/*
* Copyright (c) 2012-2017 The ANTLR Project. All rights reserved.
* Use of this file is governed by the BSD 3-clause license that
* can be found in the LICENSE.txt file in the project root.
*/
package org.antlr.v4.runtime.atn;
import org.antlr.v4.runtime.BailErrorStrategy;
import org.antlr.v4.runtime.FailedPredicateException;
import org.antlr.v4.runtime.IntStream;
import org.antlr.v4.runtime.NoViableAltException;
import org.antlr.v4.runtime.Parser;
import org.antlr.v4.runtime.ParserRuleContext;
import org.antlr.v4.runtime.RuleContext;
import org.antlr.v4.runtime.Token;
import org.antlr.v4.runtime.TokenStream;
import org.antlr.v4.runtime.Vocabulary;
import org.antlr.v4.runtime.VocabularyImpl;
import org.antlr.v4.runtime.dfa.DFA;
import org.antlr.v4.runtime.dfa.DFAState;
import org.antlr.v4.runtime.misc.DoubleKeyMap;
import org.antlr.v4.runtime.misc.Interval;
import org.antlr.v4.runtime.misc.IntervalSet;
import org.antlr.v4.runtime.misc.Pair;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.BitSet;
import java.util.Collection;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;
import static org.antlr.v4.runtime.atn.ATNState.BLOCK_END;
The embodiment of the adaptive LL(*), ALL(*), parsing strategy.
The basic complexity of the adaptive strategy makes it harder to understand.
We begin with ATN simulation to build paths in a DFA. Subsequent prediction
requests go through the DFA first. If they reach a state without an edge for
the current symbol, the algorithm fails over to the ATN simulation to
complete the DFA path for the current input (until it finds a conflict state
or uniquely predicting state).
All of that is done without using the outer context because we want to create
a DFA that is not dependent upon the rule invocation stack when we do a
prediction. One DFA works in all contexts. We avoid using context not
necessarily because it's slower, although it can be, but because of the DFA
caching problem. The closure routine only considers the rule invocation stack
created during prediction beginning in the decision rule. For example, if
prediction occurs without invoking another rule's ATN, there are no context
stacks in the configurations. When lack of context leads to a conflict, we
don't know if it's an ambiguity or a weakness in the strong LL(*) parsing
strategy (versus full LL(*)).
When SLL yields a configuration set with conflict, we rewind the input and
retry the ATN simulation, this time using full outer context without adding
to the DFA. Configuration context stacks will be the full invocation stacks
from the start rule. If we get a conflict using full context, then we can
definitively say we have a true ambiguity for that input sequence. If we
don't get a conflict, it implies that the decision is sensitive to the outer
context. (It is not context-sensitive in the sense of context-sensitive
grammars.)
The next time we reach this DFA state with an SLL conflict, through DFA
simulation, we will again retry the ATN simulation using full context mode.
This is slow because we can't save the results and have to "interpret" the
ATN each time we get that input.
CACHING FULL CONTEXT PREDICTIONS
We could cache results from full context to predicted alternative easily and
that saves a lot of time but doesn't work in presence of predicates. The set
of visible predicates from the ATN start state changes depending on the
context, because closure can fall off the end of a rule. I tried to cache
tuples (stack context, semantic context, predicted alt) but it was slower
than interpreting and much more complicated. Also required a huge amount of
memory. The goal is not to create the world's fastest parser anyway. I'd like
to keep this algorithm simple. By launching multiple threads, we can improve
the speed of parsing across a large number of files.
There is no strict ordering between the amount of input used by SLL vs LL,
which makes it really hard to build a cache for full context. Let's say that
we have input A B C that leads to an SLL conflict with full context X. That
implies that using X we might only use A B but we could also use A B C D to
resolve conflict. Input A B C D could predict alternative 1 in one position
in the input and A B C E could predict alternative 2 in another position in
input. The conflicting SLL configurations could still be non-unique in the
full context prediction, which would lead us to requiring more input than the
original A B C. To make a prediction cache work, we have to track the exact
input used during the previous prediction. That amounts to a cache that maps
X to a specific DFA for that context.
Something should be done for left-recursive expression predictions. They are
likely LL(1) + pred eval. Easier to do the whole SLL unless error and retry
with full LL thing Sam does.
AVOIDING FULL CONTEXT PREDICTION
We avoid doing full context retry when the outer context is empty, we did not
dip into the outer context by falling off the end of the decision state rule,
or when we force SLL mode.
As an example of the not dip into outer context case, consider as super
constructor calls versus function calls. One grammar might look like
this:
ctorBody
: '{' superCall? stat* '}'
;
Or, you might see something like
stat
: superCall ';'
| expression ';'
| ...
;
In both cases I believe that no closure operations will dip into the outer
context. In the first case ctorBody in the worst case will stop at the '}'.
In the 2nd case it should stop at the ';'. Both cases should stay within the
entry rule and not dip into the outer context.
PREDICATES
Predicates are always evaluated if present in either SLL or LL both. SLL and
LL simulation deals with predicates differently. SLL collects predicates as
it performs closure operations like ANTLR v3 did. It delays predicate
evaluation until it reaches and accept state. This allows us to cache the SLL
ATN simulation whereas, if we had evaluated predicates on-the-fly during
closure, the DFA state configuration sets would be different and we couldn't
build up a suitable DFA.
When building a DFA accept state during ATN simulation, we evaluate any
predicates and return the sole semantically valid alternative. If there is
more than 1 alternative, we report an ambiguity. If there are 0 alternatives,
we throw an exception. Alternatives without predicates act like they have
true predicates. The simple way to think about it is to strip away all
alternatives with false predicates and choose the minimum alternative that
remains.
When we start in the DFA and reach an accept state that's predicated, we test
those and return the minimum semantically viable alternative. If no
alternatives are viable, we throw an exception.
During full LL ATN simulation, closure always evaluates predicates and
on-the-fly. This is crucial to reducing the configuration set size during
closure. It hits a landmine when parsing with the Java grammar, for example,
without this on-the-fly evaluation.
SHARING DFA
All instances of the same parser share the same decision DFAs through a static field. Each instance gets its own ATN simulator but they share the same decisionToDFA
field. They also share a PredictionContextCache
object that makes sure that all PredictionContext
objects are shared among the DFA states. This makes a big size difference.
THREAD SAFETY
The ParserATNSimulator
locks on the decisionToDFA
field when it adds a new DFA object to that array. addDFAEdge
locks on the DFA for the current decision when setting the DFAState.edges
field. addDFAState
locks on the DFA for the current decision when looking up a DFA state to see if it already exists. We must make sure that all requests to add DFA states that are equivalent result in the same shared DFA object. This is because lots of threads will be trying to update the DFA at once. The addDFAState
method also locks inside the DFA lock but this time on the shared context cache when it rebuilds the configurations' PredictionContext
objects using cached subgraphs/nodes. No other locking occurs, even during DFA simulation. This is safe as long as we can guarantee that all threads referencing s.edge[t]
get the same physical target DFAState
, or null
. Once into the DFA, the DFA simulation does not reference the DFA.states
map. It follows the DFAState.edges
field to new targets. The DFA simulator will either find DFAState.edges
to be null
, to be non-null
and dfa.edges[t]
null, or dfa.edges[t]
to be non-null. The addDFAEdge
method could be racing to set the field but in either case the DFA simulator works; if null
, and requests ATN simulation. It could also race trying to get dfa.edges[t]
, but either way it will work because it's not doing a test and set operation.
Starting with SLL then failing to combined SLL/LL (Two-Stage
Parsing)
Sam pointed out that if SLL does not give a syntax error, then there is no point in doing full LL, which is slower. We only have to try LL if we get a syntax error. For maximum speed, Sam starts the parser set to pure SLL mode with the BailErrorStrategy
:
parser.getInterpreter()
.setPredictionMode
(
PredictionMode.SLL
)
; parser.setErrorHandler
(new BailErrorStrategy
());
If it does not get a syntax error, then we're done. If it does get a syntax
error, we need to retry with the combined SLL/LL strategy.
The reason this works is as follows. If there are no SLL conflicts, then the
grammar is SLL (at least for that input set). If there is an SLL conflict,
the full LL analysis must yield a set of viable alternatives which is a
subset of the alternatives reported by SLL. If the LL set is a singleton,
then the grammar is LL but not SLL. If the LL set is the same size as the SLL
set, the decision is SLL. If the LL set has size > 1, then that decision
is truly ambiguous on the current input. If the LL set is smaller, then the
SLL conflict resolution might choose an alternative that the full LL would
rule out as a possibility based upon better context information. If that's
the case, then the SLL parse will definitely get an error because the full LL
analysis says it's not viable. If SLL conflict resolution chooses an
alternative within the LL set, them both SLL and LL would choose the same
alternative because they both choose the minimum of multiple conflicting
alternatives.
Let's say we have a set of SLL conflicting alternatives {1, 2, 3}
and a smaller LL set called s. If s is {2, 3}
, then SLL parsing will get an error because SLL will pursue alternative 1. If s is {1, 2}
or {1, 3}
then both SLL and LL will choose the same alternative because alternative one is the minimum of either set. If s is {2}
or {3}
then SLL will get a syntax error. If s is {1}
then SLL will succeed.
Of course, if the input is invalid, then we will get an error for sure in
both SLL and LL parsing. Erroneous input will therefore require 2 passes over
the input.
/**
* The embodiment of the adaptive LL(*), ALL(*), parsing strategy.
*
* <p>
* The basic complexity of the adaptive strategy makes it harder to understand.
* We begin with ATN simulation to build paths in a DFA. Subsequent prediction
* requests go through the DFA first. If they reach a state without an edge for
* the current symbol, the algorithm fails over to the ATN simulation to
* complete the DFA path for the current input (until it finds a conflict state
* or uniquely predicting state).</p>
*
* <p>
* All of that is done without using the outer context because we want to create
* a DFA that is not dependent upon the rule invocation stack when we do a
* prediction. One DFA works in all contexts. We avoid using context not
* necessarily because it's slower, although it can be, but because of the DFA
* caching problem. The closure routine only considers the rule invocation stack
* created during prediction beginning in the decision rule. For example, if
* prediction occurs without invoking another rule's ATN, there are no context
* stacks in the configurations. When lack of context leads to a conflict, we
* don't know if it's an ambiguity or a weakness in the strong LL(*) parsing
* strategy (versus full LL(*)).</p>
*
* <p>
* When SLL yields a configuration set with conflict, we rewind the input and
* retry the ATN simulation, this time using full outer context without adding
* to the DFA. Configuration context stacks will be the full invocation stacks
* from the start rule. If we get a conflict using full context, then we can
* definitively say we have a true ambiguity for that input sequence. If we
* don't get a conflict, it implies that the decision is sensitive to the outer
* context. (It is not context-sensitive in the sense of context-sensitive
* grammars.)</p>
*
* <p>
* The next time we reach this DFA state with an SLL conflict, through DFA
* simulation, we will again retry the ATN simulation using full context mode.
* This is slow because we can't save the results and have to "interpret" the
* ATN each time we get that input.</p>
*
* <p>
* <strong>CACHING FULL CONTEXT PREDICTIONS</strong></p>
*
* <p>
* We could cache results from full context to predicted alternative easily and
* that saves a lot of time but doesn't work in presence of predicates. The set
* of visible predicates from the ATN start state changes depending on the
* context, because closure can fall off the end of a rule. I tried to cache
* tuples (stack context, semantic context, predicted alt) but it was slower
* than interpreting and much more complicated. Also required a huge amount of
* memory. The goal is not to create the world's fastest parser anyway. I'd like
* to keep this algorithm simple. By launching multiple threads, we can improve
* the speed of parsing across a large number of files.</p>
*
* <p>
* There is no strict ordering between the amount of input used by SLL vs LL,
* which makes it really hard to build a cache for full context. Let's say that
* we have input A B C that leads to an SLL conflict with full context X. That
* implies that using X we might only use A B but we could also use A B C D to
* resolve conflict. Input A B C D could predict alternative 1 in one position
* in the input and A B C E could predict alternative 2 in another position in
* input. The conflicting SLL configurations could still be non-unique in the
* full context prediction, which would lead us to requiring more input than the
* original A B C. To make a prediction cache work, we have to track the exact
* input used during the previous prediction. That amounts to a cache that maps
* X to a specific DFA for that context.</p>
*
* <p>
* Something should be done for left-recursive expression predictions. They are
* likely LL(1) + pred eval. Easier to do the whole SLL unless error and retry
* with full LL thing Sam does.</p>
*
* <p>
* <strong>AVOIDING FULL CONTEXT PREDICTION</strong></p>
*
* <p>
* We avoid doing full context retry when the outer context is empty, we did not
* dip into the outer context by falling off the end of the decision state rule,
* or when we force SLL mode.</p>
*
* <p>
* As an example of the not dip into outer context case, consider as super
* constructor calls versus function calls. One grammar might look like
* this:</p>
*
* <pre>
* ctorBody
* : '{' superCall? stat* '}'
* ;
* </pre>
*
* <p>
* Or, you might see something like</p>
*
* <pre>
* stat
* : superCall ';'
* | expression ';'
* | ...
