public void BuildWithValidGrammar( RestrictedStartSymbolGrammar grammar, LR0CanonicalSets expected )
{
var actual = LR0CanonicalSets.Build(grammar);
Assert.NotNull(actual);
Assert.True(actual.Equals(expected));
}
/// <summary>
/// Creates and returns the LR(0) canonical sets for the given <paramref name="grammar"/>.
/// </summary>
/// <param name="grammar">The grammar to build the canonical sets for.</param>
/// <returns>The built canonical sets.</returns>
/// <exception cref="ArgumentNullException"> if <paramref name="grammar"/> is <c>null</c>.</exception>
public static LR0CanonicalSets Build( RestrictedStartSymbolGrammar grammar )
{
if( grammar == null )
throw new ArgumentNullException("grammar");
var terms_and_gramms = new List<Symbol>();
terms_and_gramms.AddRange(grammar.Grammaticals);
terms_and_gramms.AddRange(grammar.Terminals);
var sets = new LR0CanonicalSets();
// Calculate the I0 canonical set from the 'starting' rule
var s0 = new HashSet<LR0Item>();
if( grammar.Rules.Count > 0 )
s0.Add(new LR0Item(grammar.StartRule, 0));
var I0 = Closure(s0, grammar);
sets.Add(I0);
// Generate LR0CanonicalSets with Read()
for( int current_set_id = 0; current_set_id < sets.Sets.Count; ++current_set_id )
// Generate new sets for every symbol read
foreach( var symbol in terms_and_gramms )
{
var new_I = Read(sets.Sets[current_set_id], symbol, grammar);
// If the set is empty then continue
if( new_I.Count == 0 )
continue;
// Identify new_I
int next_set_id = sets.IndexOf(new_I);
// If this is a new CanonicalSet then add to the group
if( next_set_id == -1 )
next_set_id = sets.Add(new_I);
// Record the state transition
sets.SetTransition(current_set_id, symbol, next_set_id);
}
return sets;
}
/// <summary>
/// Builds and return an SLR(1) <see cref="ParserTable"/> based on the given paramaters.
/// </summary>
/// <param name="grammar">The grammar.</param>
/// <param name="canonical_sets">LR(0) canonical sets generated from the <paramref name="grammar"/>.</param>
/// <param name="follow_sets">Follow sets of all the <paramref name="grammar"/>'s grammaticals
/// (see <see cref="FollowSets"/>).</param>
/// <returns>An SLR(1) parser table.</returns>
public static ParserTable Build( ExtendedGrammar grammar, LR0CanonicalSets canonical_sets,
IDictionary<GrammaticalSymbol, ISet<TerminalSymbol>> follow_sets)
{
if( object.ReferenceEquals(grammar, null) )
throw new ArgumentNullException("grammar");
if( object.ReferenceEquals(canonical_sets, null) )
throw new ArgumentNullException("canonical_sets");
if( object.ReferenceEquals(follow_sets, null) )
throw new ArgumentNullException("follow_sets");
var table = new ParserTable(canonical_sets.Sets.Count, grammar.Terminals.Count, grammar.Grammaticals.Count);
for( int state_index = 0; state_index < canonical_sets.Sets.Count; ++state_index )
{
// Steps
foreach( var term in grammar.Terminals )
{
int next_state_index = canonical_sets.GetTransition(state_index, term);
if( next_state_index != -1 )
{
if( table.GetAction(state_index, grammar.IndexOf(term)).Type != ParserTable.ActionType.Error )
throw new ArgumentException("Grammar is not SLR(1)."); //TODO: custom exception with conflict info
table.SetAction(state_index, grammar.IndexOf(term), ParserTable.Action.Step(next_state_index));
}
}
// Gotos
foreach( var gram in grammar.Grammaticals )
{
int next_state_index = canonical_sets.GetTransition(state_index, gram);
if( next_state_index != -1 )
{
if( table.GetGoto(state_index, grammar.IndexOf(gram)) != -1 )
throw new ArgumentException("Grammar is not SLR(1)."); //TODO: custom exception with conflict info
table.SetGoto(state_index, grammar.IndexOf(gram), next_state_index);
}
}
// Reductions
foreach( var item in canonical_sets.Sets[state_index] )
if( item.CursorPosition == item.Rule.RightHandSide.Count )
{
// Is it [S'->S.] ?
if( item.Rule.LeftHandSide == grammar.StartSymbol )
{
if( table.GetAction(state_index, grammar.IndexOf(grammar.EndOfSourceSymbol)).Type != ParserTable.ActionType.Error )
throw new ArgumentException("Grammar is not SLR(1)."); //TODO: custom exception with conflict info
table.SetAction(
state_index,
grammar.IndexOf(grammar.EndOfSourceSymbol),
ParserTable.Action.Accept());
}
// ... no, it's [A->alpha.]
else
{
int rule_index = grammar.Rules.IndexOf(item.Rule);
foreach( var follow_term in follow_sets[item.Rule.LeftHandSide] )
{
if( table.GetAction(state_index, grammar.IndexOf(follow_term)).Type != ParserTable.ActionType.Error )
throw new ArgumentException("Grammar is not SLR(1)."); //TODO: custom exception with conflict info
table.SetAction(
state_index,
grammar.IndexOf(follow_term),
ParserTable.Action.Reduction(rule_index));
}
}
}
}
return table;
}
public bool Equals( LR0CanonicalSets other )
{
if( object.ReferenceEquals(other, null) )
return false;
if( !other.Sets.SequenceEqual(other.Sets,
new Utils.DelegateEqualityComparer<ISet<LR0Item>>(
( x, y ) => LR0CanonicalSets.AreEqual(x, y),
( obj ) => obj.GetHashCode())) )
return false;
//TODO: Check transitions too?
