SPIGOT-6972: Root command nodes can leak to client

By: md_5 <git@md-5.net>
This commit is contained in:
CraftBukkit/Spigot 2022-04-17 18:57:35 +10:00
parent ef79686997
commit e935d60d1d
2 changed files with 722 additions and 16 deletions

View file

@ -0,0 +1,708 @@
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT license.
package com.mojang.brigadier;
// CHECKSTYLE:OFF
import com.mojang.brigadier.builder.LiteralArgumentBuilder;
import com.mojang.brigadier.context.CommandContext;
import com.mojang.brigadier.context.CommandContextBuilder;
import com.mojang.brigadier.context.SuggestionContext;
import com.mojang.brigadier.exceptions.CommandSyntaxException;
import com.mojang.brigadier.suggestion.Suggestions;
import com.mojang.brigadier.suggestion.SuggestionsBuilder;
import com.mojang.brigadier.tree.CommandNode;
import com.mojang.brigadier.tree.LiteralCommandNode;
import com.mojang.brigadier.tree.RootCommandNode;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
import java.util.LinkedHashMap;
import java.util.LinkedHashSet;
import java.util.List;
import java.util.Locale;
import java.util.Map;
import java.util.Set;
import java.util.concurrent.CompletableFuture;
import java.util.function.Predicate;
import java.util.stream.Collectors;
/**
* The core command dispatcher, for registering, parsing, and executing commands.
*
* @param <S> a custom "source" type, such as a user or originator of a command
*/
public class CommandDispatcher<S> {
/**
* The string required to separate individual arguments in an input string
*
* @see #ARGUMENT_SEPARATOR_CHAR
*/
public static final String ARGUMENT_SEPARATOR = " ";
/**
* The char required to separate individual arguments in an input string
*
* @see #ARGUMENT_SEPARATOR
*/
public static final char ARGUMENT_SEPARATOR_CHAR = ' ';
private static final String USAGE_OPTIONAL_OPEN = "[";
private static final String USAGE_OPTIONAL_CLOSE = "]";
private static final String USAGE_REQUIRED_OPEN = "(";
private static final String USAGE_REQUIRED_CLOSE = ")";
private static final String USAGE_OR = "|";
private final RootCommandNode<S> root;
private final Predicate<CommandNode<S>> hasCommand = new Predicate<CommandNode<S>>() {
@Override
public boolean test(final CommandNode<S> input) {
return input != null && (input.getCommand() != null || input.getChildren().stream().anyMatch(hasCommand));
}
};
private ResultConsumer<S> consumer = (c, s, r) -> {
};
/**
* Create a new {@link CommandDispatcher} with the specified root node.
*
* <p>This is often useful to copy existing or pre-defined command trees.</p>
*
* @param root the existing {@link RootCommandNode} to use as the basis for this tree
*/
public CommandDispatcher(final RootCommandNode<S> root) {
this.root = root;
}
/**
* Creates a new {@link CommandDispatcher} with an empty command tree.
*/
public CommandDispatcher() {
this(new RootCommandNode<>());
}
/**
* Utility method for registering new commands.
*
* <p>This is a shortcut for calling {@link RootCommandNode#addChild(CommandNode)} after building the provided {@code command}.</p>
*
* <p>As {@link RootCommandNode} can only hold literals, this method will only allow literal arguments.</p>
*
* @param command a literal argument builder to add to this command tree
* @return the node added to this tree
*/
public LiteralCommandNode<S> register(final LiteralArgumentBuilder<S> command) {
final LiteralCommandNode<S> build = command.build();
root.addChild(build);
return build;
}
/**
* Sets a callback to be informed of the result of every command.
*
* @param consumer the new result consumer to be called
*/
public void setConsumer(final ResultConsumer<S> consumer) {
this.consumer = consumer;
}
/**
* Parses and executes a given command.