* ;
* </pre>
*
* <p>
* In both cases I believe that no closure operations will dip into the outer
* context. In the first case ctorBody in the worst case will stop at the '}'.
* In the 2nd case it should stop at the ';'. Both cases should stay within the
* entry rule and not dip into the outer context.</p>
*
* <p>
* <strong>PREDICATES</strong></p>
*
* <p>
* Predicates are always evaluated if present in either SLL or LL both. SLL and
* LL simulation deals with predicates differently. SLL collects predicates as
* it performs closure operations like ANTLR v3 did. It delays predicate
* evaluation until it reaches and accept state. This allows us to cache the SLL
* ATN simulation whereas, if we had evaluated predicates on-the-fly during
* closure, the DFA state configuration sets would be different and we couldn't
* build up a suitable DFA.</p>
*
* <p>
* When building a DFA accept state during ATN simulation, we evaluate any
* predicates and return the sole semantically valid alternative. If there is
* more than 1 alternative, we report an ambiguity. If there are 0 alternatives,
* we throw an exception. Alternatives without predicates act like they have
* true predicates. The simple way to think about it is to strip away all
* alternatives with false predicates and choose the minimum alternative that
* remains.</p>
*
* <p>
* When we start in the DFA and reach an accept state that's predicated, we test
* those and return the minimum semantically viable alternative. If no
* alternatives are viable, we throw an exception.</p>
*
* <p>
* During full LL ATN simulation, closure always evaluates predicates and
* on-the-fly. This is crucial to reducing the configuration set size during
* closure. It hits a landmine when parsing with the Java grammar, for example,
* without this on-the-fly evaluation.</p>
*
* <p>
* <strong>SHARING DFA</strong></p>
*
* <p>
* All instances of the same parser share the same decision DFAs through a
* static field. Each instance gets its own ATN simulator but they share the
* same {@link #decisionToDFA} field. They also share a
* {@link PredictionContextCache} object that makes sure that all
* {@link PredictionContext} objects are shared among the DFA states. This makes
* a big size difference.</p>
*
* <p>
* <strong>THREAD SAFETY</strong></p>
*
* <p>
* The {@link ParserATNSimulator} locks on the {@link #decisionToDFA} field when
* it adds a new DFA object to that array. {@link #addDFAEdge}
* locks on the DFA for the current decision when setting the
* {@link DFAState#edges} field. {@link #addDFAState} locks on
* the DFA for the current decision when looking up a DFA state to see if it
* already exists. We must make sure that all requests to add DFA states that
* are equivalent result in the same shared DFA object. This is because lots of
* threads will be trying to update the DFA at once. The
* {@link #addDFAState} method also locks inside the DFA lock
* but this time on the shared context cache when it rebuilds the
* configurations' {@link PredictionContext} objects using cached
* subgraphs/nodes. No other locking occurs, even during DFA simulation. This is
* safe as long as we can guarantee that all threads referencing
* {@code s.edge[t]} get the same physical target {@link DFAState}, or
* {@code null}. Once into the DFA, the DFA simulation does not reference the
* {@link DFA#states} map. It follows the {@link DFAState#edges} field to new
* targets. The DFA simulator will either find {@link DFAState#edges} to be
* {@code null}, to be non-{@code null} and {@code dfa.edges[t]} null, or
* {@code dfa.edges[t]} to be non-null. The
* {@link #addDFAEdge} method could be racing to set the field
* but in either case the DFA simulator works; if {@code null}, and requests ATN
* simulation. It could also race trying to get {@code dfa.edges[t]}, but either
* way it will work because it's not doing a test and set operation.</p>
*
* <p>
* <strong>Starting with SLL then failing to combined SLL/LL (Two-Stage
* Parsing)</strong></p>
*
* <p>
* Sam pointed out that if SLL does not give a syntax error, then there is no
* point in doing full LL, which is slower. We only have to try LL if we get a
* syntax error. For maximum speed, Sam starts the parser set to pure SLL
* mode with the {@link BailErrorStrategy}:</p>
*
* <pre>
* parser.{@link Parser#getInterpreter() getInterpreter()}.{@link #setPredictionMode setPredictionMode}{@code (}{@link PredictionMode#SLL}{@code )};
* parser.{@link Parser#setErrorHandler setErrorHandler}(new {@link BailErrorStrategy}());
* </pre>
*
* <p>
* If it does not get a syntax error, then we're done. If it does get a syntax
* error, we need to retry with the combined SLL/LL strategy.</p>
*
* <p>
* The reason this works is as follows. If there are no SLL conflicts, then the
* grammar is SLL (at least for that input set). If there is an SLL conflict,
* the full LL analysis must yield a set of viable alternatives which is a
* subset of the alternatives reported by SLL. If the LL set is a singleton,
* then the grammar is LL but not SLL. If the LL set is the same size as the SLL
* set, the decision is SLL. If the LL set has size > 1, then that decision
* is truly ambiguous on the current input. If the LL set is smaller, then the
* SLL conflict resolution might choose an alternative that the full LL would
* rule out as a possibility based upon better context information. If that's
* the case, then the SLL parse will definitely get an error because the full LL
* analysis says it's not viable. If SLL conflict resolution chooses an
* alternative within the LL set, them both SLL and LL would choose the same
* alternative because they both choose the minimum of multiple conflicting
* alternatives.</p>
*
* <p>
* Let's say we have a set of SLL conflicting alternatives {@code {1, 2, 3}} and
* a smaller LL set called <em>s</em>. If <em>s</em> is {@code {2, 3}}, then SLL
* parsing will get an error because SLL will pursue alternative 1. If
* <em>s</em> is {@code {1, 2}} or {@code {1, 3}} then both SLL and LL will
* choose the same alternative because alternative one is the minimum of either
* set. If <em>s</em> is {@code {2}} or {@code {3}} then SLL will get a syntax
* error. If <em>s</em> is {@code {1}} then SLL will succeed.</p>
*
* <p>
* Of course, if the input is invalid, then we will get an error for sure in
* both SLL and LL parsing. Erroneous input will therefore require 2 passes over
* the input.</p>
*/
public class ParserATNSimulator extends ATNSimulator {
public static final boolean debug = false;
public static final boolean debug_list_atn_decisions = false;
public static final boolean dfa_debug = false;
public static final boolean retry_debug = false;
Just in case this optimization is bad, add an ENV variable to turn it off /** Just in case this optimization is bad, add an ENV variable to turn it off */
public static final boolean TURN_OFF_LR_LOOP_ENTRY_BRANCH_OPT = Boolean.parseBoolean(getSafeEnv("TURN_OFF_LR_LOOP_ENTRY_BRANCH_OPT"));
protected final Parser parser;
public final DFA[] decisionToDFA;
SLL, LL, or LL + exact ambig detection? /** SLL, LL, or LL + exact ambig detection? */
private PredictionMode mode = PredictionMode.LL;
Each prediction operation uses a cache for merge of prediction contexts.
Don't keep around as it wastes huge amounts of memory. DoubleKeyMap
isn't synchronized but we're ok since two threads shouldn't reuse same
parser/atnsim object because it can only handle one input at a time.
This maps graphs a and b to merged result c. (a,b)→c. We can avoid
the merge if we ever see a and b again. Note that (b,a)→c should
also be examined during cache lookup.
/** Each prediction operation uses a cache for merge of prediction contexts.
* Don't keep around as it wastes huge amounts of memory. DoubleKeyMap
* isn't synchronized but we're ok since two threads shouldn't reuse same
* parser/atnsim object because it can only handle one input at a time.
* This maps graphs a and b to merged result c. (a,b)→c. We can avoid
* the merge if we ever see a and b again. Note that (b,a)→c should
* also be examined during cache lookup.
*/
protected DoubleKeyMap<PredictionContext,PredictionContext,PredictionContext> mergeCache;
// LAME globals to avoid parameters!!!!! I need these down deep in predTransition
protected TokenStream _input;
protected int _startIndex;
protected ParserRuleContext _outerContext;
protected DFA _dfa;
Testing only! /** Testing only! */
public ParserATNSimulator(ATN atn, DFA[] decisionToDFA,
PredictionContextCache sharedContextCache)
{
this(null, atn, decisionToDFA, sharedContextCache);
}
public ParserATNSimulator(Parser parser, ATN atn,
DFA[] decisionToDFA,
PredictionContextCache sharedContextCache)
{
super(atn, sharedContextCache);
this.parser = parser;
this.decisionToDFA = decisionToDFA;
// DOTGenerator dot = new DOTGenerator(null);
// System.out.println(dot.getDOT(atn.rules.get(0), parser.getRuleNames()));
// System.out.println(dot.getDOT(atn.rules.get(1), parser.getRuleNames()));
}
@Override
public void reset() {
}
@Override
public void clearDFA() {
for (int d = 0; d < decisionToDFA.length; d++) {
decisionToDFA[d] = new DFA(atn.getDecisionState(d), d);
}
}
public int adaptivePredict(TokenStream input, int decision,
ParserRuleContext outerContext)
{
if ( debug || debug_list_atn_decisions ) {
System.out.println("adaptivePredict decision "+decision+
" exec LA(1)=="+ getLookaheadName(input)+
" line "+input.LT(1).getLine()+":"+input.LT(1).getCharPositionInLine());
}
_input = input;
_startIndex = input.index();
_outerContext = outerContext;
DFA dfa = decisionToDFA[decision];
_dfa = dfa;
int m = input.mark();
int index = _startIndex;
// Now we are certain to have a specific decision's DFA
// But, do we still need an initial state?
try {
DFAState s0;
if (dfa.isPrecedenceDfa()) {
// the start state for a precedence DFA depends on the current
// parser precedence, and is provided by a DFA method.
s0 = dfa.getPrecedenceStartState(parser.getPrecedence());
}
else {
// the start state for a "regular" DFA is just s0
s0 = dfa.s0;
}
if (s0 == null) {
if ( outerContext ==null ) outerContext = ParserRuleContext.EMPTY;
if ( debug || debug_list_atn_decisions ) {
System.out.println("predictATN decision "+ dfa.decision+
" exec LA(1)=="+ getLookaheadName(input) +
", outerContext="+ outerContext.toString(parser));
}
boolean fullCtx = false;
ATNConfigSet s0_closure =
computeStartState(dfa.atnStartState,
ParserRuleContext.EMPTY,
fullCtx);
if (dfa.isPrecedenceDfa()) {
/* If this is a precedence DFA, we use applyPrecedenceFilter
* to convert the computed start state to a precedence start
* state. We then use DFA.setPrecedenceStartState to set the
* appropriate start state for the precedence level rather
* than simply setting DFA.s0.
*/
dfa.s0.configs = s0_closure; // not used for prediction but useful to know start configs anyway
s0_closure = applyPrecedenceFilter(s0_closure);
s0 = addDFAState(dfa, new DFAState(s0_closure));
dfa.setPrecedenceStartState(parser.getPrecedence(), s0);
}
else {
s0 = addDFAState(dfa, new DFAState(s0_closure));
dfa.s0 = s0;
}
}
int alt = execATN(dfa, s0, input, index, outerContext);
if ( debug ) System.out.println("DFA after predictATN: "+ dfa.toString(parser.getVocabulary()));
return alt;
}
finally {
mergeCache = null; // wack cache after each prediction
_dfa = null;
input.seek(index);
input.release(m);
}
}
Performs ATN simulation to compute a predicted alternative based
upon the remaining input, but also updates the DFA cache to avoid
having to traverse the ATN again for the same input sequence.
