java_cup

Class lalr_state

public class lalr_state extends Object

This class represents a state in the LALR viable prefix recognition machine. A state consists of an LALR item set and a set of transitions to other states under terminal and non-terminal symbols. Each state represents a potential configuration of the parser. If the item set of a state includes an item such as:
    [A ::= B * C d E , {a,b,c}]
  
this indicates that when the parser is in this state it is currently looking for an A of the given form, has already seen the B, and would expect to see an a, b, or c after this sequence is complete. Note that the parser is normally looking for several things at once (represented by several items). In our example above, the state would also include items such as:
    [C ::= * X e Z, {d}]
    [X ::= * f, {e}]
  
to indicate that it was currently looking for a C followed by a d (which would be reduced into a C, matching the first symbol in our production above), and the terminal f followed by e.

At runtime, the parser uses a viable prefix recognition machine made up of these states to parse. The parser has two operations, shift and reduce. In a shift, it consumes one Symbol and makes a transition to a new state. This corresponds to "moving the dot past" a terminal in one or more items in the state (these new shifted items will then be found in the state at the end of the transition). For a reduce operation, the parser is signifying that it is recognizing the RHS of some production. To do this it first "backs up" by popping a stack of previously saved states. It pops off the same number of states as are found in the RHS of the production. This leaves the machine in the same state is was in when the parser first attempted to find the RHS. From this state it makes a transition based on the non-terminal on the LHS of the production. This corresponds to placing the parse in a configuration equivalent to having replaced all the symbols from the the input corresponding to the RHS with the symbol on the LHS.

Version: last updated: 7/3/96

Author: Frank Flannery

See Also: lalr_item lalr_item_set lalr_transition

Field Summary
protected static intnext_index
Static counter for assigning unique state indexes.
protected static Hashtable_all
Collection of all states.
protected static Hashtable_all_kernels
Hash table to find states by their kernels (i.e, the original, unclosed, set of items -- which uniquely define the state).
protected int_index
Index of this state in the parse tables
protected lalr_item_set_items
The item set for this state.
protected lalr_transition_transitions
List of transitions out of this state.
Constructor Summary
lalr_state(lalr_item_set itms)
Constructor for building a state from a set of items.
Method Summary
voidadd_transition(symbol on_sym, lalr_state to_st)
Add a transition out of this state to another.
static Enumerationall()
Collection of all states.
static lalr_statebuild_machine(production start_prod)
Build an LALR viable prefix recognition machine given a start production.
voidbuild_table_entries(parse_action_table act_table, parse_reduce_table reduce_table)
Fill in the parse table entries for this state.
static voidclear()
protected static voiddump_state(lalr_state st)
Helper routine for debugging -- produces a dump of the given state onto System.out.
booleanequals(lalr_state other)
Equality comparison.
booleanequals(Object other)
Generic equality comparison.
static lalr_statefind_state(lalr_item_set itms)
Find and return state with a given a kernel item set (or null if not found).
protected booleanfix_with_precedence(production p, int term_index, parse_action_row table_row, parse_action act)
Procedure that attempts to fix a shift/reduce error by using precedences.
inthashCode()
Produce a hash code.
intindex()
Index of this state in the parse tables
protected parse_actioninsert_action(parse_action a1, parse_action a2, int act_type)
protected parse_actioninsert_reduce(parse_action a1, parse_action a2)
protected parse_actioninsert_shift(parse_action a1, parse_action a2)
lalr_item_setitems()
The item set for this state.
static intnumber()
Indicate total number of states there are.
protected static voidpropagate_all_lookaheads()
Propagate lookahead sets through the constructed viable prefix recognizer.
protected voidpropagate_lookaheads()
Propagate lookahead sets out of this state.
protected voidreport_conflicts(terminal_set conflict_set)
Produce warning messages for all conflicts found in this state.
protected voidreport_reduce_reduce(lalr_item itm1, lalr_item itm2)
Produce a warning message for one reduce/reduce conflict.
protected voidreport_shift_reduce(lalr_item red_itm, int conflict_sym)
Produce a warning message for one shift/reduce conflict.
StringtoString()
Convert to a string.
lalr_transitiontransitions()
List of transitions out of this state.

Field Detail

next_index

protected static int next_index
Static counter for assigning unique state indexes.

_all

protected static Hashtable _all
Collection of all states.

_all_kernels

protected static Hashtable _all_kernels
Hash table to find states by their kernels (i.e, the original, unclosed, set of items -- which uniquely define the state). This table stores state objects using (a copy of) their kernel item sets as keys.

_index

protected int _index
Index of this state in the parse tables

_items

protected lalr_item_set _items
The item set for this state.

_transitions

protected lalr_transition _transitions
List of transitions out of this state.

Constructor Detail

lalr_state

public lalr_state(lalr_item_set itms)
Constructor for building a state from a set of items.

Parameters: itms the set of items that makes up this state.

Method Detail

add_transition

public void add_transition(symbol on_sym, lalr_state to_st)
Add a transition out of this state to another.