*
* <p>This is a shortcut to first {@link #parse(StringReader, Object)} and then {@link #execute(ParseResults)}.</p>
*
* <p>It is recommended to parse and execute as separate steps, as parsing is often the most expensive step, and easiest to cache.</p>
*
* <p>If this command returns a value, then it successfully executed something. If it could not parse the command, or the execution was a failure,
* then an exception will be thrown. Most exceptions will be of type {@link CommandSyntaxException}, but it is possible that a {@link RuntimeException}
* may bubble up from the result of a command. The meaning behind the returned result is arbitrary, and will depend
* entirely on what command was performed.</p>
*
* <p>If the command passes through a node that is {@link CommandNode#isFork()} then it will be 'forked'.
* A forked command will not bubble up any {@link CommandSyntaxException}s, and the 'result' returned will turn into
* 'amount of successful commands executes'.</p>
*
* <p>After each and any command is ran, a registered callback given to {@link #setConsumer(ResultConsumer)}
* will be notified of the result and success of the command. You can use that method to gather more meaningful
* results than this method will return, especially when a command forks.</p>
*
* @param input a command string to parse &amp; execute
* @param source a custom "source" object, usually representing the originator of this command
* @return a numeric result from a "command" that was performed
* @throws CommandSyntaxException if the command failed to parse or execute
* @throws RuntimeException if the command failed to execute and was not handled gracefully
* @see #parse(String, Object)
* @see #parse(StringReader, Object)
* @see #execute(ParseResults)
* @see #execute(StringReader, Object)
*/
public int execute(final String input, final S source) throws CommandSyntaxException {
return execute(new StringReader(input), source);
}
/**
* Parses and executes a given command.
*
* <p>This is a shortcut to first {@link #parse(StringReader, Object)} and then {@link #execute(ParseResults)}.</p>
*
* <p>It is recommended to parse and execute as separate steps, as parsing is often the most expensive step, and easiest to cache.</p>
*
* <p>If this command returns a value, then it successfully executed something. If it could not parse the command, or the execution was a failure,
* then an exception will be thrown. Most exceptions will be of type {@link CommandSyntaxException}, but it is possible that a {@link RuntimeException}
* may bubble up from the result of a command. The meaning behind the returned result is arbitrary, and will depend
* entirely on what command was performed.</p>
*
* <p>If the command passes through a node that is {@link CommandNode#isFork()} then it will be 'forked'.
* A forked command will not bubble up any {@link CommandSyntaxException}s, and the 'result' returned will turn into
* 'amount of successful commands executes'.</p>
*
* <p>After each and any command is ran, a registered callback given to {@link #setConsumer(ResultConsumer)}
* will be notified of the result and success of the command. You can use that method to gather more meaningful
* results than this method will return, especially when a command forks.</p>
*
* @param input a command string to parse &amp; execute
* @param source a custom "source" object, usually representing the originator of this command
* @return a numeric result from a "command" that was performed
* @throws CommandSyntaxException if the command failed to parse or execute
* @throws RuntimeException if the command failed to execute and was not handled gracefully
* @see #parse(String, Object)
* @see #parse(StringReader, Object)
* @see #execute(ParseResults)
* @see #execute(String, Object)
*/
public int execute(final StringReader input, final S source) throws CommandSyntaxException {
final ParseResults<S> parse = parse(input, source);
return execute(parse);
}
/**
* Executes a given pre-parsed command.
*
* <p>If this command returns a value, then it successfully executed something. If the execution was a failure,
* then an exception will be thrown.
* Most exceptions will be of type {@link CommandSyntaxException}, but it is possible that a {@link RuntimeException}
* may bubble up from the result of a command. The meaning behind the returned result is arbitrary, and will depend
* entirely on what command was performed.</p>
*
* <p>If the command passes through a node that is {@link CommandNode#isFork()} then it will be 'forked'.
