/*
 * Copyright (c) 2010, 2015, Oracle and/or its affiliates. All rights reserved.
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
 *
 * This code is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License version 2 only, as
 * published by the Free Software Foundation.  Oracle designates this
 * particular file as subject to the "Classpath" exception as provided
 * by Oracle in the LICENSE file that accompanied this code.
 *
 * This code is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
 * version 2 for more details (a copy is included in the LICENSE file that
 * accompanied this code).
 *
 * You should have received a copy of the GNU General Public License version
 * 2 along with this work; if not, write to the Free Software Foundation,
 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
 *
 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
 * or visit www.oracle.com if you need additional information or have any
 * questions.
 */
package jdk.nashorn.internal.ir;

import java.io.File;
import java.util.Iterator;
import java.util.NoSuchElementException;
import jdk.nashorn.internal.runtime.Debug;
import jdk.nashorn.internal.runtime.Source;

A class that tracks the current lexical context of node visitation as a stack of Block nodes. Has special methods to retrieve useful subsets of the context. This is implemented with a primitive array and a stack pointer, because it really makes a difference performance-wise. None of the collection classes were optimal.
/** * A class that tracks the current lexical context of node visitation as a stack * of {@link Block} nodes. Has special methods to retrieve useful subsets of the * context. * * This is implemented with a primitive array and a stack pointer, because it * really makes a difference performance-wise. None of the collection classes * were optimal. */
public class LexicalContext { private LexicalContextNode[] stack; private int[] flags; private int sp;
Creates a new empty lexical context.
/** * Creates a new empty lexical context. */
public LexicalContext() { stack = new LexicalContextNode[16]; flags = new int[16]; }
Set the flags for a lexical context node on the stack. Does not replace the flags, but rather adds to them.
Params:
  • node – node
  • flag – new flag to set
/** * Set the flags for a lexical context node on the stack. Does not * replace the flags, but rather adds to them. * * @param node node * @param flag new flag to set */
public void setFlag(final LexicalContextNode node, final int flag) { if (flag != 0) { // Use setBlockNeedsScope() instead assert !(flag == Block.NEEDS_SCOPE && node instanceof Block); for (int i = sp - 1; i >= 0; i--) { if (stack[i] == node) { flags[i] |= flag; return; } } } assert false; }
Marks the block as one that creates a scope. Note that this method must be used instead of setFlag(LexicalContextNode, int) with Block.NEEDS_SCOPE because it atomically also sets the FunctionNode.HAS_SCOPE_BLOCK flag on the block's containing function.
Params:
  • block – the block that needs to be marked as creating a scope.
/** * Marks the block as one that creates a scope. Note that this method must * be used instead of {@link #setFlag(LexicalContextNode, int)} with * {@link Block#NEEDS_SCOPE} because it atomically also sets the * {@link FunctionNode#HAS_SCOPE_BLOCK} flag on the block's containing * function. * * @param block the block that needs to be marked as creating a scope. */
public void setBlockNeedsScope(final Block block) { for (int i = sp - 1; i >= 0; i--) { if (stack[i] == block) { flags[i] |= Block.NEEDS_SCOPE; for(int j = i - 1; j >=0; j --) { if(stack[j] instanceof FunctionNode) { flags[j] |= FunctionNode.HAS_SCOPE_BLOCK; return; } } } } assert false; }
Get the flags for a lexical context node on the stack.
Params:
  • node – node
Returns:the flags for the node
/** * Get the flags for a lexical context node on the stack. * * @param node node * * @return the flags for the node */
public int getFlags(final LexicalContextNode node) { for (int i = sp - 1; i >= 0; i--) { if (stack[i] == node) { return flags[i]; } } throw new AssertionError("flag node not on context stack"); }
Get the function body of a function node on the lexical context stack. This will trigger an assertion if node isn't present.