Parameters: on_sym the symbol the transition is under. to_st the state the transition goes to.

all

public static Enumeration all()
Collection of all states.

build_machine

public static lalr_state build_machine(production start_prod)
Build an LALR viable prefix recognition machine given a start production. This method operates by first building a start state from the start production (based on a single item with the dot at the beginning and EOF as expected lookahead). Then for each state it attempts to extend the machine by creating transitions out of the state to new or existing states. When considering extension from a state we make a transition on each symbol that appears before the dot in some item. For example, if we have the items:
    [A ::= a b * X c, {d,e}]
    [B ::= a b * X d, {a,b}]
  
in some state, then we would be making a transition under X to a new state. This new state would be formed by a "kernel" of items corresponding to moving the dot past the X. In this case:
    [A ::= a b X * c, {d,e}]
    [B ::= a b X * Y, {a,b}]
  
The full state would then be formed by "closing" this kernel set of items so that it included items that represented productions of things the parser was now looking for. In this case we would items corresponding to productions of Y, since various forms of Y are expected next when in this state (see lalr_item_set.compute_closure() for details on closure).

The process of building the viable prefix recognizer terminates when no new states can be added. However, in order to build a smaller number of states (i.e., corresponding to LALR rather than canonical LR) the state building process does not maintain full loookaheads in all items. Consequently, after the machine is built, we go back and propagate lookaheads through the constructed machine using a call to propagate_all_lookaheads(). This makes use of propagation links constructed during the closure and transition process.

Parameters: start_prod the start production of the grammar

See Also: lalr_item_set lalr_state

build_table_entries

public void build_table_entries(parse_action_table act_table, parse_reduce_table reduce_table)
Fill in the parse table entries for this state. There are two parse tables that encode the viable prefix recognition machine, an action table and a reduce-goto table. The rows in each table correspond to states of the machine. The columns of the action table are indexed by terminal symbols and correspond to either transitions out of the state (shift entries) or reductions from the state to some previous state saved on the stack (reduce entries). All entries in the action table that are not shifts or reduces, represent errors. The reduce-goto table is indexed by non terminals and represents transitions out of a state on that non-terminal.

Conflicts occur if more than one action needs to go in one entry of the action table (this cannot happen with the reduce-goto table). Conflicts are resolved by always shifting for shift/reduce conflicts and choosing the lowest numbered production (hence the one that appeared first in the specification) in reduce/reduce conflicts. All conflicts are reported and if more conflicts are detected than were declared by the user, code generation is aborted.

Parameters: act_table the action table to put entries in. reduce_table the reduce-goto table to put entries in.

clear

public static void clear()

dump_state

protected static void dump_state(lalr_state st)
Helper routine for debugging -- produces a dump of the given state onto System.out.

equals

public boolean equals(lalr_state other)
Equality comparison.

equals

public boolean equals(Object other)
Generic equality comparison.

find_state

public static lalr_state find_state(lalr_item_set itms)
Find and return state with a given a kernel item set (or null if not found). The kernel item set is the subset of items that were used to originally create the state. These items are formed by "shifting the dot" within items of other states that have a transition to this one. The remaining elements of this state's item set are added during closure.

Parameters: itms the kernel set of the state we are looking for.

fix_with_precedence

protected boolean fix_with_precedence(production p, int term_index, parse_action_row table_row, parse_action act)
Procedure that attempts to fix a shift/reduce error by using precedences. --frankf 6/26/96 if a production (also called rule) or the lookahead terminal has a precedence, then the table can be fixed. if the rule has greater precedence than the terminal, a reduce by that rule in inserted in the table. If the terminal has a higher precedence, it is shifted. if they have equal precedence, then the associativity of the precedence is used to determine what to put in the table: if the precedence is left associative, the action is to reduce. if the precedence is right associative, the action is to shift. if the precedence is non associative, then it is a syntax error.

Parameters: p the production term_index the index of the lokahead terminal parse_action_row a row of the action table act the rule in conflict with the table entry

hashCode

public int hashCode()
Produce a hash code.

index

public int index()
Index of this state in the parse tables

insert_action

protected parse_action insert_action(parse_action a1, parse_action a2, int act_type)

insert_reduce

protected parse_action insert_reduce(parse_action a1, parse_action a2)

insert_shift

protected parse_action insert_shift(parse_action a1, parse_action a2)

items

public lalr_item_set items()
The item set for this state.

number

public static int number()
Indicate total number of states there are.

propagate_all_lookaheads

protected static void propagate_all_lookaheads()
Propagate lookahead sets through the constructed viable prefix recognizer. When the machine is constructed, each item that results in the creation of another such that its lookahead is included in the other's will have a propagate link set up for it. This allows additions to the lookahead of one item to be included in other items that it was used to directly or indirectly create.

propagate_lookaheads

protected void propagate_lookaheads()
Propagate lookahead sets out of this state. This recursively propagates to all items that have propagation links from some item in this state.

report_conflicts

protected void report_conflicts(terminal_set conflict_set)
Produce warning messages for all conflicts found in this state.

report_reduce_reduce

protected void report_reduce_reduce(lalr_item itm1, lalr_item itm2)
Produce a warning message for one reduce/reduce conflict.

Parameters: itm1 first item in conflict. itm2 second item in conflict.

report_shift_reduce

protected void report_shift_reduce(lalr_item red_itm, int conflict_sym)
Produce a warning message for one shift/reduce conflict.

Parameters: red_itm the item with the reduce. conflict_sym the index of the symbol conflict occurs under.

toString

public String toString()
Convert to a string.

transitions

public lalr_transition transitions()
List of transitions out of this state.