* A forked command will not bubble up any {@link CommandSyntaxException}s, and the 'result' returned will turn into
* 'amount of successful commands executes'.</p>
*
* <p>After each and any command is ran, a registered callback given to {@link #setConsumer(ResultConsumer)}
* will be notified of the result and success of the command. You can use that method to gather more meaningful
* results than this method will return, especially when a command forks.</p>
*
* @param parse the result of a successful {@link #parse(StringReader, Object)}
* @return a numeric result from a "command" that was performed.
* @throws CommandSyntaxException if the command failed to parse or execute
* @throws RuntimeException if the command failed to execute and was not handled gracefully
* @see #parse(String, Object)
* @see #parse(StringReader, Object)
* @see #execute(String, Object)
* @see #execute(StringReader, Object)
*/
public int execute(final ParseResults<S> parse) throws CommandSyntaxException {
if (parse.getReader().canRead()) {
if (parse.getExceptions().size() == 1) {
throw parse.getExceptions().values().iterator().next();
} else if (parse.getContext().getRange().isEmpty()) {
throw CommandSyntaxException.BUILT_IN_EXCEPTIONS.dispatcherUnknownCommand().createWithContext(parse.getReader());
} else {
throw CommandSyntaxException.BUILT_IN_EXCEPTIONS.dispatcherUnknownArgument().createWithContext(parse.getReader());
}
}
int result = 0;
int successfulForks = 0;
boolean forked = false;
boolean foundCommand = false;
final String command = parse.getReader().getString();
final CommandContext<S> original = parse.getContext().build(command);
List<CommandContext<S>> contexts = Collections.singletonList(original);
ArrayList<CommandContext<S>> next = null;
while (contexts != null) {
final int size = contexts.size();
for (int i = 0; i < size; i++) {
final CommandContext<S> context = contexts.get(i);
final CommandContext<S> child = context.getChild();
if (child != null) {
forked |= context.isForked();
if (child.hasNodes()) {
foundCommand = true;
final RedirectModifier<S> modifier = context.getRedirectModifier();
if (modifier == null) {
if (next == null) {
next = new ArrayList<>(1);
}
next.add(child.copyFor(context.getSource()));
} else {
try {
final Collection<S> results = modifier.apply(context);
if (!results.isEmpty()) {
if (next == null) {
next = new ArrayList<>(results.size());
}
for (final S source : results) {
next.add(child.copyFor(source));
}
}
} catch (final CommandSyntaxException ex) {
consumer.onCommandComplete(context, false, 0);
if (!forked) {
throw ex;
}
}
}
}
} else if (context.getCommand() != null) {
foundCommand = true;
try {
final int value = context.getCommand().run(context);
result += value;
consumer.onCommandComplete(context, true, value);
successfulForks++;
} catch (final CommandSyntaxException ex) {
consumer.onCommandComplete(context, false, 0);
if (!forked) {
throw ex;
}
}
}
}
contexts = next;
next = null;
}
if (!foundCommand) {
consumer.onCommandComplete(original, false, 0);
throw CommandSyntaxException.BUILT_IN_EXCEPTIONS.dispatcherUnknownCommand().createWithContext(parse.getReader());
}
return forked ? successfulForks : result;
}
/**
* Parses a given command.
*
* <p>The result of this method can be cached, and it is advised to do so where appropriate. Parsing is often the
* most expensive step, and this allows you to essentially "precompile" a command if it will be ran often.</p>
*
* <p>If the command passes through a node that is {@link CommandNode#isFork()} then the resulting context will be marked as 'forked'.
* Forked contexts may contain child contexts, which may be modified by the {@link RedirectModifier} attached to the fork.</p>
*
* <p>Parsing a command can never fail, you will always be provided with a new {@link ParseResults}.
* However, that does not mean that it will always parse into a valid command. You should inspect the returned results
* to check for validity. If its {@link ParseResults#getReader()} {@link StringReader#canRead()} then it did not finish
* parsing successfully. You can use that position as an indicator to the user where the command stopped being valid.
