/**
* Copyright (C) 2011-2012 Typesafe Inc. <http://typesafe.com>
*/
package com.crashnote.core.config.helper.impl;
import java.io.File;
import java.io.StringReader;
import java.net.MalformedURLException;
import java.net.URL;
import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
import java.util.Iterator;
import java.util.LinkedList;
import java.util.List;
import java.util.ListIterator;
import java.util.Map;
import java.util.Stack;
import com.crashnote.core.config.helper.ConfigException;
import com.crashnote.core.config.helper.ConfigIncludeContext;
import com.crashnote.core.config.helper.ConfigOrigin;
import com.crashnote.core.config.helper.ConfigParseOptions;
import com.crashnote.core.config.helper.ConfigSyntax;
import com.crashnote.core.config.helper.ConfigValueType;
final class Parser {
static AbstractConfigValue parse(final Iterator<Token> tokens,
final ConfigOrigin origin, final ConfigParseOptions options,
final ConfigIncludeContext includeContext) {
final ParseContext context = new ParseContext(options.getSyntax(), origin, tokens,
SimpleIncluder.makeFull(options.getIncluder()), includeContext);
return context.parse();
}
static private final class TokenWithComments {
final Token token;
final List<Token> comments;
TokenWithComments(final Token token, final List<Token> comments) {
this.token = token;
this.comments = comments;
}
TokenWithComments(final Token token) {
this(token, Collections.<Token> emptyList());
}
TokenWithComments prepend(final List<Token> earlier) {
if (this.comments.isEmpty()) {
return new TokenWithComments(token, earlier);
} else {
final List<Token> merged = new ArrayList<Token>();
merged.addAll(earlier);
merged.addAll(comments);
return new TokenWithComments(token, merged);
}
}
SimpleConfigOrigin setComments(final SimpleConfigOrigin origin) {
if (comments.isEmpty()) {
return origin;
} else {
final List<String> newComments = new ArrayList<String>();
for (final Token c : comments) {
newComments.add(Tokens.getCommentText(c));
}
return origin.setComments(newComments);
}
}
@Override
public String toString() {
// this ends up in user-visible error messages, so we don't want the
// comments
return token.toString();
}
}
static private final class ParseContext {
private int lineNumber;
final private Stack<TokenWithComments> buffer;
final private Iterator<Token> tokens;
final private FullIncluder includer;
final private ConfigIncludeContext includeContext;
final private ConfigSyntax flavor;
final private ConfigOrigin baseOrigin;
final private LinkedList<Path> pathStack;
// this is the number of "equals" we are inside,
// used to modify the error message to reflect that
// someone may think this is .properties format.
int equalsCount;
ParseContext(final ConfigSyntax flavor, final ConfigOrigin origin, final Iterator<Token> tokens,
final FullIncluder includer, final ConfigIncludeContext includeContext) {
lineNumber = 1;
buffer = new Stack<TokenWithComments>();
this.tokens = tokens;
this.flavor = flavor;
this.baseOrigin = origin;
this.includer = includer;
this.includeContext = includeContext;
this.pathStack = new LinkedList<Path>();
this.equalsCount = 0;
}
private void consolidateCommentBlock(final Token commentToken) {
// a comment block "goes with" the following token
// unless it's separated from it by a blank line.
// we want to build a list of newline tokens followed
// by a non-newline non-comment token; with all comments
// associated with that final non-newline non-comment token.