Params:
  • functionNode – function node
Returns:body of function node
/** * Get the function body of a function node on the lexical context * stack. This will trigger an assertion if node isn't present. * * @param functionNode function node * * @return body of function node */
public Block getFunctionBody(final FunctionNode functionNode) { for (int i = sp - 1; i >= 0 ; i--) { if (stack[i] == functionNode) { return (Block)stack[i + 1]; } } throw new AssertionError(functionNode.getName() + " not on context stack"); }
Returns:all nodes in the LexicalContext.
/** * @return all nodes in the LexicalContext. */
public Iterator<LexicalContextNode> getAllNodes() { return new NodeIterator<>(LexicalContextNode.class); }
Returns the outermost function in this context. It is either the program, or a lazily compiled function.
Returns:the outermost function in this context.
/** * Returns the outermost function in this context. It is either the program, * or a lazily compiled function. * * @return the outermost function in this context. */
public FunctionNode getOutermostFunction() { return (FunctionNode)stack[0]; }
Pushes a new block on top of the context, making it the innermost open block.
Params:
  • node – the new node
Type parameters:
  • <T> – the type of the new node
Returns:the node that was pushed
/** * Pushes a new block on top of the context, making it the innermost open * block. * * @param <T> the type of the new node * @param node the new node * * @return the node that was pushed */
public <T extends LexicalContextNode> T push(final T node) { assert !contains(node); if (sp == stack.length) { final LexicalContextNode[] newStack = new LexicalContextNode[sp * 2]; System.arraycopy(stack, 0, newStack, 0, sp); stack = newStack; final int[] newFlags = new int[sp * 2]; System.arraycopy(flags, 0, newFlags, 0, sp); flags = newFlags; } stack[sp] = node; flags[sp] = 0; sp++; return node; }
Is the context empty?
Returns:true if empty
/** * Is the context empty? * * @return {@code true} if empty */
public boolean isEmpty() { return sp == 0; }
Returns:the depth of the lexical context.
/** * @return the depth of the lexical context. */
public int size() { return sp; }
Pops the innermost block off the context and all nodes that has been contributed since it was put there.
Params:
  • node – the node expected to be popped, used to detect unbalanced pushes/pops
Type parameters:
  • <T> – the type of the node to be popped
Returns:the node that was popped
/** * Pops the innermost block off the context and all nodes that has been * contributed since it was put there. * * @param <T> the type of the node to be popped * @param node the node expected to be popped, used to detect unbalanced * pushes/pops * * @return the node that was popped */
@SuppressWarnings("unchecked") public <T extends Node> T pop(final T node) { --sp; final LexicalContextNode popped = stack[sp]; stack[sp] = null; if (popped instanceof Flags) { return (T)((Flags<?>)popped).setFlag(this, flags[sp]); } return (T)popped; }
Explicitly apply flags to the topmost element on the stack. This is only valid to use from a NodeVisitor.leaveXxx() method and only on the node being exited at the time. It is not mandatory to use, as pop(Node) will apply the flags automatically, but this method can be used to apply them during the leaveXxx() method in case its logic depends on the value of the flags.
Params:
  • node – the node to apply the flags to. Must be the topmost node on the stack.
Type parameters:
  • <T> – the type of the node to apply the flags to.
Returns:the passed in node, or a modified node (if any flags were modified)
/** * Explicitly apply flags to the topmost element on the stack. This is only * valid to use from a {@code NodeVisitor.leaveXxx()} method and only on the * node being exited at the time. It is not mandatory to use, as * {@link #pop(Node)} will apply the flags automatically, but this method * can be used to apply them during the {@code leaveXxx()} method in case * its logic depends on the value of the flags. * * @param <T> the type of the node to apply the flags to. * @param node the node to apply the flags to. Must be the topmost node on * the stack. * * @return the passed in node, or a modified node (if any flags were modified) */
public <T extends LexicalContextNode & Flags<T>> T applyTopFlags(final T node) { assert node == peek(); return node.setFlag(this, flags[sp - 1]); }
Return the top element in the context.
Returns:the node that was pushed last
/** * Return the top element in the context. * * @return the node that was pushed last */
public LexicalContextNode peek() { return stack[sp - 1]; }
Check if a node is in the lexical context.