* You may inspect {@link ParseResults#getExceptions()} if you know the parse failed, as it will explain why it could
* not find any valid commands. It may contain multiple exceptions, one for each "potential node" that it could have visited,
* explaining why it did not go down that node.</p>
*
* <p>When you eventually call {@link #execute(ParseResults)} with the result of this method, the above error checking
* will occur. You only need to inspect it yourself if you wish to handle that yourself.</p>
*
* @param command a command string to parse
* @param source a custom "source" object, usually representing the originator of this command
* @return the result of parsing this command
* @see #parse(StringReader, Object)
* @see #execute(ParseResults)
* @see #execute(String, Object)
*/
public ParseResults<S> parse(final String command, final S source) {
return parse(new StringReader(command), source);
}
/**
* Parses a given command.
*
* <p>The result of this method can be cached, and it is advised to do so where appropriate. Parsing is often the
* most expensive step, and this allows you to essentially "precompile" a command if it will be ran often.</p>
*
* <p>If the command passes through a node that is {@link CommandNode#isFork()} then the resulting context will be marked as 'forked'.
* Forked contexts may contain child contexts, which may be modified by the {@link RedirectModifier} attached to the fork.</p>
*
* <p>Parsing a command can never fail, you will always be provided with a new {@link ParseResults}.
* However, that does not mean that it will always parse into a valid command. You should inspect the returned results
* to check for validity. If its {@link ParseResults#getReader()} {@link StringReader#canRead()} then it did not finish
* parsing successfully. You can use that position as an indicator to the user where the command stopped being valid.
* You may inspect {@link ParseResults#getExceptions()} if you know the parse failed, as it will explain why it could
* not find any valid commands. It may contain multiple exceptions, one for each "potential node" that it could have visited,
* explaining why it did not go down that node.</p>
*
* <p>When you eventually call {@link #execute(ParseResults)} with the result of this method, the above error checking
* will occur. You only need to inspect it yourself if you wish to handle that yourself.</p>
*
* @param command a command string to parse
* @param source a custom "source" object, usually representing the originator of this command
* @return the result of parsing this command
* @see #parse(String, Object)
* @see #execute(ParseResults)
* @see #execute(String, Object)
*/
public ParseResults<S> parse(final StringReader command, final S source) {
final CommandContextBuilder<S> context = new CommandContextBuilder<>(this, source, root, command.getCursor());
return parseNodes(root, command, context);
}
private ParseResults<S> parseNodes(final CommandNode<S> node, final StringReader originalReader, final CommandContextBuilder<S> contextSoFar) {
final S source = contextSoFar.getSource();
Map<CommandNode<S>, CommandSyntaxException> errors = null;
List<ParseResults<S>> potentials = null;
final int cursor = originalReader.getCursor();
for (final CommandNode<S> child : node.getRelevantNodes(originalReader)) {
if (!child.canUse(source)) {
continue;
}
final CommandContextBuilder<S> context = contextSoFar.copy();
final StringReader reader = new StringReader(originalReader);
try {
try {
child.parse(reader, context);
} catch (final RuntimeException ex) {
throw CommandSyntaxException.BUILT_IN_EXCEPTIONS.dispatcherParseException().createWithContext(reader, ex.getMessage());
}
if (reader.canRead()) {
if (reader.peek() != ARGUMENT_SEPARATOR_CHAR) {
throw CommandSyntaxException.BUILT_IN_EXCEPTIONS.dispatcherExpectedArgumentSeparator().createWithContext(reader);
}
}
} catch (final CommandSyntaxException ex) {
if (errors == null) {