There are some key conditions we're looking for after computing a new
set of ATN configs (proposed DFA state):
if the set is empty, there is no viable alternative for current symbol
does the state uniquely predict an alternative?
does the state have a conflict that would prevent us from
putting it on the work list?
We also have some key operations to do:
add an edge from previous DFA state to potentially new DFA state, D,
upon current symbol but only if adding to work list, which means in all
cases except no viable alternative (and possibly non-greedy decisions?)
collecting predicates and adding semantic context to DFA accept states
adding rule context to context-sensitive DFA accept states
consuming an input symbol
reporting a conflict
reporting an ambiguity
reporting a context sensitivity
reporting insufficient predicates
cover these cases:
dead end
single alt
single alt + preds
conflict
conflict + preds
/** Performs ATN simulation to compute a predicted alternative based
* upon the remaining input, but also updates the DFA cache to avoid
* having to traverse the ATN again for the same input sequence.
There are some key conditions we're looking for after computing a new
set of ATN configs (proposed DFA state):
* if the set is empty, there is no viable alternative for current symbol
* does the state uniquely predict an alternative?
* does the state have a conflict that would prevent us from
putting it on the work list?
We also have some key operations to do:
* add an edge from previous DFA state to potentially new DFA state, D,
upon current symbol but only if adding to work list, which means in all
cases except no viable alternative (and possibly non-greedy decisions?)
* collecting predicates and adding semantic context to DFA accept states
* adding rule context to context-sensitive DFA accept states
* consuming an input symbol
* reporting a conflict
* reporting an ambiguity
* reporting a context sensitivity
* reporting insufficient predicates
cover these cases:
dead end
single alt
single alt + preds
conflict
conflict + preds
*/
protected int execATN(DFA dfa, DFAState s0,
TokenStream input, int startIndex,
ParserRuleContext outerContext)
{
if ( debug || debug_list_atn_decisions) {
System.out.println("execATN decision "+dfa.decision+
" exec LA(1)=="+ getLookaheadName(input)+
" line "+input.LT(1).getLine()+":"+input.LT(1).getCharPositionInLine());
}
DFAState previousD = s0;
if ( debug ) System.out.println("s0 = "+s0);
int t = input.LA(1);
while (true) { // while more work
DFAState D = getExistingTargetState(previousD, t);
if (D == null) {
D = computeTargetState(dfa, previousD, t);
}
if (D == ERROR) {
// if any configs in previous dipped into outer context, that
// means that input up to t actually finished entry rule
// at least for SLL decision. Full LL doesn't dip into outer
// so don't need special case.
// We will get an error no matter what so delay until after
// decision; better error message. Also, no reachable target
// ATN states in SLL implies LL will also get nowhere.
// If conflict in states that dip out, choose min since we
// will get error no matter what.
NoViableAltException e = noViableAlt(input, outerContext, previousD.configs, startIndex);
input.seek(startIndex);
int alt = getSynValidOrSemInvalidAltThatFinishedDecisionEntryRule(previousD.configs, outerContext);
if ( alt!=ATN.INVALID_ALT_NUMBER ) {
return alt;
}
throw e;
}
if ( D.requiresFullContext && mode != PredictionMode.SLL ) {
// IF PREDS, MIGHT RESOLVE TO SINGLE ALT => SLL (or syntax error)
BitSet conflictingAlts = D.configs.conflictingAlts;
if ( D.predicates!=null ) {
if ( debug ) System.out.println("DFA state has preds in DFA sim LL failover");
int conflictIndex = input.index();
if (conflictIndex != startIndex) {
input.seek(startIndex);
}
conflictingAlts = evalSemanticContext(D.predicates, outerContext, true);
if ( conflictingAlts.cardinality()==1 ) {
if ( debug ) System.out.println("Full LL avoided");
return conflictingAlts.nextSetBit(0);
}
if (conflictIndex != startIndex) {
// restore the index so reporting the fallback to full
// context occurs with the index at the correct spot
input.seek(conflictIndex);
}
}
if ( dfa_debug ) System.out.println("ctx sensitive state "+outerContext+" in "+D);
boolean fullCtx = true;
ATNConfigSet s0_closure =
computeStartState(dfa.atnStartState, outerContext,
fullCtx);
reportAttemptingFullContext(dfa, conflictingAlts, D.configs, startIndex, input.index());
int alt = execATNWithFullContext(dfa, D, s0_closure,
input, startIndex,
outerContext);
return alt;
}
if ( D.isAcceptState ) {
if (D.predicates == null) {
return D.prediction;
}
int stopIndex = input.index();
input.seek(startIndex);
BitSet alts = evalSemanticContext(D.predicates, outerContext, true);
switch (alts.cardinality()) {
case 0:
throw noViableAlt(input, outerContext, D.configs, startIndex);
case 1:
return alts.nextSetBit(0);
default:
// report ambiguity after predicate evaluation to make sure the correct
// set of ambig alts is reported.
reportAmbiguity(dfa, D, startIndex, stopIndex, false, alts, D.configs);
return alts.nextSetBit(0);
}
}
previousD = D;
if (t != IntStream.EOF) {
input.consume();
t = input.LA(1);
}
}
}
Get an existing target state for an edge in the DFA. If the target state for the edge has not yet been computed or is otherwise not available, this method returns null
. Params: - previousD – The current DFA state
- t – The next input symbol
Returns: The existing target DFA state for the given input symbol t
, or null
if the target state for this edge is not already cached
/**
* Get an existing target state for an edge in the DFA. If the target state
* for the edge has not yet been computed or is otherwise not available,
* this method returns {@code null}.
*
* @param previousD The current DFA state
* @param t The next input symbol
* @return The existing target DFA state for the given input symbol
* {@code t}, or {@code null} if the target state for this edge is not
* already cached
*/
protected DFAState getExistingTargetState(DFAState previousD, int t) {
DFAState[] edges = previousD.edges;
if (edges == null || t + 1 < 0 || t + 1 >= edges.length) {
return null;
}
return edges[t + 1];
}
Compute a target state for an edge in the DFA, and attempt to add the
computed state and corresponding edge to the DFA.
Params: - dfa – The DFA
- previousD – The current DFA state
- t – The next input symbol
Returns: The computed target DFA state for the given input symbol t
. If t
does not lead to a valid DFA state, this method returns ATNSimulator.ERROR
.
/**
* Compute a target state for an edge in the DFA, and attempt to add the
* computed state and corresponding edge to the DFA.
*
* @param dfa The DFA
* @param previousD The current DFA state
* @param t The next input symbol
*
* @return The computed target DFA state for the given input symbol
* {@code t}. If {@code t} does not lead to a valid DFA state, this method
* returns {@link #ERROR}.
*/
protected DFAState computeTargetState(DFA dfa, DFAState previousD, int t) {
ATNConfigSet reach = computeReachSet(previousD.configs, t, false);
if ( reach==null ) {
addDFAEdge(dfa, previousD, t, ERROR);
return ERROR;
}
// create new target state; we'll add to DFA after it's complete
DFAState D = new DFAState(reach);
int predictedAlt = getUniqueAlt(reach);
if ( debug ) {
Collection<BitSet> altSubSets = PredictionMode.getConflictingAltSubsets(reach);
System.out.println("SLL altSubSets="+altSubSets+
", configs="+reach+
", predict="+predictedAlt+", allSubsetsConflict="+
PredictionMode.allSubsetsConflict(altSubSets)+", conflictingAlts="+
getConflictingAlts(reach));
}
if ( predictedAlt!=ATN.INVALID_ALT_NUMBER ) {
// NO CONFLICT, UNIQUELY PREDICTED ALT
D.isAcceptState = true;
D.configs.uniqueAlt = predictedAlt;
D.prediction = predictedAlt;
}
else if ( PredictionMode.hasSLLConflictTerminatingPrediction(mode, reach) ) {
// MORE THAN ONE VIABLE ALTERNATIVE
D.configs.conflictingAlts = getConflictingAlts(reach);
D.requiresFullContext = true;
// in SLL-only mode, we will stop at this state and return the minimum alt
D.isAcceptState = true;
D.prediction = D.configs.conflictingAlts.nextSetBit(0);
}
if ( D.isAcceptState && D.configs.hasSemanticContext ) {
predicateDFAState(D, atn.getDecisionState(dfa.decision));
if (D.predicates != null) {
D.prediction = ATN.INVALID_ALT_NUMBER;
}
}
// all adds to dfa are done after we've created full D state
D = addDFAEdge(dfa, previousD, t, D);
return D;
}
protected void predicateDFAState(DFAState dfaState, DecisionState decisionState) {
// We need to test all predicates, even in DFA states that
// uniquely predict alternative.
int nalts = decisionState.getNumberOfTransitions();
// Update DFA so reach becomes accept state with (predicate,alt)
// pairs if preds found for conflicting alts
BitSet altsToCollectPredsFrom = getConflictingAltsOrUniqueAlt(dfaState.configs);
SemanticContext[] altToPred = getPredsForAmbigAlts(altsToCollectPredsFrom, dfaState.configs, nalts);
if ( altToPred!=null ) {
dfaState.predicates = getPredicatePredictions(altsToCollectPredsFrom, altToPred);
dfaState.prediction = ATN.INVALID_ALT_NUMBER; // make sure we use preds
}
else {
// There are preds in configs but they might go away
// when OR'd together like {p}? || NONE == NONE. If neither
// alt has preds, resolve to min alt
dfaState.prediction = altsToCollectPredsFrom.nextSetBit(0);
}
}
// comes back with reach.uniqueAlt set to a valid alt
protected int execATNWithFullContext(DFA dfa,
DFAState D, // how far we got in SLL DFA before failing over
ATNConfigSet s0,
TokenStream input, int startIndex,
ParserRuleContext outerContext)
{
if ( debug || debug_list_atn_decisions ) {
System.out.println("execATNWithFullContext "+s0);
}
boolean fullCtx = true;
boolean foundExactAmbig = false;
ATNConfigSet reach = null;
ATNConfigSet previous = s0;
input.seek(startIndex);
int t = input.LA(1);
int predictedAlt;
while (true) { // while more work
// System.out.println("LL REACH "+getLookaheadName(input)+
// " from configs.size="+previous.size()+
// " line "+input.LT(1).getLine()+":"+input.LT(1).getCharPositionInLine());
reach = computeReachSet(previous, t, fullCtx);
if ( reach==null ) {
// if any configs in previous dipped into outer context, that
// means that input up to t actually finished entry rule
// at least for LL decision. Full LL doesn't dip into outer
// so don't need special case.
// We will get an error no matter what so delay until after
// decision; better error message. Also, no reachable target
// ATN states in SLL implies LL will also get nowhere.
// If conflict in states that dip out, choose min since we
// will get error no matter what.
NoViableAltException e = noViableAlt(input, outerContext, previous, startIndex);
input.seek(startIndex);
int alt = getSynValidOrSemInvalidAltThatFinishedDecisionEntryRule(previous, outerContext);
if ( alt!=ATN.INVALID_ALT_NUMBER ) {
return alt;
}
throw e;
}
Collection<BitSet> altSubSets = PredictionMode.getConflictingAltSubsets(reach);
if ( debug ) {
System.out.println("LL altSubSets="+altSubSets+
", predict="+PredictionMode.getUniqueAlt(altSubSets)+
", resolvesToJustOneViableAlt="+
PredictionMode.resolvesToJustOneViableAlt(altSubSets));
}
// System.out.println("altSubSets: "+altSubSets);
// System.err.println("reach="+reach+", "+reach.conflictingAlts);
reach.uniqueAlt = getUniqueAlt(reach);
// unique prediction?
if ( reach.uniqueAlt!=ATN.INVALID_ALT_NUMBER ) {
predictedAlt = reach.uniqueAlt;
break;
}
if ( mode != PredictionMode.LL_EXACT_AMBIG_DETECTION ) {
predictedAlt = PredictionMode.resolvesToJustOneViableAlt(altSubSets);
if ( predictedAlt != ATN.INVALID_ALT_NUMBER ) {
break;
}
}
else {
// In exact ambiguity mode, we never try to terminate early.
// Just keeps scarfing until we know what the conflict is
if ( PredictionMode.allSubsetsConflict(altSubSets) &&
PredictionMode.allSubsetsEqual(altSubSets) )
{
foundExactAmbig = true;
predictedAlt = PredictionMode.getSingleViableAlt(altSubSets);
break;
}
// else there are multiple non-conflicting subsets or
// we're not sure what the ambiguity is yet.
// So, keep going.