final List<Token> newlines = new ArrayList<Token>();
final List<Token> comments = new ArrayList<Token>();
Token previous = null;
Token next = commentToken;
while (true) {
if (Tokens.isNewline(next)) {
if (previous != null && Tokens.isNewline(previous)) {
// blank line; drop all comments to this point and
// start a new comment block
comments.clear();
}
newlines.add(next);
} else if (Tokens.isComment(next)) {
comments.add(next);
} else {
// a non-newline non-comment token
break;
}
previous = next;
next = tokens.next();
}
// put our concluding token in the queue with all the comments
// attached
buffer.push(new TokenWithComments(next, comments));
// now put all the newlines back in front of it
final ListIterator<Token> li = newlines.listIterator(newlines.size());
while (li.hasPrevious()) {
buffer.push(new TokenWithComments(li.previous()));
}
}
private TokenWithComments popToken() {
if (buffer.isEmpty()) {
final Token t = tokens.next();
if (Tokens.isComment(t)) {
consolidateCommentBlock(t);
return buffer.pop();
} else {
return new TokenWithComments(t);
}
} else {
return buffer.pop();
}
}
private TokenWithComments nextToken() {
TokenWithComments withComments = null;
withComments = popToken();
final Token t = withComments.token;
if (Tokens.isProblem(t)) {
final ConfigOrigin origin = t.origin();
String message = Tokens.getProblemMessage(t);
final Throwable cause = Tokens.getProblemCause(t);
final boolean suggestQuotes = Tokens.getProblemSuggestQuotes(t);
if (suggestQuotes) {
message = addQuoteSuggestion(t.toString(), message);
} else {
message = addKeyName(message);
}
throw new ConfigException.Parse(origin, message, cause);
} else {
if (flavor == ConfigSyntax.JSON) {
if (Tokens.isUnquotedText(t)) {
throw parseError(addKeyName("Token not allowed in valid JSON: '"
+ Tokens.getUnquotedText(t) + "'"));
} else if (Tokens.isSubstitution(t)) {
throw parseError(addKeyName("Substitutions (${} syntax) not allowed in JSON"));
}
}
return withComments;
}
}
private void putBack(final TokenWithComments token) {
buffer.push(token);
}
private TokenWithComments nextTokenIgnoringNewline() {
TokenWithComments t = nextToken();
while (Tokens.isNewline(t.token)) {
// line number tokens have the line that was _ended_ by the
// newline, so we have to add one.
lineNumber = t.token.lineNumber() + 1;
t = nextToken();
}
return t;
}
// In arrays and objects, comma can be omitted
// as long as there's at least one newline instead.
// this skips any newlines in front of a comma,
// skips the comma, and returns true if it found
// either a newline or a comma. The iterator
// is left just after the comma or the newline.
private boolean checkElementSeparator() {
if (flavor == ConfigSyntax.JSON) {
final TokenWithComments t = nextTokenIgnoringNewline();
if (t.token == Tokens.COMMA) {
return true;
} else {
putBack(t);
return false;
}
} else {
boolean sawSeparatorOrNewline = false;
TokenWithComments t = nextToken();
while (true) {
if (Tokens.isNewline(t.token)) {
// newline number is the line just ended, so add one
lineNumber = t.token.lineNumber() + 1;
sawSeparatorOrNewline = true;
// we want to continue to also eat
// a comma if there is one.
} else if (t.token == Tokens.COMMA) {
return true;
} else {
// non-newline-or-comma
putBack(t);
return sawSeparatorOrNewline;
}
t = nextToken();
}
}
}
private static SubstitutionExpression tokenToSubstitutionExpression(final Token valueToken) {
final List<Token> expression = Tokens.getSubstitutionPathExpression(valueToken);
final Path path = parsePathExpression(expression.iterator(), valueToken.origin());
final boolean optional = Tokens.getSubstitutionOptional(valueToken);
return new SubstitutionExpression(path, optional);
}
// merge a bunch of adjacent values into one
// value; change unquoted text into a string
// value.
private void consolidateValueTokens() {
// this trick is not done in JSON
if (flavor == ConfigSyntax.JSON)
return;
// create only if we have value tokens
List<AbstractConfigValue> values = null;
TokenWithComments firstValueWithComments = null;
// ignore a newline up front
TokenWithComments t = nextTokenIgnoringNewline();
while (true) {
AbstractConfigValue v = null;
if (Tokens.isValue(t.token)) {
// if we consolidateValueTokens() multiple times then
// this value could be a concatenation, object, array,
// or substitution already.