errors = new LinkedHashMap<>();
}
errors.put(child, ex);
reader.setCursor(cursor);
continue;
}
context.withCommand(child.getCommand());
if (reader.canRead(child.getRedirect() == null ? 2 : 1)) {
reader.skip();
if (child.getRedirect() != null) {
final CommandContextBuilder<S> childContext = new CommandContextBuilder<>(this, source, child.getRedirect(), reader.getCursor());
final ParseResults<S> parse = parseNodes(child.getRedirect(), reader, childContext);
context.withChild(parse.getContext());
return new ParseResults<>(context, parse.getReader(), parse.getExceptions());
} else {
final ParseResults<S> parse = parseNodes(child, reader, context);
if (potentials == null) {
potentials = new ArrayList<>(1);
}
potentials.add(parse);
}
} else {
if (potentials == null) {
potentials = new ArrayList<>(1);
}
potentials.add(new ParseResults<>(context, reader, Collections.emptyMap()));
}
}
if (potentials != null) {
if (potentials.size() > 1) {
potentials.sort((a, b) -> {
if (!a.getReader().canRead() && b.getReader().canRead()) {
return -1;
}
if (a.getReader().canRead() && !b.getReader().canRead()) {
return 1;
}
if (a.getExceptions().isEmpty() && !b.getExceptions().isEmpty()) {
return -1;
}
if (!a.getExceptions().isEmpty() && b.getExceptions().isEmpty()) {
return 1;
}
return 0;
});
}
return potentials.get(0);
}
return new ParseResults<>(contextSoFar, originalReader, errors == null ? Collections.emptyMap() : errors);
}
/**
* Gets all possible executable commands following the given node.
*
* <p>You may use {@link #getRoot()} as a target to get all usage data for the entire command tree.</p>
*
* <p>The returned syntax will be in "simple" form: {@code <param>} and {@code literal}. "Optional" nodes will be
* listed as multiple entries: the parent node, and the child nodes.
* For example, a required literal "foo" followed by an optional param "int" will be two nodes:</p>
* <ul>
* <li>{@code foo}</li>
* <li>{@code foo <int>}</li>
* </ul>
*
* <p>The path to the specified node will <b>not</b> be prepended to the output, as there can theoretically be many
* ways to reach a given node. It will only give you paths relative to the specified node, not absolute from root.</p>
*
* @param node target node to get child usage strings for
* @param source a custom "source" object, usually representing the originator of this command
* @param restricted if true, commands that the {@code source} cannot access will not be mentioned
* @return array of full usage strings under the target node
*/
public String[] getAllUsage(final CommandNode<S> node, final S source, final boolean restricted) {
final ArrayList<String> result = new ArrayList<>();
getAllUsage(node, source, result, "", restricted);
return result.toArray(new String[result.size()]);
}
private void getAllUsage(final CommandNode<S> node, final S source, final ArrayList<String> result, final String prefix, final boolean restricted) {
if (restricted && !node.canUse(source)) {
return;
}
if (node.getCommand() != null) {
result.add(prefix);
}
if (node.getRedirect() != null) {
final String redirect = node.getRedirect() == root ? "..." : "-> " + node.getRedirect().getUsageText();
result.add(prefix.isEmpty() ? node.getUsageText() + ARGUMENT_SEPARATOR + redirect : prefix + ARGUMENT_SEPARATOR + redirect);
} else if (!node.getChildren().isEmpty()) {
for (final CommandNode<S> child : node.getChildren()) {
getAllUsage(child, source, result, prefix.isEmpty() ? child.getUsageText() : prefix + ARGUMENT_SEPARATOR + child.getUsageText(), restricted);
}
}
}
/**
* Gets the possible executable commands from a specified node.
*
* <p>You may use {@link #getRoot()} as a target to get usage data for the entire command tree.</p>
*
* <p>The returned syntax will be in "smart" form: {@code <param>}, {@code literal}, {@code [optional]} and {@code (either|or)}.
* These forms may be mixed and matched to provide as much information about the child nodes as it can, without being too verbose.
