Syntax Directed Translation

1
Syntax Directed Translation
2
Why do we need Syntax-directed?

• To translate a programming language construct, a
  compiler may need to keep track of many
  quantities besides the code generated for the
  construct.
• Syntax-directed
• We augment a context-free grammar by adding
  semantic rules or actions to each production. 
• These rules define the values of attributes and
  how to compute them. 
• We call this augmented grammar an attribute
  grammar.
• Two types of attributes
• synthesized attribute which computes the value
  based on the children
• inherited attribute computed based on the parent
  and the older siblings.
• We can use two methods to translate the
  programming language
• a formalism, syntax- directed definition
• a more procedural notation called translation
  scheme.

3
Syntax-directed definition(1/2)

• Syntax-directed definition
• A grammar -oriented compiling technique
• It is very helpful for organizing the compiler's
  front end and it translates infix expression into
  post form
• Postfix notation
• 1) If E is a variable or constant, then the
  postfix notation for E is E itself.
• 2) If E is an expression of the form E1 op E2
  where op is any binary operator, then the postfix
  notation foe E is E1' E2' op.
• 3) If E is an expression of the form (E1), then
  the postfix notation for E1 is also the postfix
  notation for E.

4
Syntax-directed definition(2/2)

• Essentially a formal treatment of two simple
  concepts
• Each node of a parse tree can have a set of
  associated values (attributes).
• The creation and traversal of parse trees can be
  automated
• Example
• if we have a production
• A --gt X1 X2 ... Xn
• we might have a rule
• A.a f(X1.a,X1.b,X2.a,X2.b,...,Xn.a,Xn.b)
• which describes how the attribute a  of A is
  defined in terms of attributes of the children of
  A.

5
Translation scheme(1/2)

• A translation is an iuput-output mapping
• The output for each input x is specified in the
  following way
• construct a parse tree for x
• compute the value using semantic rule for
  attribute associated with the X-production used
  at node n
• two types of attributes
• synthesized attribute which computes the value
  based on the children
• inherited attribute computed based on the parent
  and the older siblings
• A translation scheme gives a procedural
  description of an action to be taken when a
  production is applied

6
Translation scheme(2/2)

• The actions may be interspersed with the symbols
  on the right-hand side of the production
• thus indicating an evaluation order
• For example, in a translation scheme used to
  generate assembly code
• while-stmt  --gt  while   generate label
  L1,L2 emit("L1"  )                         
  expression emit("beqz L2")
                           statement
  emit("b   L1) emit("L2")
• Attributes may be evaluated during parsing or in
  a separate pass over the parse tree
• Different strategies are used, depending whether
  a top-down or bottom-up parser is used

7
How attributes are evaluated depends on whether
they are synthsized or inherited

• A synthesized attribute of a symbol is one which
  depends only on the attributes of its children
  (in the parse tree)
• An inherited attribute is one which depends on
  attributes of the parent or siblings (ex. the
  data type of an identifier)
• In general, inherited attributes are more
  difficult to deal with, particularly for
  bottom-up parsers

8
How attributes are evaluated depends on whether
they are synthesized or inherited. 

• A synthesized attribute of a symbol is one which
  depends only on the attributes of its children
  (in the parse tree)
• for example, the value of an expression. 
• An inherited attribute is one which depends on
  attributes of the parent or siblings
• for example the data type of an identifier. 
• In general, inherited attributes are more
  difficult to deal with, particularly for
  bottom-up parsers.

9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13

• When the syntax-directed definition has only
  synthesized attribute, it is a S-attribute
  definitions. It can be computed by using
  bottom-up parser. While a syntax-directed
  definition is L-attribute if each inherited
  attribute of X_j, 1lt j lt n, on the right side
  of A-gt X1X2Xn, depends only on
• 1)      the attribute of the symbols
  X1,X2,,Xj-1 to the left of Xj in the production
  and
• 2)      the inherited attributes of A.

14

• Every S-attributed definition is L-attributed,
  because the restriction 1) and 2) apply only to
  inherited attributes. It can be computed in
  depth-first order. All syntax-directed
  definition based on LL(1) grammars are
  L-attributed. 
• Following are some examples of S-attributed and
  L-attributed
• 1)     synthesized attribute on the parser
  stack
• 2)     A non-L-attributed syntax-directed
  definition

X X.x
Y Y.y
Z Z.z

Production Semantic rule
A -gt LM L.i l(A,I)
M.i m(L.s)
A.s f(M.s)

A -gt QR R.i r(A.i)
Q.i q(R.s)
A.s f(Q.s)