Params:
  • node – node to check for
Returns:true if in the context
/** * Check if a node is in the lexical context. * * @param node node to check for * * @return {@code true} if in the context */
public boolean contains(final LexicalContextNode node) { for (int i = 0; i < sp; i++) { if (stack[i] == node) { return true; } } return false; }
Replace a node on the lexical context with a new one. Normally you should try to engineer IR traversals so this isn't needed
Params:
  • oldNode – old node
  • newNode – new node
Returns:the new node
/** * Replace a node on the lexical context with a new one. Normally * you should try to engineer IR traversals so this isn't needed * * @param oldNode old node * @param newNode new node * * @return the new node */
public LexicalContextNode replace(final LexicalContextNode oldNode, final LexicalContextNode newNode) { for (int i = sp - 1; i >= 0; i--) { if (stack[i] == oldNode) { assert i == sp - 1 : "violation of contract - we always expect to find the replacement node on top of the lexical context stack: " + newNode + " has " + stack[i + 1].getClass() + " above it"; stack[i] = newNode; break; } } return newNode; }
Returns an iterator over all blocks in the context, with the top block (innermost lexical context) first.
Returns:an iterator over all blocks in the context.
/** * Returns an iterator over all blocks in the context, with the top block * (innermost lexical context) first. * * @return an iterator over all blocks in the context. */
public Iterator<Block> getBlocks() { return new NodeIterator<>(Block.class); }
Returns an iterator over all functions in the context, with the top (innermost open) function first.
Returns:an iterator over all functions in the context.
/** * Returns an iterator over all functions in the context, with the top * (innermost open) function first. * * @return an iterator over all functions in the context. */
public Iterator<FunctionNode> getFunctions() { return new NodeIterator<>(FunctionNode.class); }
Get the parent block for the current lexical context block
Returns:parent block
/** * Get the parent block for the current lexical context block * * @return parent block */
public Block getParentBlock() { final Iterator<Block> iter = new NodeIterator<>(Block.class, getCurrentFunction()); iter.next(); return iter.hasNext() ? iter.next() : null; }
Gets the label node of the current block.
Returns:the label node of the current block, if it is labeled. Otherwise returns null.
/** * Gets the label node of the current block. * * @return the label node of the current block, if it is labeled. Otherwise * returns {@code null}. */
public LabelNode getCurrentBlockLabelNode() { assert stack[sp - 1] instanceof Block; if(sp < 2) { return null; } final LexicalContextNode parent = stack[sp - 2]; return parent instanceof LabelNode ? (LabelNode)parent : null; }
Returns an iterator over all ancestors block of the given block, with its parent block first.
Params:
  • block – the block whose ancestors are returned
Returns:an iterator over all ancestors block of the given block.
/** * Returns an iterator over all ancestors block of the given block, with its * parent block first. * * @param block the block whose ancestors are returned * * @return an iterator over all ancestors block of the given block. */
public Iterator<Block> getAncestorBlocks(final Block block) { final Iterator<Block> iter = getBlocks(); while (iter.hasNext()) { final Block b = iter.next(); if (block == b) { return iter; } } throw new AssertionError("Block is not on the current lexical context stack"); }
Returns an iterator over a block and all its ancestors blocks, with the block first.
Params:
  • block – the block that is the starting point of the iteration.
Returns:an iterator over a block and all its ancestors.
/** * Returns an iterator over a block and all its ancestors blocks, with the * block first. * * @param block the block that is the starting point of the iteration. * * @return an iterator over a block and all its ancestors. */
public Iterator<Block> getBlocks(final Block block) { final Iterator<Block> iter = getAncestorBlocks(block); return new Iterator<Block>() { boolean blockReturned = false; @Override public boolean hasNext() { return iter.hasNext() || !blockReturned; } @Override public Block next() { if (blockReturned) { return iter.next(); } blockReturned = true; return block; } @Override public void remove() { throw new UnsupportedOperationException(); } }; }
Get the function for this block.
