/* * Copyright (C) 2015 Google, Inc. * * Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except * in compliance with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software distributed under the License * is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express * or implied. See the License for the specific language governing permissions and limitations under * the License. */ /* * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ package com.google.escapevelocity; import static com.google.escapevelocity.Node.emptyNode; import com.google.escapevelocity.DirectiveNode.ForEachNode; import com.google.escapevelocity.DirectiveNode.IfNode; import com.google.escapevelocity.DirectiveNode.MacroCallNode; import com.google.escapevelocity.DirectiveNode.SetNode; import com.google.escapevelocity.TokenNode.CommentTokenNode; import com.google.escapevelocity.TokenNode.ElseIfTokenNode; import com.google.escapevelocity.TokenNode.ElseTokenNode; import com.google.escapevelocity.TokenNode.EndTokenNode; import com.google.escapevelocity.TokenNode.EofNode; import com.google.escapevelocity.TokenNode.ForEachTokenNode; import com.google.escapevelocity.TokenNode.IfOrElseIfTokenNode; import com.google.escapevelocity.TokenNode.IfTokenNode; import com.google.escapevelocity.TokenNode.MacroDefinitionTokenNode; import com.google.escapevelocity.TokenNode.NestedTokenNode; import com.google.common.base.CharMatcher; import com.google.common.collect.ImmutableList; import com.google.common.collect.ImmutableSet; import com.google.common.collect.Iterables; import java.util.Map; import java.util.Set; import java.util.TreeMap; /** * The second phase of parsing. See {@link Parser#parse()} for a description of the phases and why * we need them. * * @author emcmanus@google.com (Éamonn McManus) */ class Reparser { private static final ImmutableSet> END_SET = ImmutableSet.>of(EndTokenNode.class); private static final ImmutableSet> EOF_SET = ImmutableSet.>of(EofNode.class); private static final ImmutableSet> ELSE_ELSE_IF_END_SET = ImmutableSet.>of( ElseTokenNode.class, ElseIfTokenNode.class, EndTokenNode.class); /** * The nodes that make up the input sequence. Nodes are removed one by one from this list as * parsing proceeds. At any time, {@link #currentNode} is the node being examined. */ private final ImmutableList nodes; /** * The index of the node we are currently looking at while parsing. */ private int nodeIndex; /** * Macros are removed from the input as they are found. They do not appear in the output parse * tree. Macro definitions are not executed in place but are all applied before template rendering * starts. This means that a macro can be referenced before it is defined. */ private final Map macros; Reparser(ImmutableList nodes) { this(nodes, new TreeMap<>()); } private Reparser(ImmutableList nodes, Map macros) { this.nodes = removeSpaceBeforeSet(nodes); this.nodeIndex = 0; this.macros = macros; } Template reparse() { Node root = reparseNodes(); linkMacroCalls(); return new Template(root); } private Node reparseNodes() { return parseTo(EOF_SET, new EofNode((String) null, 1)); } /** * Returns a copy of the given list where spaces have been moved where appropriate after {@code * #set}. This hack is needed to match what appears to be special treatment in Apache Velocity of * spaces before {@code #set} directives. If you have thing whitespace {@code #set}, * then the whitespace is deleted if the thing is a comment ({@code ##...\n}); a reference * ({@code $x} or {@code $x.foo} etc); a macro definition; or another {@code #set}. */ private static ImmutableList removeSpaceBeforeSet(ImmutableList nodes) { assert Iterables.getLast(nodes) instanceof EofNode; // Since the last node is EofNode, the i + 1 and i + 2 accesses below are safe. ImmutableList.Builder newNodes = ImmutableList.builder(); for (int i = 0; i < nodes.size(); i++) { Node nodeI = nodes.get(i); newNodes.add(nodeI); if (shouldDeleteSpaceBetweenThisAndSet(nodeI) && isWhitespaceLiteral(nodes.get(i + 1)) && nodes.get(i + 2) instanceof SetNode) { // Skip the space. i++; } } return newNodes.build(); } private static boolean shouldDeleteSpaceBetweenThisAndSet(Node node) { return node instanceof CommentTokenNode || node instanceof ReferenceNode || node instanceof SetNode || node instanceof MacroDefinitionTokenNode; } private static boolean isWhitespaceLiteral(Node node) { if (node instanceof ConstantExpressionNode) { Object constant = node.evaluate(null); return constant instanceof String && CharMatcher.whitespace().matchesAllOf((String) constant); } return false; } /** * Parse subtrees until one of the token types in {@code stopSet} is encountered. * If this is the top level, {@code stopSet} will include {@link EofNode} so parsing will stop * when it reaches the end of the input. Otherwise, if an {@code EofNode} is encountered it is an * error because we have something like {@code #if} without {@code #end}. * * @param stopSet the kinds of tokens that will stop the parse. For example, if we are parsing * after an {@code #if}, we will stop at any of {@code #else}, {@code #elseif}, * or {@code #end}. * @param forWhat the token that triggered this call, for example the {@code #if} whose * {@code #end} etc we are looking for. * * @return a Node that is the concatenation of the parsed subtrees */ private Node parseTo(Set> stopSet, TokenNode forWhat) { ImmutableList.Builder nodeList = ImmutableList.builder(); while (true) { Node currentNode = currentNode(); if (stopSet.contains(currentNode.getClass())) { break; } if (currentNode instanceof EofNode) { throw new ParseException( "Reached end of file while parsing " + forWhat.name(), forWhat.resourceName, forWhat.lineNumber); } Node parsed; if (currentNode instanceof TokenNode) { parsed = parseTokenNode(); } else { parsed = currentNode; nextNode(); } nodeList.add(parsed); } return Node.cons(forWhat.resourceName, forWhat.lineNumber, nodeList.build()); } private Node currentNode() { return nodes.get(nodeIndex); } private Node nextNode() { Node currentNode = currentNode(); if (currentNode instanceof EofNode) { return currentNode; } else { nodeIndex++; return currentNode(); } } private Node parseTokenNode() { TokenNode tokenNode = (TokenNode) currentNode(); nextNode(); if (tokenNode instanceof CommentTokenNode) { return emptyNode(tokenNode.resourceName, tokenNode.lineNumber); } else if (tokenNode instanceof IfTokenNode) { return parseIfOrElseIf((IfTokenNode) tokenNode); } else if (tokenNode instanceof ForEachTokenNode) { return parseForEach((ForEachTokenNode) tokenNode); } else if (tokenNode instanceof NestedTokenNode) { return parseNested((NestedTokenNode) tokenNode); } else if (tokenNode instanceof MacroDefinitionTokenNode) { return parseMacroDefinition((MacroDefinitionTokenNode) tokenNode); } else { throw new IllegalArgumentException( "Unexpected token: " + tokenNode.name() + " on line " + tokenNode.lineNumber); } } private Node parseForEach(ForEachTokenNode forEach) { Node body = parseTo(END_SET, forEach); nextNode(); // Skip #end return new ForEachNode( forEach.resourceName, forEach.lineNumber, forEach.var, forEach.collection, body); } private Node parseIfOrElseIf(IfOrElseIfTokenNode ifOrElseIf) { Node truePart = parseTo(ELSE_ELSE_IF_END_SET, ifOrElseIf); Node falsePart; Node token = currentNode(); nextNode(); // Skip #else or #elseif (cond) or #end. if (token instanceof EndTokenNode) { falsePart = emptyNode(token.resourceName, token.lineNumber); } else if (token instanceof ElseTokenNode) { falsePart = parseTo(END_SET, ifOrElseIf); nextNode(); // Skip #end } else if (token instanceof ElseIfTokenNode) { // We've seen #if (condition1) ... #elseif (condition2). currentToken is the first token // after (condition2). We pretend that we've just seen #if (condition2) and parse out // the remainder (which might have further #elseif and final #else). Then we pretend that // we actually saw #if (condition1) ... #else #if (condition2) ...remainder ... #end #end. falsePart = parseIfOrElseIf((ElseIfTokenNode) token); } else { throw new AssertionError(currentNode()); } return new IfNode( ifOrElseIf.resourceName, ifOrElseIf.lineNumber, ifOrElseIf.condition, truePart, falsePart); } // This is a #parse("foo.vm") directive. We've already done the first phase of parsing on the // contents of foo.vm. Now we need to do the second phase, and insert the result into the // reparsed nodes. We can call Reparser recursively, but we must ensure that any macros found // are added to the containing Reparser's macro definitions. private Node parseNested(NestedTokenNode nested) { Reparser reparser = new Reparser(nested.nodes, this.macros); return reparser.reparseNodes(); } private Node parseMacroDefinition(MacroDefinitionTokenNode macroDefinition) { Node body = parseTo(END_SET, macroDefinition); nextNode(); // Skip #end if (!macros.containsKey(macroDefinition.name)) { Macro macro = new Macro( macroDefinition.lineNumber, macroDefinition.name, macroDefinition.parameterNames, body); macros.put(macroDefinition.name, macro); } return emptyNode(macroDefinition.resourceName, macroDefinition.lineNumber); } private void linkMacroCalls() { for (Node node : nodes) { if (node instanceof MacroCallNode) { linkMacroCall((MacroCallNode) node); } } } private void linkMacroCall(MacroCallNode macroCall) { Macro macro = macros.get(macroCall.name()); if (macro == null) { throw new ParseException( "#" + macroCall.name() + " is neither a standard directive nor a macro that has been defined", macroCall.resourceName, macroCall.lineNumber); } if (macro.parameterCount() != macroCall.argumentCount()) { throw new ParseException( "Wrong number of arguments to #" + macroCall.name() + ": expected " + macro.parameterCount() + ", got " + macroCall.argumentCount(), macroCall.resourceName, macroCall.lineNumber); } macroCall.setMacro(macro); } }