Lines Matching refs:closure

54  * caching problem. The closure routine only considers the rule invocation stack
80  * context, because closure can fall off the end of a rule. I tried to cache
127  * In both cases I believe that no closure operations will dip into the outer
136  * it performs closure operations like ANTLR v3 did. It delays predicate
139  * closure, the DFA state configuration sets would be different and we couldn't
154  * During full LL ATN simulation, closure always evaluates predicates and
156  * closure. It hits a landmine when parsing with the Java grammar, for example,
890     protected function computeReachSet(ATNConfigSet $closure, int $t, bool $fullCtx) : ?ATNConfigSet
893             $this->log[] = \sprintf('In computeReachSet, starting closure: %s', (string) $closure);
906          * closure operation, so they are handled separately for the performance
907          * advantage of having a smaller intermediate set when calling closure.
916         foreach ($closure->elements() as $c) {
970                 // Don't pursue the closure if there is just one state.
972                 // Also don't pursue the closure if there is unique alternative
977                 // Also don't pursue the closure if there is unique alternative among the configurations.
982         // If the reach set could not be trivially determined, perform a closure
990                 $this->closure($item, $reach, $closureBusy, false, $fullCtx, $treatEofAsEpsilon);
1002              * When reach==intermediate, no closure operation was performed. In
1020          * closure operation reached such a state. This ensures adaptivePredict
1103             $this->closure($c, $configs, $closureBusy, true, $fullCtx, false);
1191      * flag is carried along in closure(). so to avoid adding field, set bit
1211      * - The closure block cannot contain any epsilon transitions which bypass
1212      *    the body of the closure, so all states reachable via alternative 1 are
1231      * of both the primary and closure blocks of the left-recursive rule
1585      *       closure operations if we reach a DFA state that uniquely predicts
1587      *       waste to pursue the closure. Might have to advance when we do
1590     protected function closure(
1625             $this->log[] = \sprintf('closure(%s)', $config->toString(true));
1806      * during closure operations. See antlr/antlr4#1398.
1811      * (traversing only epsilon edges, so we're still in closure, in
1820      * closure is at a StarLoopEntryState generated during LR removal.
1831      * If any is true for each stack top, then closure does not add a
1842      * Do we need to evaluate predicates ever in closure for this case?
1849      * the "primary" part (ID here). If closure is in
1856      * closure starting at edges[0], edges[1] emanating from
1857      * StarLoopEntryState. That means it is not performing closure on
1870      * The lookahead language has not changed because closure chooses
1875      * closure had chosen to enter the choice block immediately.
2092                  * during closure, which dramatically reduces the size of
2152                 // during closure, which dramatically reduces the size of