// If the sets are the same, the transition must be the same too.
return true;
}
static Grammar2()
{
var terminals = new TerminalSymbol[]
{
new TerminalSymbol("+"),
new TerminalSymbol("*"),
new TerminalSymbol("("),
new TerminalSymbol(")"),
new TerminalSymbol("i"),
};
var grammaticals = new GrammaticalSymbol[]
{
new GrammaticalSymbol("S"),
new GrammaticalSymbol("E"),
new GrammaticalSymbol("E'"),
new GrammaticalSymbol("T"),
new GrammaticalSymbol("T'"),
new GrammaticalSymbol("F"),
};
var rules = new Rule[]
{
// S->E
new Rule(
grammaticals[0],
new Symbol[]
{
grammaticals[1],
}),
// E->TE'
new Rule(
grammaticals[1],
new Symbol[]
{
grammaticals[3],
grammaticals[2],
}),
// E'->+TE'
new Rule(
grammaticals[2],
new Symbol[]
{
terminals[0],
grammaticals[3],
grammaticals[2],
}),
// E'->
new Rule(
grammaticals[2],
new Symbol[] { }),
// T->FT'
new Rule(
grammaticals[3],
new Symbol[]
{
grammaticals[5],
grammaticals[4],
}),
// T'->*FT'
new Rule(
grammaticals[4],
new Symbol[]
{
terminals[1],
grammaticals[5],
grammaticals[4],
}),
// T'->
new Rule(
grammaticals[4],
new Symbol[] { }),
// F->(E)
new Rule(
grammaticals[5],
new Symbol[]
{
terminals[2],
grammaticals[1],
terminals[3],
}),
// F->i
new Rule(
grammaticals[5],
new Symbol[]
{
terminals[4],
}),
};
BaseGrammar = new Grammar(terminals, grammaticals, rules, 0);
RestrictedGrammar = new RestrictedStartSymbolGrammar(BaseGrammar);
ExtendedGrammar = new ExtendedGrammar(RestrictedGrammar);
EpsilonGrammaticals = new HashSet<GrammaticalSymbol>(
new GrammaticalSymbol[] { grammaticals[2], grammaticals[4] });
FirstSets = new Dictionary<GrammaticalSymbol, ISet<TerminalSymbol>>();
// S: (,i
FirstSets[BaseGrammar.Grammaticals[0]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
BaseGrammar.Terminals[2],
BaseGrammar.Terminals[4],
});
// E: (,i
FirstSets[BaseGrammar.Grammaticals[1]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
BaseGrammar.Terminals[2],
BaseGrammar.Terminals[4],
});
// E': +,Eps
FirstSets[BaseGrammar.Grammaticals[2]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
BaseGrammar.Terminals[0],
EpsilonSymbol.Instance,
});
// T: (,i
FirstSets[BaseGrammar.Grammaticals[3]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
BaseGrammar.Terminals[2],
BaseGrammar.Terminals[4],
});
// T': *,Eps
FirstSets[BaseGrammar.Grammaticals[4]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
BaseGrammar.Terminals[1],
EpsilonSymbol.Instance,
});
// F: (,i
FirstSets[BaseGrammar.Grammaticals[5]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
BaseGrammar.Terminals[2],
BaseGrammar.Terminals[4],
});
FollowSets = new Dictionary<GrammaticalSymbol, ISet<TerminalSymbol>>();
// S:
FollowSets[ExtendedGrammar.Grammaticals[0]] = new HashSet<TerminalSymbol>();
// E: ),EoS
FollowSets[ExtendedGrammar.Grammaticals[1]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
ExtendedGrammar.Terminals[3],
ExtendedGrammar.Terminals[5],
});
// E': ),EoS
FollowSets[ExtendedGrammar.Grammaticals[2]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
ExtendedGrammar.Terminals[3],
ExtendedGrammar.Terminals[5],
});
// T: +,),EoS
FollowSets[ExtendedGrammar.Grammaticals[3]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
ExtendedGrammar.Terminals[0],
ExtendedGrammar.Terminals[3],
ExtendedGrammar.Terminals[5],
});
// T': +,),EoS
FollowSets[ExtendedGrammar.Grammaticals[4]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
ExtendedGrammar.Terminals[0],
ExtendedGrammar.Terminals[3],
ExtendedGrammar.Terminals[5],
});
// F: +,*,),EoS
FollowSets[ExtendedGrammar.Grammaticals[5]] = new HashSet<TerminalSymbol>(new TerminalSymbol[]
{
ExtendedGrammar.Terminals[0],
ExtendedGrammar.Terminals[1],
ExtendedGrammar.Terminals[3],
ExtendedGrammar.Terminals[5],
});
//TODO: Use real fixture!!
LR0CanonicalSets = LR0CanonicalSets.Build(RestrictedGrammar);
//TODO: Use real fixtures!!
SLR1ParserTable = SLR1ParserTableBuilder.Build(ExtendedGrammar, LR0CanonicalSets, FollowSets);
}