* For example, a required literal "foo" followed by an optional param "int" can be compressed into one string:</p>
* <ul>
* <li>{@code foo [<int>]}</li>
* </ul>
*
* <p>The path to the specified node will <b>not</b> be prepended to the output, as there can theoretically be many
* ways to reach a given node. It will only give you paths relative to the specified node, not absolute from root.</p>
*
* <p>The returned usage will be restricted to only commands that the provided {@code source} can use.</p>
*
* @param node target node to get child usage strings for
* @param source a custom "source" object, usually representing the originator of this command
* @return array of full usage strings under the target node
*/
public Map<CommandNode<S>, String> getSmartUsage(final CommandNode<S> node, final S source) {
final Map<CommandNode<S>, String> result = new LinkedHashMap<>();
final boolean optional = node.getCommand() != null;
for (final CommandNode<S> child : node.getChildren()) {
final String usage = getSmartUsage(child, source, optional, false);
if (usage != null) {
result.put(child, usage);
}
}
return result;
}
private String getSmartUsage(final CommandNode<S> node, final S source, final boolean optional, final boolean deep) {
if (!node.canUse(source)) {
return null;
}
final String self = optional ? USAGE_OPTIONAL_OPEN + node.getUsageText() + USAGE_OPTIONAL_CLOSE : node.getUsageText();
final boolean childOptional = node.getCommand() != null;
final String open = childOptional ? USAGE_OPTIONAL_OPEN : USAGE_REQUIRED_OPEN;
final String close = childOptional ? USAGE_OPTIONAL_CLOSE : USAGE_REQUIRED_CLOSE;
if (!deep) {
if (node.getRedirect() != null) {
final String redirect = node.getRedirect() == root ? "..." : "-> " + node.getRedirect().getUsageText();
return self + ARGUMENT_SEPARATOR + redirect;
} else {
final Collection<CommandNode<S>> children = node.getChildren().stream().filter(c -> c.canUse(source)).collect(Collectors.toList());
if (children.size() == 1) {
final String usage = getSmartUsage(children.iterator().next(), source, childOptional, childOptional);
if (usage != null) {
return self + ARGUMENT_SEPARATOR + usage;
}
} else if (children.size() > 1) {
final Set<String> childUsage = new LinkedHashSet<>();
for (final CommandNode<S> child : children) {
final String usage = getSmartUsage(child, source, childOptional, true);
if (usage != null) {
childUsage.add(usage);
}
}
if (childUsage.size() == 1) {
final String usage = childUsage.iterator().next();
return self + ARGUMENT_SEPARATOR + (childOptional ? USAGE_OPTIONAL_OPEN + usage + USAGE_OPTIONAL_CLOSE : usage);
} else if (childUsage.size() > 1) {
final StringBuilder builder = new StringBuilder(open);
int count = 0;
for (final CommandNode<S> child : children) {
if (count > 0) {
builder.append(USAGE_OR);
}
builder.append(child.getUsageText());
count++;
}
if (count > 0) {
builder.append(close);
return self + ARGUMENT_SEPARATOR + builder.toString();
}
}
}
}
}
return self;
}
/**
* Gets suggestions for a parsed input string on what comes next.