v = Tokens.getValue(t.token);
} else if (Tokens.isUnquotedText(t.token)) {
v = new ConfigString(t.token.origin(), Tokens.getUnquotedText(t.token));
} else if (Tokens.isSubstitution(t.token)) {
v = new ConfigReference(t.token.origin(),
tokenToSubstitutionExpression(t.token));
} else if (t.token == Tokens.OPEN_CURLY || t.token == Tokens.OPEN_SQUARE) {
// there may be newlines _within_ the objects and arrays
v = parseValue(t);
} else {
break;
}
if (v == null)
throw new ConfigException.BugOrBroken("no value");
if (values == null) {
values = new ArrayList<AbstractConfigValue>();
firstValueWithComments = t;
}
values.add(v);
t = nextToken(); // but don't consolidate across a newline
}
// the last one wasn't a value token
putBack(t);
if (values == null)
return;
final AbstractConfigValue consolidated = ConfigConcatenation.concatenate(values);
putBack(new TokenWithComments(Tokens.newValue(consolidated),
firstValueWithComments.comments));
}
private ConfigOrigin lineOrigin() {
return ((SimpleConfigOrigin) baseOrigin).setLineNumber(lineNumber);
}
private ConfigException parseError(final String message) {
return parseError(message, null);
}
private ConfigException parseError(final String message, final Throwable cause) {
return new ConfigException.Parse(lineOrigin(), message, cause);
}
private String previousFieldName(final Path lastPath) {
if (lastPath != null) {
return lastPath.render();
} else if (pathStack.isEmpty())
return null;
else
return pathStack.peek().render();
}
private Path fullCurrentPath() {
Path full = null;
// pathStack has top of stack at front
for (final Path p : pathStack) {
if (full == null)
full = p;
else
full = full.prepend(p);
}
return full;
}
private String previousFieldName() {
return previousFieldName(null);
}
private String addKeyName(final String message) {
final String previousFieldName = previousFieldName();
if (previousFieldName != null) {
return "in value for key '" + previousFieldName + "': " + message;
} else {
return message;
}
}
private String addQuoteSuggestion(final String badToken, final String message) {
return addQuoteSuggestion(null, equalsCount > 0, badToken, message);
}
private String addQuoteSuggestion(final Path lastPath, final boolean insideEquals, final String badToken,
final String message) {
final String previousFieldName = previousFieldName(lastPath);
final String part;
if (badToken.equals(Tokens.END.toString())) {
// EOF requires special handling for the error to make sense.
if (previousFieldName != null)
part = message + " (if you intended '" + previousFieldName
+ "' to be part of a value, instead of a key, "
+ "try adding double quotes around the whole value";
else
return message;
} else {
if (previousFieldName != null) {
part = message + " (if you intended " + badToken
+ " to be part of the value for '" + previousFieldName + "', "
+ "try enclosing the value in double quotes";
} else {
part = message + " (if you intended " + badToken
+ " to be part of a key or string value, "
+ "try enclosing the key or value in double quotes";
}
}
if (insideEquals)
return part
+ ", or you may be able to rename the file .properties rather than .conf)";
else
return part + ")";
}
private AbstractConfigValue parseValue(final TokenWithComments t) {
AbstractConfigValue v;
if (Tokens.isValue(t.token)) {
v = Tokens.getValue(t.token);
} else if (t.token == Tokens.OPEN_CURLY) {
v = parseObject(true);
} else if (t.token == Tokens.OPEN_SQUARE) {
v = parseArray();
} else {
throw parseError(addQuoteSuggestion(t.token.toString(),
"Expecting a value but got wrong token: " + t.token));
}
v = v.withOrigin(t.setComments(v.origin()));
return v;
}
private static AbstractConfigObject createValueUnderPath(final Path path,
final AbstractConfigValue value) {
// for path foo.bar, we are creating
// { "foo" : { "bar" : value } }
final List<String> keys = new ArrayList<String>();
String key = path.first();
Path remaining = path.remainder();
while (key != null) {
keys.add(key);
if (remaining == null) {
break;
} else {
key = remaining.first();
remaining = remaining.remainder();
}
}
// the setComments(null) is to ensure comments are only
// on the exact leaf node they apply to.