Params:
  • block – block for which to get function
Returns:function for block
/** * Get the function for this block. * * @param block block for which to get function * * @return function for block */
public FunctionNode getFunction(final Block block) { final Iterator<LexicalContextNode> iter = new NodeIterator<>(LexicalContextNode.class); while (iter.hasNext()) { final LexicalContextNode next = iter.next(); if (next == block) { while (iter.hasNext()) { final LexicalContextNode next2 = iter.next(); if (next2 instanceof FunctionNode) { return (FunctionNode)next2; } } } } assert false; return null; }
Returns:the innermost block in the context.
/** * @return the innermost block in the context. */
public Block getCurrentBlock() { return getBlocks().next(); }
Returns:the innermost function in the context.
/** * @return the innermost function in the context. */
public FunctionNode getCurrentFunction() { for (int i = sp - 1; i >= 0; i--) { if (stack[i] instanceof FunctionNode) { return (FunctionNode) stack[i]; } } return null; }
Get the block in which a symbol is defined.
Params:
  • symbol – symbol
Returns:block in which the symbol is defined, assert if no such block in context.
/** * Get the block in which a symbol is defined. * * @param symbol symbol * * @return block in which the symbol is defined, assert if no such block in * context. */
public Block getDefiningBlock(final Symbol symbol) { final String name = symbol.getName(); for (final Iterator<Block> it = getBlocks(); it.hasNext();) { final Block next = it.next(); if (next.getExistingSymbol(name) == symbol) { return next; } } throw new AssertionError("Couldn't find symbol " + name + " in the context"); }
Get the function in which a symbol is defined.
Params:
  • symbol – symbol
Returns:function node in which this symbol is defined, assert if no such symbol exists in context.
/** * Get the function in which a symbol is defined. * * @param symbol symbol * * @return function node in which this symbol is defined, assert if no such * symbol exists in context. */
public FunctionNode getDefiningFunction(final Symbol symbol) { final String name = symbol.getName(); for (final Iterator<LexicalContextNode> iter = new NodeIterator<>(LexicalContextNode.class); iter.hasNext();) { final LexicalContextNode next = iter.next(); if (next instanceof Block && ((Block)next).getExistingSymbol(name) == symbol) { while (iter.hasNext()) { final LexicalContextNode next2 = iter.next(); if (next2 instanceof FunctionNode) { return (FunctionNode)next2; } } throw new AssertionError("Defining block for symbol " + name + " has no function in the context"); } } throw new AssertionError("Couldn't find symbol " + name + " in the context"); }
Is the topmost lexical context element a function body?
Returns:true if function body.
/** * Is the topmost lexical context element a function body? * * @return {@code true} if function body. */
public boolean isFunctionBody() { return getParentBlock() == null; }
Is the topmost lexical context element body of a SplitNode?
Returns:true if it's the body of a split node.
/** * Is the topmost lexical context element body of a SplitNode? * * @return {@code true} if it's the body of a split node. */
public boolean isSplitBody() { return sp >= 2 && stack[sp - 1] instanceof Block && stack[sp - 2] instanceof SplitNode; }
Get the parent function for a function in the lexical context.
Params:
  • functionNode – function for which to get parent
Returns:parent function of functionNode or null if none (e.g., if functionNode is the program).
/** * Get the parent function for a function in the lexical context. * * @param functionNode function for which to get parent * * @return parent function of functionNode or {@code null} if none (e.g., if * functionNode is the program). */
public FunctionNode getParentFunction(final FunctionNode functionNode) { final Iterator<FunctionNode> iter = new NodeIterator<>(FunctionNode.class); while (iter.hasNext()) { final FunctionNode next = iter.next(); if (next == functionNode) { return iter.hasNext() ? iter.next() : null; } } assert false; return null; }
Count the number of scopes until a given node. Note that this method is solely used to figure out the number of scopes that need to be explicitly popped in order to perform a break or continue jump within the current bytecode method. For this reason, the method returns 0 if it encounters a SplitNode between the current location and the break/continue target.
Params:
  • until – node to stop counting at. Must be within the current function.
Returns:number of with scopes encountered in the context.