}
previous = reach;
if (t != IntStream.EOF) {
input.consume();
t = input.LA(1);
}
}
// If the configuration set uniquely predicts an alternative,
// without conflict, then we know that it's a full LL decision
// not SLL.
if ( reach.uniqueAlt != ATN.INVALID_ALT_NUMBER ) {
reportContextSensitivity(dfa, predictedAlt, reach, startIndex, input.index());
return predictedAlt;
}
// We do not check predicates here because we have checked them
// on-the-fly when doing full context prediction.
/*
In non-exact ambiguity detection mode, we might actually be able to
detect an exact ambiguity, but I'm not going to spend the cycles
needed to check. We only emit ambiguity warnings in exact ambiguity
mode.
For example, we might know that we have conflicting configurations.
But, that does not mean that there is no way forward without a
conflict. It's possible to have nonconflicting alt subsets as in:
LL altSubSets=[{1, 2}, {1, 2}, {1}, {1, 2}]
from
[(17,1,[5 $]), (13,1,[5 10 $]), (21,1,[5 10 $]), (11,1,[$]),
(13,2,[5 10 $]), (21,2,[5 10 $]), (11,2,[$])]
In this case, (17,1,[5 $]) indicates there is some next sequence that
would resolve this without conflict to alternative 1. Any other viable
next sequence, however, is associated with a conflict. We stop
looking for input because no amount of further lookahead will alter
the fact that we should predict alternative 1. We just can't say for
sure that there is an ambiguity without looking further.
*/
reportAmbiguity(dfa, D, startIndex, input.index(), foundExactAmbig,
reach.getAlts(), reach);
return predictedAlt;
}
protected ATNConfigSet computeReachSet(ATNConfigSet closure, int t,
boolean fullCtx)
{
if ( debug )
System.out.println("in computeReachSet, starting closure: " + closure);
if (mergeCache == null) {
mergeCache = new DoubleKeyMap<PredictionContext, PredictionContext, PredictionContext>();
}
ATNConfigSet intermediate = new ATNConfigSet(fullCtx);
/* Configurations already in a rule stop state indicate reaching the end
* of the decision rule (local context) or end of the start rule (full
* context). Once reached, these configurations are never updated by a
* closure operation, so they are handled separately for the performance
* advantage of having a smaller intermediate set when calling closure.
*
* For full-context reach operations, separate handling is required to
* ensure that the alternative matching the longest overall sequence is
* chosen when multiple such configurations can match the input.
*/
List<ATNConfig> skippedStopStates = null;
// First figure out where we can reach on input t
for (ATNConfig c : closure) {
if ( debug ) System.out.println("testing "+getTokenName(t)+" at "+c.toString());
if (c.state instanceof RuleStopState) {
assert c.context.isEmpty();
if (fullCtx || t == IntStream.EOF) {
if (skippedStopStates == null) {
skippedStopStates = new ArrayList<ATNConfig>();
}
skippedStopStates.add(c);
}
continue;
}
int n = c.state.getNumberOfTransitions();
for (int ti=0; ti<n; ti++) { // for each transition
Transition trans = c.state.transition(ti);
ATNState target = getReachableTarget(trans, t);
if ( target!=null ) {
intermediate.add(new ATNConfig(c, target), mergeCache);
}
}
}
// Now figure out where the reach operation can take us...
ATNConfigSet reach = null;
/* This block optimizes the reach operation for intermediate sets which
* trivially indicate a termination state for the overall
* adaptivePredict operation.
*
* The conditions assume that intermediate
* contains all configurations relevant to the reach set, but this
* condition is not true when one or more configurations have been
* withheld in skippedStopStates, or when the current symbol is EOF.
*/
if (skippedStopStates == null && t != Token.EOF) {
if ( intermediate.size()==1 ) {
// Don't pursue the closure if there is just one state.
// It can only have one alternative; just add to result
// Also don't pursue the closure if there is unique alternative
// among the configurations.
reach = intermediate;
}
else if ( getUniqueAlt(intermediate)!=ATN.INVALID_ALT_NUMBER ) {
// Also don't pursue the closure if there is unique alternative
// among the configurations.
reach = intermediate;
}
}
/* If the reach set could not be trivially determined, perform a closure
* operation on the intermediate set to compute its initial value.
*/
if (reach == null) {
reach = new ATNConfigSet(fullCtx);
Set<ATNConfig> closureBusy = new HashSet<ATNConfig>();
boolean treatEofAsEpsilon = t == Token.EOF;
for (ATNConfig c : intermediate) {
closure(c, reach, closureBusy, false, fullCtx, treatEofAsEpsilon);
}
}
if (t == IntStream.EOF) {
/* After consuming EOF no additional input is possible, so we are
* only interested in configurations which reached the end of the
* decision rule (local context) or end of the start rule (full
* context). Update reach to contain only these configurations. This
* handles both explicit EOF transitions in the grammar and implicit
* EOF transitions following the end of the decision or start rule.
*
* When reach==intermediate, no closure operation was performed. In
* this case, removeAllConfigsNotInRuleStopState needs to check for
* reachable rule stop states as well as configurations already in
* a rule stop state.
*
* This is handled before the configurations in skippedStopStates,
* because any configurations potentially added from that list are
* already guaranteed to meet this condition whether or not it's
* required.
*/
reach = removeAllConfigsNotInRuleStopState(reach, reach == intermediate);
}
/* If skippedStopStates is not null, then it contains at least one
* configuration. For full-context reach operations, these
* configurations reached the end of the start rule, in which case we
* only add them back to reach if no configuration during the current
* closure operation reached such a state. This ensures adaptivePredict
* chooses an alternative matching the longest overall sequence when
* multiple alternatives are viable.
*/
if (skippedStopStates != null && (!fullCtx || !PredictionMode.hasConfigInRuleStopState(reach))) {
assert !skippedStopStates.isEmpty();
for (ATNConfig c : skippedStopStates) {
reach.add(c, mergeCache);
}
}
if ( reach.isEmpty() ) return null;
return reach;
}
Return a configuration set containing only the configurations from configs
which are in a RuleStopState
. If all configurations in configs
are already in a rule stop state, this method simply returns configs
. When lookToEndOfRule
is true, this method uses ATN.nextTokens
for each configuration in configs
which is not already in a rule stop state to see if a rule stop state is reachable from the configuration via epsilon-only transitions.
Params: - configs – the configuration set to update
- lookToEndOfRule – when true, this method checks for rule stop states reachable by epsilon-only transitions from each configuration in
configs
.
Returns: configs
if all configurations in configs
are in a rule stop state, otherwise return a new configuration set containing only the configurations from configs
which are in a rule stop state
/**
* Return a configuration set containing only the configurations from
* {@code configs} which are in a {@link RuleStopState}. If all
* configurations in {@code configs} are already in a rule stop state, this
* method simply returns {@code configs}.
*
* <p>When {@code lookToEndOfRule} is true, this method uses
* {@link ATN#nextTokens} for each configuration in {@code configs} which is
* not already in a rule stop state to see if a rule stop state is reachable
* from the configuration via epsilon-only transitions.</p>
*
* @param configs the configuration set to update
* @param lookToEndOfRule when true, this method checks for rule stop states
* reachable by epsilon-only transitions from each configuration in
* {@code configs}.
*
* @return {@code configs} if all configurations in {@code configs} are in a
* rule stop state, otherwise return a new configuration set containing only
* the configurations from {@code configs} which are in a rule stop state
*/
protected ATNConfigSet removeAllConfigsNotInRuleStopState(ATNConfigSet configs, boolean lookToEndOfRule) {
if (PredictionMode.allConfigsInRuleStopStates(configs)) {
return configs;
}
ATNConfigSet result = new ATNConfigSet(configs.fullCtx);
for (ATNConfig config : configs) {
if (config.state instanceof RuleStopState) {
result.add(config, mergeCache);
continue;
}
if (lookToEndOfRule && config.state.onlyHasEpsilonTransitions()) {
IntervalSet nextTokens = atn.nextTokens(config.state);
if (nextTokens.contains(Token.EPSILON)) {
ATNState endOfRuleState = atn.ruleToStopState[config.state.ruleIndex];
result.add(new ATNConfig(config, endOfRuleState), mergeCache);
}
}
}
return result;
}
protected ATNConfigSet computeStartState(ATNState p,
RuleContext ctx,
boolean fullCtx)
{
// always at least the implicit call to start rule
PredictionContext initialContext = PredictionContext.fromRuleContext(atn, ctx);
ATNConfigSet configs = new ATNConfigSet(fullCtx);
for (int i=0; i<p.getNumberOfTransitions(); i++) {
ATNState target = p.transition(i).target;
ATNConfig c = new ATNConfig(target, i+1, initialContext);
Set<ATNConfig> closureBusy = new HashSet<ATNConfig>();
closure(c, configs, closureBusy, true, fullCtx, false);
}
return configs;
}
/* parrt internal source braindump that doesn't mess up
* external API spec.
context-sensitive in that they can only be properly evaluated
in the context of the proper prec argument. Without pruning,
these predicates are normal predicates evaluated when we reach
conflict state (or unique prediction). As we cannot evaluate
these predicates out of context, the resulting conflict leads
to full LL evaluation and nonlinear prediction which shows up
very clearly with fairly large expressions.
Example grammar:
e : e '*' e
| e '+' e
| INT
;
We convert that to the following:
e[int prec]
: INT
( {3>=prec}? '*' e[4]
| {2>=prec}? '+' e[3]
)*
;
The (..)* loop has a decision for the inner block as well as
an enter or exit decision, which is what concerns us here. At
the 1st + of input 1+2+3, the loop entry sees both predicates
and the loop exit also sees both predicates by falling off the
edge of e. This is because we have no stack information with
SLL and find the follow of e, which will hit the return states
inside the loop after e[4] and e[3], which brings it back to
the enter or exit decision. In this case, we know that we
cannot evaluate those predicates because we have fallen off
the edge of the stack and will in general not know which prec
parameter is the right one to use in the predicate.
Because we have special information, that these are precedence
predicates, we can resolve them without failing over to full
LL despite their context sensitive nature. We make an
assumption that prec[-1] <= prec[0], meaning that the current
precedence level is greater than or equal to the precedence
level of recursive invocations above us in the stack. For
example, if predicate {3>=prec}? is true of the current prec,
then one option is to enter the loop to match it now. The
other option is to exit the loop and the left recursive rule
to match the current operator in rule invocation further up
the stack. But, we know that all of those prec are lower or
the same value and so we can decide to enter the loop instead
of matching it later. That means we can strip out the other
configuration for the exit branch.
So imagine we have (14,1,$,{2>=prec}?) and then
(14,2,$-dipsIntoOuterContext,{2>=prec}?). The optimization
allows us to collapse these two configurations. We know that
if {2>=prec}? is true for the current prec parameter, it will
also be true for any prec from an invoking e call, indicated
by dipsIntoOuterContext. As the predicates are both true, we
have the option to evaluate them early in the decision start
state. We do this by stripping both predicates and choosing to
enter the loop as it is consistent with the notion of operator
precedence. It's also how the full LL conflict resolution
would work.
The solution requires a different DFA start state for each
precedence level.
The basic filter mechanism is to remove configurations of the
form (p, 2, pi) if (p, 1, pi) exists for the same p and pi. In
other words, for the same ATN state and predicate context,
remove any configuration associated with an exit branch if
there is a configuration associated with the enter branch.
It's also the case that the filter evaluates precedence
predicates and resolves conflicts according to precedence
levels. For example, for input 1+2+3 at the first +, we see
prediction filtering
[(11,1,[$],{3>=prec}?), (14,1,[$],{2>=prec}?), (5,2,[$],up=1),
(11,2,[$],up=1), (14,2,[$],up=1)],hasSemanticContext=true,dipsIntoOuterContext
to
[(11,1,[$]), (14,1,[$]), (5,2,[$],up=1)],dipsIntoOuterContext
This filters because {3>=prec}? evals to true and collapses
(11,1,[$],{3>=prec}?) and (11,2,[$],up=1) since early conflict
resolution based upon rules of operator precedence fits with
our usual match first alt upon conflict.
We noticed a problem where a recursive call resets precedence
to 0. Sam's fix: each config has flag indicating if it has
returned from an expr[0] call. then just don't filter any
config with that flag set. flag is carried along in
closure(). so to avoid adding field, set bit just under sign
bit of dipsIntoOuterContext (SUPPRESS_PRECEDENCE_FILTER).