*
* <p>As it is ultimately up to custom argument types to provide suggestions, it may be an asynchronous operation,
* for example getting in-game data or player names etc. As such, this method returns a future and no guarantees
* are made to when or how the future completes.</p>
*
* <p>The suggestions provided will be in the context of the end of the parsed input string, but may suggest
* new or replacement strings for earlier in the input string. For example, if the end of the string was
* {@code foobar} but an argument preferred it to be {@code minecraft:foobar}, it will suggest a replacement for that
* whole segment of the input.</p>
*
* @param parse the result of a {@link #parse(StringReader, Object)}
* @return a future that will eventually resolve into a {@link Suggestions} object
*/
public CompletableFuture<Suggestions> getCompletionSuggestions(final ParseResults<S> parse) {
return getCompletionSuggestions(parse, parse.getReader().getTotalLength());
}
public CompletableFuture<Suggestions> getCompletionSuggestions(final ParseResults<S> parse, int cursor) {
final CommandContextBuilder<S> context = parse.getContext();
final SuggestionContext<S> nodeBeforeCursor = context.findSuggestionContext(cursor);
final CommandNode<S> parent = nodeBeforeCursor.parent;
final int start = Math.min(nodeBeforeCursor.startPos, cursor);
final String fullInput = parse.getReader().getString();
final String truncatedInput = fullInput.substring(0, cursor);
final String truncatedInputLowerCase = truncatedInput.toLowerCase(Locale.ROOT);
@SuppressWarnings("unchecked") final CompletableFuture<Suggestions>[] futures = new CompletableFuture[parent.getChildren().size()];
int i = 0;
for (final CommandNode<S> node : parent.getChildren()) {
CompletableFuture<Suggestions> future = Suggestions.empty();
try {
if (node.canUse(parse.getContext().getSource())) future = node.listSuggestions(context.build(truncatedInput), new SuggestionsBuilder(truncatedInput, truncatedInputLowerCase, start)); // CraftBukkit
} catch (final CommandSyntaxException ignored) {
}
futures[i++] = future;
}
final CompletableFuture<Suggestions> result = new CompletableFuture<>();
CompletableFuture.allOf(futures).thenRun(() -> {
final List<Suggestions> suggestions = new ArrayList<>();
for (final CompletableFuture<Suggestions> future : futures) {
suggestions.add(future.join());
}
result.complete(Suggestions.merge(fullInput, suggestions));
});
return result;
}
/**
* Gets the root of this command tree.
*
* <p>This is often useful as a target of a {@link com.mojang.brigadier.builder.ArgumentBuilder#redirect(CommandNode)},
* {@link #getAllUsage(CommandNode, Object, boolean)} or {@link #getSmartUsage(CommandNode, Object)}.
* You may also use it to clone the command tree via {@link #CommandDispatcher(RootCommandNode)}.</p>
*
* @return root of the command tree
*/
public RootCommandNode<S> getRoot() {
return root;
}
/**
* Finds a valid path to a given node on the command tree.
*
* <p>There may theoretically be multiple paths to a node on the tree, especially with the use of forking or redirecting.
* As such, this method makes no guarantees about which path it finds. It will not look at forks or redirects,
* and find the first instance of the target node on the tree.</p>
*
* <p>The only guarantee made is that for the same command tree and the same version of this library, the result of
* this method will <b>always</b> be a valid input for {@link #findNode(Collection)}, which should return the same node
* as provided to this method.</p>
*
* @param target the target node you are finding a path for
* @return a path to the resulting node, or an empty list if it was not found
*/
public Collection<String> getPath(final CommandNode<S> target) {
final List<List<CommandNode<S>>> nodes = new ArrayList<>();
addPaths(root, nodes, new ArrayList<>());
for (final List<CommandNode<S>> list : nodes) {
if (list.get(list.size() - 1) == target) {
final List<String> result = new ArrayList<>(list.size());
for (final CommandNode<S> node : list) {
if (node != root) {
result.add(node.getName());
}
}
return result;
}
}
return Collections.emptyList();
}
/**
* Finds a node by its path
*
* <p>Paths may be generated with {@link #getPath(CommandNode)}, and are guaranteed (for the same tree, and the
* same version of this library) to always produce the same valid node by this method.</p>
*
* <p>If a node could not be found at the specified path, then {@code null} will be returned.</p>
*
* @param path a generated path to a node
* @return the node at the given path, or null if not found
*/
public CommandNode<S> findNode(final Collection<String> path) {
CommandNode<S> node = root;
for (final String name : path) {
node = node.getChild(name);
if (node == null) {
return null;
}
}
return node;
}
/**
* Scans the command tree for potential ambiguous commands.