/** * Count the number of scopes until a given node. Note that this method is * solely used to figure out the number of scopes that need to be explicitly * popped in order to perform a break or continue jump within the current * bytecode method. For this reason, the method returns 0 if it encounters a * {@code SplitNode} between the current location and the break/continue * target. * * @param until node to stop counting at. Must be within the current function. * * @return number of with scopes encountered in the context. */
public int getScopeNestingLevelTo(final LexicalContextNode until) { assert until != null; //count the number of with nodes until "until" is hit int n = 0; for (final Iterator<LexicalContextNode> iter = getAllNodes(); iter.hasNext();) { final LexicalContextNode node = iter.next(); if (node == until) { break; } assert !(node instanceof FunctionNode); // Can't go outside current function if (node instanceof WithNode || node instanceof Block && ((Block)node).needsScope()) { n++; } } return n; } private BreakableNode getBreakable() { for (final NodeIterator<BreakableNode> iter = new NodeIterator<>(BreakableNode.class, getCurrentFunction()); iter.hasNext(); ) { final BreakableNode next = iter.next(); if (next.isBreakableWithoutLabel()) { return next; } } return null; }
Check whether the lexical context is currently inside a loop.
Returns:true if inside a loop
/** * Check whether the lexical context is currently inside a loop. * * @return {@code true} if inside a loop */
public boolean inLoop() { return getCurrentLoop() != null; }
Returns:the loop header of the current loop, or null if not inside a loop.
/** * @return the loop header of the current loop, or {@code null} if not * inside a loop. */
public LoopNode getCurrentLoop() { final Iterator<LoopNode> iter = new NodeIterator<>(LoopNode.class, getCurrentFunction()); return iter.hasNext() ? iter.next() : null; }
Find the breakable node corresponding to this label.
Params:
  • labelName – name of the label to search for. If null, the closest breakable node will be returned unconditionally, e.g., a while loop with no label.
Returns:closest breakable node.
/** * Find the breakable node corresponding to this label. * * @param labelName name of the label to search for. If {@code null}, the * closest breakable node will be returned unconditionally, e.g., a * while loop with no label. * * @return closest breakable node. */
public BreakableNode getBreakable(final String labelName) { if (labelName != null) { final LabelNode foundLabel = findLabel(labelName); if (foundLabel != null) { // iterate to the nearest breakable to the foundLabel BreakableNode breakable = null; for (final NodeIterator<BreakableNode> iter = new NodeIterator<>(BreakableNode.class, foundLabel); iter.hasNext(); ) { breakable = iter.next(); } return breakable; } return null; } return getBreakable(); } private LoopNode getContinueTo() { return getCurrentLoop(); }
Find the continue target node corresponding to this label.
Params:
  • labelName – label name to search for. If null the closest loop node will be returned unconditionally, e.g., a while loop with no label.
Returns:closest continue target node.
/** * Find the continue target node corresponding to this label. * * @param labelName label name to search for. If {@code null} the closest * loop node will be returned unconditionally, e.g., a while loop * with no label. * * @return closest continue target node. */
public LoopNode getContinueTo(final String labelName) { if (labelName != null) { final LabelNode foundLabel = findLabel(labelName); if (foundLabel != null) { // iterate to the nearest loop to the foundLabel LoopNode loop = null; for (final NodeIterator<LoopNode> iter = new NodeIterator<>(LoopNode.class, foundLabel); iter.hasNext(); ) { loop = iter.next(); } return loop; } return null; } return getContinueTo(); }
Find the inlined finally block node corresponding to this label.
Params:
  • labelName – label name to search for. Must not be null.
Returns:closest inlined finally block with the given label.
/** * Find the inlined finally block node corresponding to this label. * * @param labelName label name to search for. Must not be {@code null}. * * @return closest inlined finally block with the given label. */
public Block getInlinedFinally(final String labelName) { for (final NodeIterator<TryNode> iter = new NodeIterator<>(TryNode.class); iter.hasNext(); ) { final Block inlinedFinally = iter.next().getInlinedFinally(labelName); if (inlinedFinally != null) { return inlinedFinally; } } return null; }
Find the try node for an inlined finally block corresponding to this label.
Params:
  • labelName – label name to search for. Must not be null.