With the change you filter "unless (p, 2, pi) was reached
after leaving the rule stop state of the LR rule containing
state p, corresponding to a rule invocation with precedence
level 0"
*/
This method transforms the start state computed by computeStartState
to the special start state used by a precedence DFA for a particular precedence value. The transformation process applies the following changes to the start state's configuration set.
- Evaluate the precedence predicates for each configuration using
SemanticContext.evalPrecedence
.
- When
ATNConfig.isPrecedenceFilterSuppressed
is false
, remove all configurations which predict an alternative greater than 1, for which another configuration that predicts alternative 1 is in the same ATN state with the same prediction context. This transformation is valid for the following reasons:
- The closure block cannot contain any epsilon transitions which bypass
the body of the closure, so all states reachable via alternative 1 are
part of the precedence alternatives of the transformed left-recursive
rule.
- The "primary" portion of a left recursive rule cannot contain an epsilon transition, so the only way an alternative other than 1 can exist in a state that is also reachable via alternative 1 is by nesting calls to the left-recursive rule, with the outer calls not being at the preferred precedence level. The
ATNConfig.isPrecedenceFilterSuppressed
property marks ATN configurations which do not meet this condition, and therefore are not eligible for elimination during the filtering process.
The prediction context must be considered by this filter to address
situations like the following.
grammar TA;
prog: statement* EOF;
statement: letterA | statement letterA 'b' ;
letterA: 'a';
If the above grammar, the ATN state immediately before the token reference 'a'
in letterA
is reachable from the left edge of both the primary and closure blocks of the left-recursive rule statement
. The prediction context associated with each of these configurations distinguishes between them, and prevents the alternative which stepped out to prog
(and then back in to statement
from being eliminated by the filter.
Params: - configs – The configuration set computed by
computeStartState
as the start state for the DFA.
Returns: The transformed configuration set representing the start state for a precedence DFA at a particular precedence level (determined by calling Parser.getPrecedence
).
/**
* This method transforms the start state computed by
* {@link #computeStartState} to the special start state used by a
* precedence DFA for a particular precedence value. The transformation
* process applies the following changes to the start state's configuration
* set.
*
* <ol>
* <li>Evaluate the precedence predicates for each configuration using
* {@link SemanticContext#evalPrecedence}.</li>
* <li>When {@link ATNConfig#isPrecedenceFilterSuppressed} is {@code false},
* remove all configurations which predict an alternative greater than 1,
* for which another configuration that predicts alternative 1 is in the
* same ATN state with the same prediction context. This transformation is
* valid for the following reasons:
* <ul>
* <li>The closure block cannot contain any epsilon transitions which bypass
* the body of the closure, so all states reachable via alternative 1 are
* part of the precedence alternatives of the transformed left-recursive
* rule.</li>
* <li>The "primary" portion of a left recursive rule cannot contain an
* epsilon transition, so the only way an alternative other than 1 can exist
* in a state that is also reachable via alternative 1 is by nesting calls
* to the left-recursive rule, with the outer calls not being at the
* preferred precedence level. The
* {@link ATNConfig#isPrecedenceFilterSuppressed} property marks ATN
* configurations which do not meet this condition, and therefore are not
* eligible for elimination during the filtering process.</li>
* </ul>
* </li>
* </ol>
*
* <p>
* The prediction context must be considered by this filter to address
* situations like the following.
* </p>
* <code>
* <pre>
* grammar TA;
* prog: statement* EOF;
* statement: letterA | statement letterA 'b' ;
* letterA: 'a';
* </pre>
* </code>
* <p>
* If the above grammar, the ATN state immediately before the token
* reference {@code 'a'} in {@code letterA} is reachable from the left edge
* of both the primary and closure blocks of the left-recursive rule
* {@code statement}. The prediction context associated with each of these
* configurations distinguishes between them, and prevents the alternative
* which stepped out to {@code prog} (and then back in to {@code statement}
* from being eliminated by the filter.
* </p>
*
* @param configs The configuration set computed by
* {@link #computeStartState} as the start state for the DFA.
* @return The transformed configuration set representing the start state
* for a precedence DFA at a particular precedence level (determined by
* calling {@link Parser#getPrecedence}).
*/
protected ATNConfigSet applyPrecedenceFilter(ATNConfigSet configs) {
Map<Integer, PredictionContext> statesFromAlt1 = new HashMap<Integer, PredictionContext>();
ATNConfigSet configSet = new ATNConfigSet(configs.fullCtx);
for (ATNConfig config : configs) {
// handle alt 1 first
if (config.alt != 1) {
continue;
}
SemanticContext updatedContext = config.semanticContext.evalPrecedence(parser, _outerContext);
if (updatedContext == null) {
// the configuration was eliminated
continue;
}
statesFromAlt1.put(config.state.stateNumber, config.context);
if (updatedContext != config.semanticContext) {
configSet.add(new ATNConfig(config, updatedContext), mergeCache);
}
else {
configSet.add(config, mergeCache);
}
}
for (ATNConfig config : configs) {
if (config.alt == 1) {
// already handled
continue;
}
if (!config.isPrecedenceFilterSuppressed()) {
/* In the future, this elimination step could be updated to also
* filter the prediction context for alternatives predicting alt>1
* (basically a graph subtraction algorithm).
*/
PredictionContext context = statesFromAlt1.get(config.state.stateNumber);
if (context != null && context.equals(config.context)) {
// eliminated
continue;
}
}
configSet.add(config, mergeCache);
}
return configSet;
}
protected ATNState getReachableTarget(Transition trans, int ttype) {
if (trans.matches(ttype, 0, atn.maxTokenType)) {
return trans.target;
}
return null;
}
protected SemanticContext[] getPredsForAmbigAlts(BitSet ambigAlts,
ATNConfigSet configs,
int nalts)
{
// REACH=[1|1|[]|0:0, 1|2|[]|0:1]
/* altToPred starts as an array of all null contexts. The entry at index i
* corresponds to alternative i. altToPred[i] may have one of three values:
* 1. null: no ATNConfig c is found such that c.alt==i
* 2. SemanticContext.NONE: At least one ATNConfig c exists such that
* c.alt==i and c.semanticContext==SemanticContext.NONE. In other words,
* alt i has at least one unpredicated config.
* 3. Non-NONE Semantic Context: There exists at least one, and for all
* ATNConfig c such that c.alt==i, c.semanticContext!=SemanticContext.NONE.
*
* From this, it is clear that NONE||anything==NONE.
*/
SemanticContext[] altToPred = new SemanticContext[nalts + 1];
for (ATNConfig c : configs) {
if ( ambigAlts.get(c.alt) ) {
altToPred[c.alt] = SemanticContext.or(altToPred[c.alt], c.semanticContext);
}
}
int nPredAlts = 0;
for (int i = 1; i <= nalts; i++) {
if (altToPred[i] == null) {
altToPred[i] = SemanticContext.NONE;
}
else if (altToPred[i] != SemanticContext.NONE) {
nPredAlts++;
}
}
// // Optimize away p||p and p&&p TODO: optimize() was a no-op
// for (int i = 0; i < altToPred.length; i++) {
// altToPred[i] = altToPred[i].optimize();
// }
// nonambig alts are null in altToPred
if ( nPredAlts==0 ) altToPred = null;
if ( debug ) System.out.println("getPredsForAmbigAlts result "+Arrays.toString(altToPred));
return altToPred;
}
protected DFAState.PredPrediction[] getPredicatePredictions(BitSet ambigAlts,
SemanticContext[] altToPred)
{
List<DFAState.PredPrediction> pairs = new ArrayList<DFAState.PredPrediction>();
boolean containsPredicate = false;
for (int i = 1; i < altToPred.length; i++) {
SemanticContext pred = altToPred[i];
// unpredicated is indicated by SemanticContext.NONE
assert pred != null;
if (ambigAlts!=null && ambigAlts.get(i)) {
pairs.add(new DFAState.PredPrediction(pred, i));
}
if ( pred!=SemanticContext.NONE ) containsPredicate = true;
}
if ( !containsPredicate ) {
return null;
}
// System.out.println(Arrays.toString(altToPred)+"->"+pairs);
return pairs.toArray(new DFAState.PredPrediction[pairs.size()]);
}
This method is used to improve the localization of error messages by choosing an alternative rather than throwing a NoViableAltException
in particular prediction scenarios where the ATNSimulator.ERROR
state was reached during ATN simulation. The default implementation of this method uses the following algorithm to identify an ATN configuration which successfully parsed the decision entry rule. Choosing such an alternative ensures that the ParserRuleContext
returned by the calling rule will be complete and valid, and the syntax error will be reported later at a more localized location.
- If a syntactically valid path or paths reach the end of the decision rule and
they are semantically valid if predicated, return the min associated alt.
- Else, if a semantically invalid but syntactically valid path exist
or paths exist, return the minimum associated alt.
- Otherwise, return
ATN.INVALID_ALT_NUMBER
.
In some scenarios, the algorithm described above could predict an alternative which will result in a FailedPredicateException
in the parser. Specifically, this could occur if the only configuration capable of successfully parsing to the end of the decision rule is blocked by a semantic predicate. By choosing this alternative within adaptivePredict
instead of throwing a NoViableAltException
, the resulting FailedPredicateException
in the parser will identify the specific predicate which is preventing the parser from successfully parsing the decision rule, which helps developers identify and correct logic errors in semantic predicates.
Params: - configs – The ATN configurations which were valid immediately before the
ATNSimulator.ERROR
state was reached - outerContext – The is the \gamma_0 initial parser context from the paper
or the parser stack at the instant before prediction commences.
Returns: The value to return from adaptivePredict
, or ATN.INVALID_ALT_NUMBER
if a suitable alternative was not identified and adaptivePredict
should report an error instead.
/**
* This method is used to improve the localization of error messages by
* choosing an alternative rather than throwing a
* {@link NoViableAltException} in particular prediction scenarios where the
* {@link #ERROR} state was reached during ATN simulation.
*
* <p>
* The default implementation of this method uses the following
* algorithm to identify an ATN configuration which successfully parsed the
* decision entry rule. Choosing such an alternative ensures that the
* {@link ParserRuleContext} returned by the calling rule will be complete
* and valid, and the syntax error will be reported later at a more
* localized location.</p>
*
* <ul>
* <li>If a syntactically valid path or paths reach the end of the decision rule and
* they are semantically valid if predicated, return the min associated alt.</li>
* <li>Else, if a semantically invalid but syntactically valid path exist
* or paths exist, return the minimum associated alt.
* </li>
* <li>Otherwise, return {@link ATN#INVALID_ALT_NUMBER}.</li>
* </ul>
*
* <p>
* In some scenarios, the algorithm described above could predict an
* alternative which will result in a {@link FailedPredicateException} in
* the parser. Specifically, this could occur if the <em>only</em> configuration
* capable of successfully parsing to the end of the decision rule is
* blocked by a semantic predicate. By choosing this alternative within
* {@link #adaptivePredict} instead of throwing a
* {@link NoViableAltException}, the resulting
* {@link FailedPredicateException} in the parser will identify the specific
* predicate which is preventing the parser from successfully parsing the
* decision rule, which helps developers identify and correct logic errors
* in semantic predicates.
* </p>
*
* @param configs The ATN configurations which were valid immediately before
* the {@link #ERROR} state was reached
* @param outerContext The is the \gamma_0 initial parser context from the paper
* or the parser stack at the instant before prediction commences.
*
* @return The value to return from {@link #adaptivePredict}, or
* {@link ATN#INVALID_ALT_NUMBER} if a suitable alternative was not
* identified and {@link #adaptivePredict} should report an error instead.