// a comment before "foo.bar" applies to the full setting
// "foo.bar" not also to "foo"
final ListIterator<String> i = keys.listIterator(keys.size());
final String deepest = i.previous();
AbstractConfigObject o = new SimpleConfigObject(value.origin().setComments(null),
Collections.<String, AbstractConfigValue> singletonMap(
deepest, value));
while (i.hasPrevious()) {
final Map<String, AbstractConfigValue> m = Collections.<String, AbstractConfigValue> singletonMap(
i.previous(), o);
o = new SimpleConfigObject(value.origin().setComments(null), m);
}
return o;
}
private Path parseKey(final TokenWithComments token) {
if (flavor == ConfigSyntax.JSON) {
if (Tokens.isValueWithType(token.token, ConfigValueType.STRING)) {
final String key = (String) Tokens.getValue(token.token).unwrapped();
return Path.newKey(key);
} else {
throw parseError(addKeyName("Expecting close brace } or a field name here, got "
+ token));
}
} else {
final List<Token> expression = new ArrayList<Token>();
TokenWithComments t = token;
while (Tokens.isValue(t.token) || Tokens.isUnquotedText(t.token)) {
expression.add(t.token);
t = nextToken(); // note: don't cross a newline
}
if (expression.isEmpty()) {
throw parseError(addKeyName("expecting a close brace or a field name here, got "
+ t));
}
putBack(t); // put back the token we ended with
return parsePathExpression(expression.iterator(), lineOrigin());
}
}
private static boolean isIncludeKeyword(final Token t) {
return Tokens.isUnquotedText(t)
&& Tokens.getUnquotedText(t).equals("include");
}
private static boolean isUnquotedWhitespace(final Token t) {
if (!Tokens.isUnquotedText(t))
return false;
final String s = Tokens.getUnquotedText(t);
for (int i = 0; i < s.length(); ++i) {
final char c = s.charAt(i);
if (!ConfigImplUtil.isWhitespace(c))
return false;
}
return true;
}
private void parseInclude(final Map<String, AbstractConfigValue> values) {
TokenWithComments t = nextTokenIgnoringNewline();
while (isUnquotedWhitespace(t.token)) {
t = nextTokenIgnoringNewline();
}
AbstractConfigObject obj;
// we either have a quoted string or the "file()" syntax
if (Tokens.isUnquotedText(t.token)) {
// get foo(
final String kind = Tokens.getUnquotedText(t.token);
if (kind.equals("url(")) {
} else if (kind.equals("file(")) {
} else if (kind.equals("classpath(")) {
} else {
throw parseError("expecting include parameter to be quoted filename, file(), classpath(), or url(). No spaces are allowed before the open paren. Not expecting: "
+ t);
}
// skip space inside parens
t = nextTokenIgnoringNewline();
while (isUnquotedWhitespace(t.token)) {
t = nextTokenIgnoringNewline();
}
// quoted string
final String name;
if (Tokens.isValueWithType(t.token, ConfigValueType.STRING)) {
name = (String) Tokens.getValue(t.token).unwrapped();
} else {
throw parseError("expecting a quoted string inside file(), classpath(), or url(), rather than: "
+ t);
}
// skip space after string, inside parens
t = nextTokenIgnoringNewline();
while (isUnquotedWhitespace(t.token)) {
t = nextTokenIgnoringNewline();
}
if (Tokens.isUnquotedText(t.token) && Tokens.getUnquotedText(t.token).equals(")")) {
// OK, close paren
} else {
throw parseError("expecting a close parentheses ')' here, not: " + t);
}
if (kind.equals("url(")) {
final URL url;
try {
url = new URL(name);
} catch (MalformedURLException e) {
throw parseError("include url() specifies an invalid URL: " + name, e);
}
obj = (AbstractConfigObject) includer.includeURL(includeContext, url);
} else if (kind.equals("file(")) {
obj = (AbstractConfigObject) includer.includeFile(includeContext,
new File(name));
} else if (kind.equals("classpath(")) {
obj = (AbstractConfigObject) includer.includeResources(includeContext, name);