Returns:the try node to which the labelled inlined finally block belongs.
/** * Find the try node for an inlined finally block corresponding to this label. * * @param labelName label name to search for. Must not be {@code null}. * * @return the try node to which the labelled inlined finally block belongs. */
public TryNode getTryNodeForInlinedFinally(final String labelName) { for (final NodeIterator<TryNode> iter = new NodeIterator<>(TryNode.class); iter.hasNext(); ) { final TryNode tryNode = iter.next(); if (tryNode.getInlinedFinally(labelName) != null) { return tryNode; } } return null; }
Check the lexical context for a given label node by name.
Params:
  • name – name of the label.
Returns:LabelNode if found, null otherwise.
/** * Check the lexical context for a given label node by name. * * @param name name of the label. * * @return LabelNode if found, {@code null} otherwise. */
private LabelNode findLabel(final String name) { for (final Iterator<LabelNode> iter = new NodeIterator<>(LabelNode.class, getCurrentFunction()); iter.hasNext(); ) { final LabelNode next = iter.next(); if (next.getLabelName().equals(name)) { return next; } } return null; }
Checks whether a given target is a jump destination that lies outside a given split node.
Params:
  • splitNode – the split node.
  • target – the target node.
Returns:true if target resides outside the split node.
/** * Checks whether a given target is a jump destination that lies outside a * given split node. * * @param splitNode the split node. * @param target the target node. * * @return {@code true} if target resides outside the split node. */
public boolean isExternalTarget(final SplitNode splitNode, final BreakableNode target) { for (int i = sp; i-- > 0;) { final LexicalContextNode next = stack[i]; if (next == splitNode) { return true; } else if (next == target) { return false; } else if (next instanceof TryNode) { for(final Block inlinedFinally: ((TryNode)next).getInlinedFinallies()) { if (TryNode.getLabelledInlinedFinallyBlock(inlinedFinally) == target) { return false; } } } } throw new AssertionError(target + " was expected in lexical context " + LexicalContext.this + " but wasn't"); }
Checks whether the current context is inside a switch statement without explicit blocks (curly braces).
Returns:true if in unprotected switch statement.
/** * Checks whether the current context is inside a switch statement without * explicit blocks (curly braces). * * @return {@code true} if in unprotected switch statement. */
public boolean inUnprotectedSwitchContext() { for (int i = sp; i > 0; i--) { final LexicalContextNode next = stack[i]; if (next instanceof Block) { return stack[i - 1] instanceof SwitchNode; } } return false; } @Override public String toString() { final StringBuffer sb = new StringBuffer(); sb.append("[ "); for (int i = 0; i < sp; i++) { final Object node = stack[i]; sb.append(node.getClass().getSimpleName()); sb.append('@'); sb.append(Debug.id(node)); sb.append(':'); if (node instanceof FunctionNode) { final FunctionNode fn = (FunctionNode)node; final Source source = fn.getSource(); String src = source.toString(); if (src.contains(File.pathSeparator)) { src = src.substring(src.lastIndexOf(File.pathSeparator)); } src += ' '; src += fn.getLineNumber(); sb.append(src); } sb.append(' '); } sb.append(" ==> ]"); return sb.toString(); } private class NodeIterator <T extends LexicalContextNode> implements Iterator<T> { private int index; private T next; private final Class<T> clazz; private LexicalContextNode until; NodeIterator(final Class<T> clazz) { this(clazz, null); } NodeIterator(final Class<T> clazz, final LexicalContextNode until) { this.index = sp - 1; this.clazz = clazz; this.until = until; this.next = findNext(); } @Override public boolean hasNext() { return next != null; } @Override public T next() { if (next == null) { throw new NoSuchElementException(); } final T lnext = next; next = findNext(); return lnext; } @SuppressWarnings("unchecked") private T findNext() { for (int i = index; i >= 0; i--) { final Object node = stack[i]; if (node == until) { return null; } if (clazz.isAssignableFrom(node.getClass())) { index = i - 1; return (T)node; } } return null; } @Override public void remove() { throw new UnsupportedOperationException(); } } }