*/
protected int getSynValidOrSemInvalidAltThatFinishedDecisionEntryRule(ATNConfigSet configs,
ParserRuleContext outerContext)
{
Pair<ATNConfigSet,ATNConfigSet> sets =
splitAccordingToSemanticValidity(configs, outerContext);
ATNConfigSet semValidConfigs = sets.a;
ATNConfigSet semInvalidConfigs = sets.b;
int alt = getAltThatFinishedDecisionEntryRule(semValidConfigs);
if ( alt!=ATN.INVALID_ALT_NUMBER ) { // semantically/syntactically viable path exists
return alt;
}
// Is there a syntactically valid path with a failed pred?
if ( semInvalidConfigs.size()>0 ) {
alt = getAltThatFinishedDecisionEntryRule(semInvalidConfigs);
if ( alt!=ATN.INVALID_ALT_NUMBER ) { // syntactically viable path exists
return alt;
}
}
return ATN.INVALID_ALT_NUMBER;
}
protected int getAltThatFinishedDecisionEntryRule(ATNConfigSet configs) {
IntervalSet alts = new IntervalSet();
for (ATNConfig c : configs) {
if ( c.getOuterContextDepth()>0 || (c.state instanceof RuleStopState && c.context.hasEmptyPath()) ) {
alts.add(c.alt);
}
}
if ( alts.size()==0 ) return ATN.INVALID_ALT_NUMBER;
return alts.getMinElement();
}
Walk the list of configurations and split them according to
those that have preds evaluating to true/false. If no pred, assume
true pred and include in succeeded set. Returns Pair of sets.
Create a new set so as not to alter the incoming parameter.
Assumption: the input stream has been restored to the starting point
prediction, which is where predicates need to evaluate.
/** Walk the list of configurations and split them according to
* those that have preds evaluating to true/false. If no pred, assume
* true pred and include in succeeded set. Returns Pair of sets.
*
* Create a new set so as not to alter the incoming parameter.
*
* Assumption: the input stream has been restored to the starting point
* prediction, which is where predicates need to evaluate.
*/
protected Pair<ATNConfigSet,ATNConfigSet> splitAccordingToSemanticValidity(
ATNConfigSet configs,
ParserRuleContext outerContext)
{
ATNConfigSet succeeded = new ATNConfigSet(configs.fullCtx);
ATNConfigSet failed = new ATNConfigSet(configs.fullCtx);
for (ATNConfig c : configs) {
if ( c.semanticContext!=SemanticContext.NONE ) {
boolean predicateEvaluationResult = evalSemanticContext(c.semanticContext, outerContext, c.alt, configs.fullCtx);
if ( predicateEvaluationResult ) {
succeeded.add(c);
}
else {
failed.add(c);
}
}
else {
succeeded.add(c);
}
}
return new Pair<ATNConfigSet, ATNConfigSet>(succeeded,failed);
}
Look through a list of predicate/alt pairs, returning alts for the pairs that win. A NONE
predicate indicates an alt containing an unpredicated config which behaves as "always true." If !complete then we stop at the first predicate that evaluates to true. This includes pairs with null predicates. /** Look through a list of predicate/alt pairs, returning alts for the
* pairs that win. A {@code NONE} predicate indicates an alt containing an
* unpredicated config which behaves as "always true." If !complete
* then we stop at the first predicate that evaluates to true. This
* includes pairs with null predicates.
*/
protected BitSet evalSemanticContext(DFAState.PredPrediction[] predPredictions,
ParserRuleContext outerContext,
boolean complete)
{
BitSet predictions = new BitSet();
for (DFAState.PredPrediction pair : predPredictions) {
if ( pair.pred==SemanticContext.NONE ) {
predictions.set(pair.alt);
if (!complete) {
break;
}
continue;
}
boolean fullCtx = false; // in dfa
boolean predicateEvaluationResult = evalSemanticContext(pair.pred, outerContext, pair.alt, fullCtx);
if ( debug || dfa_debug ) {
System.out.println("eval pred "+pair+"="+predicateEvaluationResult);
}
if ( predicateEvaluationResult ) {
if ( debug || dfa_debug ) System.out.println("PREDICT "+pair.alt);
predictions.set(pair.alt);
if (!complete) {
break;
}
}
}
return predictions;
}
Evaluate a semantic context within a specific parser context.
This method might not be called for every semantic context evaluated
during the prediction process. In particular, we currently do not
evaluate the following but it may change in the future:
- Precedence predicates (represented by
PrecedencePredicate
) are not currently evaluated through this method.
- Operator predicates (represented by
AND
and OR
) are evaluated as a single semantic context, rather than evaluating the operands individually. Implementations which require evaluation results from individual predicates should override this method to explicitly handle evaluation of the operands within operator predicates.
Params: - pred – The semantic context to evaluate
- parserCallStack – The parser context in which to evaluate the
semantic context
- alt – The alternative which is guarded by
pred
- fullCtx –
true
if the evaluation is occurring during LL prediction; otherwise, false
if the evaluation is occurring during SLL prediction
Since: 4.3
/**
* Evaluate a semantic context within a specific parser context.
*
* <p>
* This method might not be called for every semantic context evaluated
* during the prediction process. In particular, we currently do not
* evaluate the following but it may change in the future:</p>
*
* <ul>
* <li>Precedence predicates (represented by
* {@link SemanticContext.PrecedencePredicate}) are not currently evaluated
* through this method.</li>
* <li>Operator predicates (represented by {@link SemanticContext.AND} and
* {@link SemanticContext.OR}) are evaluated as a single semantic
* context, rather than evaluating the operands individually.
* Implementations which require evaluation results from individual
* predicates should override this method to explicitly handle evaluation of
* the operands within operator predicates.</li>
* </ul>
*
* @param pred The semantic context to evaluate
* @param parserCallStack The parser context in which to evaluate the
* semantic context
* @param alt The alternative which is guarded by {@code pred}
* @param fullCtx {@code true} if the evaluation is occurring during LL
* prediction; otherwise, {@code false} if the evaluation is occurring
* during SLL prediction
*
* @since 4.3
*/
protected boolean evalSemanticContext(SemanticContext pred, ParserRuleContext parserCallStack, int alt, boolean fullCtx) {
return pred.eval(parser, parserCallStack);
}
/* TODO: If we are doing predicates, there is no point in pursuing
closure operations if we reach a DFA state that uniquely predicts
alternative. We will not be caching that DFA state and it is a
waste to pursue the closure. Might have to advance when we do
ambig detection thought :(
*/
protected void closure(ATNConfig config,
ATNConfigSet configs,
Set<ATNConfig> closureBusy,
boolean collectPredicates,
boolean fullCtx,
boolean treatEofAsEpsilon)
{
final int initialDepth = 0;
closureCheckingStopState(config, configs, closureBusy, collectPredicates,
fullCtx,
initialDepth, treatEofAsEpsilon);
assert !fullCtx || !configs.dipsIntoOuterContext;
}
protected void closureCheckingStopState(ATNConfig config,
ATNConfigSet configs,
Set<ATNConfig> closureBusy,
boolean collectPredicates,
boolean fullCtx,
int depth,
boolean treatEofAsEpsilon)
{
if ( debug ) System.out.println("closure("+config.toString(parser,true)+")");
if ( config.state instanceof RuleStopState ) {
// We hit rule end. If we have context info, use it
// run thru all possible stack tops in ctx
if ( !config.context.isEmpty() ) {
for (int i = 0; i < config.context.size(); i++) {
if ( config.context.getReturnState(i)==PredictionContext.EMPTY_RETURN_STATE ) {
if (fullCtx) {
configs.add(new ATNConfig(config, config.state, PredictionContext.EMPTY), mergeCache);
continue;
}
else {
// we have no context info, just chase follow links (if greedy)
if ( debug ) System.out.println("FALLING off rule "+
getRuleName(config.state.ruleIndex));
closure_(config, configs, closureBusy, collectPredicates,
fullCtx, depth, treatEofAsEpsilon);
}
continue;
}
ATNState returnState = atn.states.get(config.context.getReturnState(i));
PredictionContext newContext = config.context.getParent(i); // "pop" return state
ATNConfig c = new ATNConfig(returnState, config.alt, newContext,
config.semanticContext);
// While we have context to pop back from, we may have
// gotten that context AFTER having falling off a rule.
// Make sure we track that we are now out of context.
//
// This assignment also propagates the
// isPrecedenceFilterSuppressed() value to the new
// configuration.
c.reachesIntoOuterContext = config.reachesIntoOuterContext;
assert depth > Integer.MIN_VALUE;
closureCheckingStopState(c, configs, closureBusy, collectPredicates,
fullCtx, depth - 1, treatEofAsEpsilon);
}
return;
}
else if (fullCtx) {
// reached end of start rule
configs.add(config, mergeCache);
return;
}
else {
// else if we have no context info, just chase follow links (if greedy)
if ( debug ) System.out.println("FALLING off rule "+
getRuleName(config.state.ruleIndex));
}
}
closure_(config, configs, closureBusy, collectPredicates,
fullCtx, depth, treatEofAsEpsilon);
}
Do the actual work of walking epsilon edges /** Do the actual work of walking epsilon edges */
protected void closure_(ATNConfig config,
ATNConfigSet configs,
Set<ATNConfig> closureBusy,
boolean collectPredicates,
boolean fullCtx,
int depth,
boolean treatEofAsEpsilon)
{
ATNState p = config.state;
// optimization
if ( !p.onlyHasEpsilonTransitions() ) {
configs.add(config, mergeCache);
// make sure to not return here, because EOF transitions can act as
// both epsilon transitions and non-epsilon transitions.
// if ( debug ) System.out.println("added config "+configs);
}
for (int i=0; i<p.getNumberOfTransitions(); i++) {
if ( i==0 && canDropLoopEntryEdgeInLeftRecursiveRule(config) ) continue;
Transition t = p.transition(i);
boolean continueCollecting =
!(t instanceof ActionTransition) && collectPredicates;
ATNConfig c = getEpsilonTarget(config, t, continueCollecting,
depth == 0, fullCtx, treatEofAsEpsilon);
if ( c!=null ) {
int newDepth = depth;
if ( config.state instanceof RuleStopState) {
assert !fullCtx;
// target fell off end of rule; mark resulting c as having dipped into outer context
// We can't get here if incoming config was rule stop and we had context
// track how far we dip into outer context. Might
// come in handy and we avoid evaluating context dependent
// preds if this is > 0.
if (_dfa != null && _dfa.isPrecedenceDfa()) {
int outermostPrecedenceReturn = ((EpsilonTransition)t).outermostPrecedenceReturn();
if (outermostPrecedenceReturn == _dfa.atnStartState.ruleIndex) {
c.setPrecedenceFilterSuppressed(true);
}
}
c.reachesIntoOuterContext++;
if (!closureBusy.add(c)) {
// avoid infinite recursion for right-recursive rules
continue;
}
configs.dipsIntoOuterContext = true; // TODO: can remove? only care when we add to set per middle of this method
assert newDepth > Integer.MIN_VALUE;
newDepth--;
if ( debug ) System.out.println("dips into outer ctx: "+c);
}
else {
if (!t.isEpsilon() && !closureBusy.add(c)) {
// avoid infinite recursion for EOF* and EOF+
continue;
}
if (t instanceof RuleTransition) {
// latch when newDepth goes negative - once we step out of the entry context we can't return
if (newDepth >= 0) {
newDepth++;
}
}
}
closureCheckingStopState(c, configs, closureBusy, continueCollecting,
fullCtx, newDepth, treatEofAsEpsilon);
}
}
}
Implements first-edge (loop entry) elimination as an optimization
during closure operations. See antlr/antlr4#1398.
The optimization is to avoid adding the loop entry config when
the exit path can only lead back to the same
StarLoopEntryState after popping context at the rule end state
(traversing only epsilon edges, so we're still in closure, in
this same rule).
We need to detect any state that can reach loop entry on
epsilon w/o exiting rule. We don't have to look at FOLLOW
links, just ensure that all stack tops for config refer to key
states in LR rule.
To verify we are in the right situation we must first check
closure is at a StarLoopEntryState generated during LR removal.
Then we check that each stack top of context is a return state
from one of these cases:
1. 'not' expr, '(' type ')' expr. The return state points at loop entry state
2. expr op expr. The return state is the block end of internal block of (...)*
3. 'between' expr 'and' expr. The return state of 2nd expr reference.
That state points at block end of internal block of (...)*.
4. expr '?' expr ':' expr. The return state points at block end,
which points at loop entry state.
If any is true for each stack top, then closure does not add a
config to the current config set for edge[0], the loop entry branch.
Conditions fail if any context for the current config is:
a. empty (we'd fall out of expr to do a global FOLLOW which could
even be to some weird spot in expr) or,
b. lies outside of expr or,
c. lies within expr but at a state not the BlockEndState
generated during LR removal
Do we need to evaluate predicates ever in closure for this case?
No. Predicates, including precedence predicates, are only
evaluated when computing a DFA start state. I.e., only before
the lookahead (but not parser) consumes a token.
