Resolutions
• Backpatching
• Leave space for unknown value and fill it later
• Multiple pass resolution
• Resolve it at the later passes