*
* <p>This is a shortcut for {@link CommandNode#findAmbiguities(AmbiguityConsumer)} on {@link #getRoot()}.</p>
*
* <p>Ambiguities are detected by testing every {@link CommandNode#getExamples()} on one node verses every sibling
* node. This is not fool proof, and relies a lot on the providers of the used argument types to give good examples.</p>
*
* @param consumer a callback to be notified of potential ambiguities
*/
public void findAmbiguities(final AmbiguityConsumer<S> consumer) {
root.findAmbiguities(consumer);
}
private void addPaths(final CommandNode<S> node, final List<List<CommandNode<S>>> result, final List<CommandNode<S>> parents) {
final List<CommandNode<S>> current = new ArrayList<>(parents);
current.add(node);
result.add(current);
for (final CommandNode<S> child : node.getChildren()) {
addPaths(child, result, current);
}
}
}

View file

@ -1,9 +1,9 @@
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT license.
package com.mojang.brigadier.tree;
// CHECKSTYLE:OFF
import com.google.common.collect.ComparisonChain;
import com.google.common.collect.Maps;
import com.google.common.collect.Sets;
import com.mojang.brigadier.AmbiguityConsumer;
import com.mojang.brigadier.Command;
import com.mojang.brigadier.RedirectModifier;
@ -17,19 +17,19 @@ import com.mojang.brigadier.suggestion.SuggestionsBuilder;
import java.util.Collection;
import java.util.Collections;
import java.util.HashSet;
import java.util.LinkedHashMap;
import java.util.Map;
import java.util.Set;
import java.util.concurrent.CompletableFuture;
import java.util.function.Predicate;
import java.util.stream.Collectors;
import net.minecraft.commands.CommandListenerWrapper; // CraftBukkit
public abstract class CommandNode<S> implements Comparable<CommandNode<S>> {
private Map<String, CommandNode<S>> children = Maps.newLinkedHashMap();
private Map<String, LiteralCommandNode<S>> literals = Maps.newLinkedHashMap();
private Map<String, ArgumentCommandNode<S, ?>> arguments = Maps.newLinkedHashMap();
private final Map<String, CommandNode<S>> children = new LinkedHashMap<>();
private final Map<String, LiteralCommandNode<S>> literals = new LinkedHashMap<>();
private final Map<String, ArgumentCommandNode<S, ?>> arguments = new LinkedHashMap<>();
private final Predicate<S> requirement;
private final CommandNode<S> redirect;
private final RedirectModifier<S> modifier;
@ -107,12 +107,10 @@ public abstract class CommandNode<S> implements Comparable<CommandNode<S>> {
arguments.put(node.getName(), (ArgumentCommandNode<S, ?>) node);
}
}
children = children.entrySet().stream().sorted(Map.Entry.comparingByValue()).collect(Collectors.toMap(Map.Entry::getKey, Map.Entry::getValue, (e1, e2) -> e1, LinkedHashMap::new));
}
public void findAmbiguities(final AmbiguityConsumer<S> consumer) {
Set<String> matches = Sets.newHashSet();
Set<String> matches = new HashSet<>();
for (final CommandNode<S> child : children.values()) {
for (final CommandNode<S> sibling : children.values()) {
@ -128,7 +126,7 @@ public abstract class CommandNode<S> implements Comparable<CommandNode<S>> {
if (matches.size() > 0) {
consumer.ambiguous(this, child, sibling, matches);
matches = Sets.newHashSet();
matches = new HashSet<>();
}
}
@ -193,11 +191,11 @@ public abstract class CommandNode<S> implements Comparable<CommandNode<S>> {
@Override
public int compareTo(final CommandNode<S> o) {
return ComparisonChain
.start()
.compareTrueFirst(this instanceof LiteralCommandNode, o instanceof LiteralCommandNode)
.compare(getSortedKey(), o.getSortedKey())
.result();
if (this instanceof LiteralCommandNode == o instanceof LiteralCommandNode) {
return getSortedKey().compareTo(o.getSortedKey());
}
return (o instanceof LiteralCommandNode) ? 1 : -1;
}
public boolean isFork() {