There are no epsilon edges allowed in LR rule alt blocks or in
the "primary" part (ID here). If closure is in
StarLoopEntryState any lookahead operation will have consumed a
token as there are no epsilon-paths that lead to
StarLoopEntryState. We do not have to evaluate predicates
therefore if we are in the generated StarLoopEntryState of a LR
rule. Note that when making a prediction starting at that
decision point, decision d=2, compute-start-state performs
closure starting at edges[0], edges[1] emanating from
StarLoopEntryState. That means it is not performing closure on
StarLoopEntryState during compute-start-state.
How do we know this always gives same prediction answer?
Without predicates, loop entry and exit paths are ambiguous
upon remaining input +b (in, say, a+b). Either paths lead to
valid parses. Closure can lead to consuming + immediately or by
falling out of this call to expr back into expr and loop back
again to StarLoopEntryState to match +b. In this special case,
we choose the more efficient path, which is to take the bypass
path.
The lookahead language has not changed because closure chooses
one path over the other. Both paths lead to consuming the same
remaining input during a lookahead operation. If the next token
is an operator, lookahead will enter the choice block with
operators. If it is not, lookahead will exit expr. Same as if
closure had chosen to enter the choice block immediately.
Closure is examining one config (some loopentrystate, some alt,
context) which means it is considering exactly one alt. Closure
always copies the same alt to any derived configs.
How do we know this optimization doesn't mess up precedence in
our parse trees?
Looking through expr from left edge of stat only has to confirm
that an input, say, a+b+c; begins with any valid interpretation
of an expression. The precedence actually doesn't matter when
making a decision in stat seeing through expr. It is only when
parsing rule expr that we must use the precedence to get the
right interpretation and, hence, parse tree.
Since: 4.6
/** Implements first-edge (loop entry) elimination as an optimization
* during closure operations. See antlr/antlr4#1398.
*
* The optimization is to avoid adding the loop entry config when
* the exit path can only lead back to the same
* StarLoopEntryState after popping context at the rule end state
* (traversing only epsilon edges, so we're still in closure, in
* this same rule).
*
* We need to detect any state that can reach loop entry on
* epsilon w/o exiting rule. We don't have to look at FOLLOW
* links, just ensure that all stack tops for config refer to key
* states in LR rule.
*
* To verify we are in the right situation we must first check
* closure is at a StarLoopEntryState generated during LR removal.
* Then we check that each stack top of context is a return state
* from one of these cases:
*
* 1. 'not' expr, '(' type ')' expr. The return state points at loop entry state
* 2. expr op expr. The return state is the block end of internal block of (...)*
* 3. 'between' expr 'and' expr. The return state of 2nd expr reference.
* That state points at block end of internal block of (...)*.
* 4. expr '?' expr ':' expr. The return state points at block end,
* which points at loop entry state.
*
* If any is true for each stack top, then closure does not add a
* config to the current config set for edge[0], the loop entry branch.
*
* Conditions fail if any context for the current config is:
*
* a. empty (we'd fall out of expr to do a global FOLLOW which could
* even be to some weird spot in expr) or,
* b. lies outside of expr or,
* c. lies within expr but at a state not the BlockEndState
* generated during LR removal
*
* Do we need to evaluate predicates ever in closure for this case?
*
* No. Predicates, including precedence predicates, are only
* evaluated when computing a DFA start state. I.e., only before
* the lookahead (but not parser) consumes a token.
*
* There are no epsilon edges allowed in LR rule alt blocks or in
* the "primary" part (ID here). If closure is in
* StarLoopEntryState any lookahead operation will have consumed a
* token as there are no epsilon-paths that lead to
* StarLoopEntryState. We do not have to evaluate predicates
* therefore if we are in the generated StarLoopEntryState of a LR
* rule. Note that when making a prediction starting at that
* decision point, decision d=2, compute-start-state performs
* closure starting at edges[0], edges[1] emanating from
* StarLoopEntryState. That means it is not performing closure on
* StarLoopEntryState during compute-start-state.
*
* How do we know this always gives same prediction answer?
*
* Without predicates, loop entry and exit paths are ambiguous
* upon remaining input +b (in, say, a+b). Either paths lead to
* valid parses. Closure can lead to consuming + immediately or by
* falling out of this call to expr back into expr and loop back
* again to StarLoopEntryState to match +b. In this special case,
* we choose the more efficient path, which is to take the bypass
* path.
*
* The lookahead language has not changed because closure chooses
* one path over the other. Both paths lead to consuming the same
* remaining input during a lookahead operation. If the next token
* is an operator, lookahead will enter the choice block with
* operators. If it is not, lookahead will exit expr. Same as if
* closure had chosen to enter the choice block immediately.
*
* Closure is examining one config (some loopentrystate, some alt,
* context) which means it is considering exactly one alt. Closure
* always copies the same alt to any derived configs.
*
* How do we know this optimization doesn't mess up precedence in
* our parse trees?
*
* Looking through expr from left edge of stat only has to confirm
* that an input, say, a+b+c; begins with any valid interpretation
* of an expression. The precedence actually doesn't matter when
* making a decision in stat seeing through expr. It is only when
* parsing rule expr that we must use the precedence to get the
* right interpretation and, hence, parse tree.
*
* @since 4.6
*/
protected boolean canDropLoopEntryEdgeInLeftRecursiveRule(ATNConfig config) {
if ( TURN_OFF_LR_LOOP_ENTRY_BRANCH_OPT ) return false;
ATNState p = config.state;
// First check to see if we are in StarLoopEntryState generated during
// left-recursion elimination. For efficiency, also check if
// the context has an empty stack case. If so, it would mean
// global FOLLOW so we can't perform optimization
if ( p.getStateType() != ATNState.STAR_LOOP_ENTRY ||
!((StarLoopEntryState)p).isPrecedenceDecision || // Are we the special loop entry/exit state?
config.context.isEmpty() || // If SLL wildcard
config.context.hasEmptyPath())
{
return false;
}
// Require all return states to return back to the same rule
// that p is in.
int numCtxs = config.context.size();
for (int i = 0; i < numCtxs; i++) { // for each stack context
ATNState returnState = atn.states.get(config.context.getReturnState(i));
if ( returnState.ruleIndex != p.ruleIndex ) return false;
}
BlockStartState decisionStartState = (BlockStartState)p.transition(0).target;
int blockEndStateNum = decisionStartState.endState.stateNumber;
BlockEndState blockEndState = (BlockEndState)atn.states.get(blockEndStateNum);
// Verify that the top of each stack context leads to loop entry/exit
// state through epsilon edges and w/o leaving rule.
for (int i = 0; i < numCtxs; i++) { // for each stack context
int returnStateNumber = config.context.getReturnState(i);
ATNState returnState = atn.states.get(returnStateNumber);
// all states must have single outgoing epsilon edge
if ( returnState.getNumberOfTransitions()!=1 ||
!returnState.transition(0).isEpsilon() )
{
return false;
}
// Look for prefix op case like 'not expr', (' type ')' expr
ATNState returnStateTarget = returnState.transition(0).target;
if ( returnState.getStateType()==BLOCK_END && returnStateTarget==p ) {
continue;
}
// Look for 'expr op expr' or case where expr's return state is block end
// of (...)* internal block; the block end points to loop back
// which points to p but we don't need to check that
if ( returnState==blockEndState ) {
continue;
}
// Look for ternary expr ? expr : expr. The return state points at block end,
// which points at loop entry state
if ( returnStateTarget==blockEndState ) {
continue;
}
// Look for complex prefix 'between expr and expr' case where 2nd expr's
// return state points at block end state of (...)* internal block
if ( returnStateTarget.getStateType() == BLOCK_END &&
returnStateTarget.getNumberOfTransitions()==1 &&
returnStateTarget.transition(0).isEpsilon() &&
returnStateTarget.transition(0).target == p )
{
continue;
}
// anything else ain't conforming
return false;
}
return true;
}
public String getRuleName(int index) {
if ( parser!=null && index>=0 ) return parser.getRuleNames()[index];
return "<rule "+index+">";
}
protected ATNConfig getEpsilonTarget(ATNConfig config,
Transition t,
boolean collectPredicates,
boolean inContext,
boolean fullCtx,
boolean treatEofAsEpsilon)
{
switch (t.getSerializationType()) {
case Transition.RULE:
return ruleTransition(config, (RuleTransition)t);
case Transition.PRECEDENCE:
return precedenceTransition(config, (PrecedencePredicateTransition)t, collectPredicates, inContext, fullCtx);
case Transition.PREDICATE:
return predTransition(config, (PredicateTransition)t,
collectPredicates,
inContext,
fullCtx);
case Transition.ACTION:
return actionTransition(config, (ActionTransition)t);
case Transition.EPSILON:
return new ATNConfig(config, t.target);
case Transition.ATOM:
case Transition.RANGE:
case Transition.SET:
// EOF transitions act like epsilon transitions after the first EOF
// transition is traversed
if (treatEofAsEpsilon) {
if (t.matches(Token.EOF, 0, 1)) {
return new ATNConfig(config, t.target);
}
}
return null;
default:
return null;
}
}
protected ATNConfig actionTransition(ATNConfig config, ActionTransition t) {
if ( debug ) System.out.println("ACTION edge "+t.ruleIndex+":"+t.actionIndex);
return new ATNConfig(config, t.target);
}
public ATNConfig precedenceTransition(ATNConfig config,
PrecedencePredicateTransition pt,
boolean collectPredicates,
boolean inContext,
boolean fullCtx)
{
if ( debug ) {
System.out.println("PRED (collectPredicates="+collectPredicates+") "+
pt.precedence+">=_p"+
", ctx dependent=true");
if ( parser != null ) {
System.out.println("context surrounding pred is "+
parser.getRuleInvocationStack());
}
}
ATNConfig c = null;
if (collectPredicates && inContext) {
if ( fullCtx ) {
// In full context mode, we can evaluate predicates on-the-fly
// during closure, which dramatically reduces the size of
// the config sets. It also obviates the need to test predicates
// later during conflict resolution.
int currentPosition = _input.index();
_input.seek(_startIndex);
boolean predSucceeds = evalSemanticContext(pt.getPredicate(), _outerContext, config.alt, fullCtx);
_input.seek(currentPosition);
if ( predSucceeds ) {
c = new ATNConfig(config, pt.target); // no pred context
}
}
else {
SemanticContext newSemCtx =
SemanticContext.and(config.semanticContext, pt.getPredicate());
c = new ATNConfig(config, pt.target, newSemCtx);
}
}
else {
c = new ATNConfig(config, pt.target);
}
if ( debug ) System.out.println("config from pred transition="+c);
return c;
}
protected ATNConfig predTransition(ATNConfig config,
PredicateTransition pt,
boolean collectPredicates,
boolean inContext,
boolean fullCtx)
{
if ( debug ) {
System.out.println("PRED (collectPredicates="+collectPredicates+") "+
pt.ruleIndex+":"+pt.predIndex+
", ctx dependent="+pt.isCtxDependent);
if ( parser != null ) {
System.out.println("context surrounding pred is "+
parser.getRuleInvocationStack());
}
}
ATNConfig c = null;
if ( collectPredicates &&
(!pt.isCtxDependent || (pt.isCtxDependent&&inContext)) )
{
if ( fullCtx ) {
// In full context mode, we can evaluate predicates on-the-fly
// during closure, which dramatically reduces the size of
// the config sets. It also obviates the need to test predicates
// later during conflict resolution.
int currentPosition = _input.index();
_input.seek(_startIndex);
boolean predSucceeds = evalSemanticContext(pt.getPredicate(), _outerContext, config.alt, fullCtx);
_input.seek(currentPosition);
if ( predSucceeds ) {
c = new ATNConfig(config, pt.target); // no pred context
}
}
else {
SemanticContext newSemCtx =
SemanticContext.and(config.semanticContext, pt.getPredicate());
c = new ATNConfig(config, pt.target, newSemCtx);
}
}
else {
c = new ATNConfig(config, pt.target);
}
if ( debug ) System.out.println("config from pred transition="+c);
return c;
}
protected ATNConfig ruleTransition(ATNConfig config, RuleTransition t) {
if ( debug ) {
System.out.println("CALL rule "+getRuleName(t.target.ruleIndex)+
", ctx="+config.context);
}
ATNState returnState = t.followState;
PredictionContext newContext =
SingletonPredictionContext.create(config.context, returnState.stateNumber);
return new ATNConfig(config, t.target, newContext);
}
Gets a BitSet
containing the alternatives in configs
which are part of one or more conflicting alternative subsets. Params: - configs – The
ATNConfigSet
to analyze.
Returns: The alternatives in configs
which are part of one or more conflicting alternative subsets. If configs
does not contain any conflicting subsets, this method returns an empty BitSet
.
/**
* Gets a {@link BitSet} containing the alternatives in {@code configs}
* which are part of one or more conflicting alternative subsets.
*
* @param configs The {@link ATNConfigSet} to analyze.
* @return The alternatives in {@code configs} which are part of one or more
* conflicting alternative subsets. If {@code configs} does not contain any
* conflicting subsets, this method returns an empty {@link BitSet}.
*/
protected BitSet getConflictingAlts(ATNConfigSet configs) {
Collection<BitSet> altsets = PredictionMode.getConflictingAltSubsets(configs);
return PredictionMode.getAlts(altsets);
}
Sam pointed out a problem with the previous definition, v3, of
ambiguous states. If we have another state associated with conflicting
alternatives, we should keep going. For example, the following grammar
s : (ID | ID ID?) ';' ;
When the ATN simulation reaches the state before ';', it has a DFA
state that looks like: [12|1|[], 6|2|[], 12|2|[]]. Naturally
12|1|[] and 12|2|[] conflict, but we cannot stop processing this node
because alternative to has another way to continue, via [6|2|[]].
The key is that we have a single state that has config's only associated
with a single alternative, 2, and crucially the state transitions
among the configurations are all non-epsilon transitions. That means
we don't consider any conflicts that include alternative 2. So, we
ignore the conflict between alts 1 and 2. We ignore a set of
conflicting alts when there is an intersection with an alternative
associated with a single alt state in the state→config-list map.
It's also the case that we might have two conflicting configurations but
also a 3rd nonconflicting configuration for a different alternative:
[1|1|[], 1|2|[], 8|3|[]]. This can come about from grammar:
a : A | A | A B ;
After matching input A, we reach the stop state for rule A, state 1.
State 8 is the state right before B. Clearly alternatives 1 and 2
conflict and no amount of further lookahead will separate the two.
However, alternative 3 will be able to continue and so we do not
stop working on this state. In the previous example, we're concerned
with states associated with the conflicting alternatives. Here alt
3 is not associated with the conflicting configs, but since we can continue
looking for input reasonably, I don't declare the state done. We
ignore a set of conflicting alts when we have an alternative
that we still need to pursue.
/**
Sam pointed out a problem with the previous definition, v3, of
ambiguous states. If we have another state associated with conflicting
alternatives, we should keep going. For example, the following grammar
s : (ID | ID ID?) ';' ;
When the ATN simulation reaches the state before ';', it has a DFA
state that looks like: [12|1|[], 6|2|[], 12|2|[]]. Naturally
12|1|[] and 12|2|[] conflict, but we cannot stop processing this node
because alternative to has another way to continue, via [6|2|[]].
The key is that we have a single state that has config's only associated
with a single alternative, 2, and crucially the state transitions
among the configurations are all non-epsilon transitions. That means
we don't consider any conflicts that include alternative 2. So, we
ignore the conflict between alts 1 and 2. We ignore a set of
conflicting alts when there is an intersection with an alternative
associated with a single alt state in the state→config-list map.
It's also the case that we might have two conflicting configurations but
also a 3rd nonconflicting configuration for a different alternative:
[1|1|[], 1|2|[], 8|3|[]]. This can come about from grammar:
a : A | A | A B ;
After matching input A, we reach the stop state for rule A, state 1.
State 8 is the state right before B. Clearly alternatives 1 and 2
conflict and no amount of further lookahead will separate the two.
However, alternative 3 will be able to continue and so we do not
stop working on this state. In the previous example, we're concerned
with states associated with the conflicting alternatives. Here alt
3 is not associated with the conflicting configs, but since we can continue
looking for input reasonably, I don't declare the state done. We
ignore a set of conflicting alts when we have an alternative
that we still need to pursue.
*/
protected BitSet getConflictingAltsOrUniqueAlt(ATNConfigSet configs) {
BitSet conflictingAlts;
if ( configs.uniqueAlt!= ATN.INVALID_ALT_NUMBER ) {
conflictingAlts = new BitSet();
conflictingAlts.set(configs.uniqueAlt);
}
else {
conflictingAlts = configs.conflictingAlts;
}
return conflictingAlts;
}
public String getTokenName(int t) {
if (t == Token.EOF) {
return "EOF";
}
Vocabulary vocabulary = parser != null ? parser.getVocabulary() : VocabularyImpl.EMPTY_VOCABULARY;
String displayName = vocabulary.getDisplayName(t);
if (displayName.equals(Integer.toString(t))) {
return displayName;
}
return displayName + "<" + t + ">";
}
public String getLookaheadName(TokenStream input) {
return getTokenName(input.LA(1));
}
Used for debugging in adaptivePredict around execATN but I cut
it out for clarity now that alg. works well. We can leave this
"dead" code for a bit.
/** Used for debugging in adaptivePredict around execATN but I cut
* it out for clarity now that alg. works well. We can leave this
* "dead" code for a bit.
*/
public void dumpDeadEndConfigs(NoViableAltException nvae) {
System.err.println("dead end configs: ");
for (ATNConfig c : nvae.getDeadEndConfigs()) {
String trans = "no edges";
if ( c.state.getNumberOfTransitions()>0 ) {
Transition t = c.state.transition(0);
if ( t instanceof AtomTransition) {
AtomTransition at = (AtomTransition)t;
trans = "Atom "+getTokenName(at.label);
}
else if ( t instanceof SetTransition ) {
SetTransition st = (SetTransition)t;
boolean not = st instanceof NotSetTransition;
trans = (not?"~":"")+"Set "+st.set.toString();
}
}
System.err.println(c.toString(parser, true)+":"+trans);
}
}
protected NoViableAltException noViableAlt(TokenStream input,
ParserRuleContext outerContext,
ATNConfigSet configs,
int startIndex)
{
return new NoViableAltException(parser, input,
input.get(startIndex),
input.LT(1),
configs, outerContext);
}
protected static int getUniqueAlt(ATNConfigSet configs) {
int alt = ATN.INVALID_ALT_NUMBER;
for (ATNConfig c : configs) {
if ( alt == ATN.INVALID_ALT_NUMBER ) {
alt = c.alt; // found first alt
}
else if ( c.alt!=alt ) {
return ATN.INVALID_ALT_NUMBER;
}
}
return alt;
}
Add an edge to the DFA, if possible. This method calls addDFAState
to ensure the to
state is present in the DFA. If from
is null
, or if t
is outside the range of edges that can be represented in the DFA tables, this method returns without adding the edge to the DFA. If to
is null
, this method returns null
. Otherwise, this method returns the DFAState
returned by calling addDFAState
for the to
state.
Params: - dfa – The DFA
- from – The source state for the edge
- t – The input symbol
- to – The target state for the edge
Returns: If to
is null
, this method returns null
; otherwise this method returns the result of calling addDFAState
on to
/**
* Add an edge to the DFA, if possible. This method calls
* {@link #addDFAState} to ensure the {@code to} state is present in the
* DFA. If {@code from} is {@code null}, or if {@code t} is outside the
* range of edges that can be represented in the DFA tables, this method
* returns without adding the edge to the DFA.
*
* <p>If {@code to} is {@code null}, this method returns {@code null}.
* Otherwise, this method returns the {@link DFAState} returned by calling
* {@link #addDFAState} for the {@code to} state.</p>
*
* @param dfa The DFA
* @param from The source state for the edge
* @param t The input symbol
* @param to The target state for the edge
*
* @return If {@code to} is {@code null}, this method returns {@code null};
* otherwise this method returns the result of calling {@link #addDFAState}
* on {@code to}
*/
protected DFAState addDFAEdge(DFA dfa,
DFAState from,
int t,
DFAState to)
{
if ( debug ) {
System.out.println("EDGE "+from+" -> "+to+" upon "+getTokenName(t));
}
if (to == null) {
return null;
}
to = addDFAState(dfa, to); // used existing if possible not incoming
if (from == null || t < -1 || t > atn.maxTokenType) {
return to;
}
synchronized (from) {
if ( from.edges==null ) {
from.edges = new DFAState[atn.maxTokenType+1+1];
}
from.edges[t+1] = to; // connect
}
if ( debug ) {
System.out.println("DFA=\n"+dfa.toString(parser!=null?parser.getVocabulary():VocabularyImpl.EMPTY_VOCABULARY));
}
return to;
}
Add state D
to the DFA if it is not already present, and return the actual instance stored in the DFA. If a state equivalent to D
is already in the DFA, the existing state is returned. Otherwise this method returns D
after adding it to the DFA. If D
is ATNSimulator.ERROR
, this method returns ATNSimulator.ERROR
and does not change the DFA.
Params: - dfa – The dfa
- D – The DFA state to add
Returns: The state stored in the DFA. This will be either the existing state if D
is already in the DFA, or D
itself if the state was not already present.
/**
* Add state {@code D} to the DFA if it is not already present, and return
* the actual instance stored in the DFA. If a state equivalent to {@code D}
* is already in the DFA, the existing state is returned. Otherwise this
* method returns {@code D} after adding it to the DFA.
*
* <p>If {@code D} is {@link #ERROR}, this method returns {@link #ERROR} and
* does not change the DFA.</p>
*
* @param dfa The dfa
* @param D The DFA state to add
* @return The state stored in the DFA. This will be either the existing
* state if {@code D} is already in the DFA, or {@code D} itself if the
* state was not already present.
*/
protected DFAState addDFAState(DFA dfa, DFAState D) {
if (D == ERROR) {
return D;
}
synchronized (dfa.states) {
DFAState existing = dfa.states.get(D);
if ( existing!=null ) return existing;
D.stateNumber = dfa.states.size();
if (!D.configs.isReadonly()) {
D.configs.optimizeConfigs(this);
D.configs.setReadonly(true);
}
dfa.states.put(D, D);
if ( debug ) System.out.println("adding new DFA state: "+D);
return D;
}
}
protected void reportAttemptingFullContext(DFA dfa, BitSet conflictingAlts, ATNConfigSet configs, int startIndex, int stopIndex) {
if ( debug || retry_debug ) {
Interval interval = Interval.of(startIndex, stopIndex);
System.out.println("reportAttemptingFullContext decision="+dfa.decision+":"+configs+
", input="+parser.getTokenStream().getText(interval));
}
if ( parser!=null ) parser.getErrorListenerDispatch().reportAttemptingFullContext(parser, dfa, startIndex, stopIndex, conflictingAlts, configs);
}
protected void reportContextSensitivity(DFA dfa, int prediction, ATNConfigSet configs, int startIndex, int stopIndex) {
if ( debug || retry_debug ) {
Interval interval = Interval.of(startIndex, stopIndex);
System.out.println("reportContextSensitivity decision="+dfa.decision+":"+configs+
", input="+parser.getTokenStream().getText(interval));
}
if ( parser!=null ) parser.getErrorListenerDispatch().reportContextSensitivity(parser, dfa, startIndex, stopIndex, prediction, configs);
}
If context sensitive parsing, we know it's ambiguity not conflict /** If context sensitive parsing, we know it's ambiguity not conflict */
protected void reportAmbiguity(DFA dfa,
DFAState D, // the DFA state from execATN() that had SLL conflicts
int startIndex, int stopIndex,
boolean exact,
BitSet ambigAlts,
ATNConfigSet configs) // configs that LL not SLL considered conflicting
{
if ( debug || retry_debug ) {
Interval interval = Interval.of(startIndex, stopIndex);
System.out.println("reportAmbiguity "+
ambigAlts+":"+configs+
", input="+parser.getTokenStream().getText(interval));
}
if ( parser!=null ) parser.getErrorListenerDispatch().reportAmbiguity(parser, dfa, startIndex, stopIndex,
exact, ambigAlts, configs);
}
public final void setPredictionMode(PredictionMode mode) {
this.mode = mode;
}
public final PredictionMode getPredictionMode() {
return mode;
}
Since: 4.3
/**
* @since 4.3
*/
public Parser getParser() {
return parser;
}
public static String getSafeEnv(String envName) {
try {
return System.getenv(envName);
}
catch(SecurityException e) {
// use the default value
}
return null;
}
}