} else {
throw new ConfigException.BugOrBroken("should not be reached");
}
} else if (Tokens.isValueWithType(t.token, ConfigValueType.STRING)) {
final String name = (String) Tokens.getValue(t.token).unwrapped();
obj = (AbstractConfigObject) includer
.include(includeContext, name);
} else {
throw parseError("include keyword is not followed by a quoted string, but by: " + t);
}
if (!pathStack.isEmpty()) {
final Path prefix = new Path(pathStack);
obj = obj.relativized(prefix);
}
for (final String key : obj.keySet()) {
final AbstractConfigValue v = obj.get(key);
final AbstractConfigValue existing = values.get(key);
if (existing != null) {
values.put(key, v.withFallback(existing));
} else {
values.put(key, v);
}
}
}
private boolean isKeyValueSeparatorToken(final Token t) {
if (flavor == ConfigSyntax.JSON) {
return t == Tokens.COLON;
} else {
return t == Tokens.COLON || t == Tokens.EQUALS || t == Tokens.PLUS_EQUALS;
}
}
private AbstractConfigObject parseObject(final boolean hadOpenCurly) {
// invoked just after the OPEN_CURLY (or START, if !hadOpenCurly)
final Map<String, AbstractConfigValue> values = new HashMap<String, AbstractConfigValue>();
final ConfigOrigin objectOrigin = lineOrigin();
boolean afterComma = false;
Path lastPath = null;
boolean lastInsideEquals = false;
while (true) {
TokenWithComments t = nextTokenIgnoringNewline();
if (t.token == Tokens.CLOSE_CURLY) {
if (flavor == ConfigSyntax.JSON && afterComma) {
throw parseError(addQuoteSuggestion(t.toString(),
"expecting a field name after a comma, got a close brace } instead"));
} else if (!hadOpenCurly) {
throw parseError(addQuoteSuggestion(t.toString(),
"unbalanced close brace '}' with no open brace"));
}
break;
} else if (t.token == Tokens.END && !hadOpenCurly) {
putBack(t);
break;
} else if (flavor != ConfigSyntax.JSON && isIncludeKeyword(t.token)) {
parseInclude(values);
afterComma = false;
} else {
final TokenWithComments keyToken = t;
final Path path = parseKey(keyToken);
final TokenWithComments afterKey = nextTokenIgnoringNewline();
boolean insideEquals = false;
// path must be on-stack while we parse the value
pathStack.push(path);
final TokenWithComments valueToken;
AbstractConfigValue newValue;
if (flavor == ConfigSyntax.CONF && afterKey.token == Tokens.OPEN_CURLY) {
// can omit the ':' or '=' before an object value
valueToken = afterKey;
} else {
if (!isKeyValueSeparatorToken(afterKey.token)) {
throw parseError(addQuoteSuggestion(afterKey.toString(),
"Key '" + path.render() + "' may not be followed by token: "
+ afterKey));
}
if (afterKey.token == Tokens.EQUALS) {
insideEquals = true;
equalsCount += 1;
}
consolidateValueTokens();
valueToken = nextTokenIgnoringNewline();
}
newValue = parseValue(valueToken.prepend(keyToken.comments));
if (afterKey.token == Tokens.PLUS_EQUALS) {
final List<AbstractConfigValue> concat = new ArrayList<AbstractConfigValue>(2);
final AbstractConfigValue previousRef = new ConfigReference(newValue.origin(),
new SubstitutionExpression(fullCurrentPath(), true /* optional */));
final AbstractConfigValue list = new SimpleConfigList(newValue.origin(),
Collections.singletonList(newValue));
concat.add(previousRef);
concat.add(list);
newValue = ConfigConcatenation.concatenate(concat);
}
lastPath = pathStack.pop();
if (insideEquals) {
equalsCount -= 1;
}
lastInsideEquals = insideEquals;
final String key = path.first();
final Path remaining = path.remainder();
if (remaining == null) {
final AbstractConfigValue existing = values.get(key);
if (existing != null) {
// In strict JSON, dups should be an error; while in
// our custom config language, they should be merged
// if the value is an object (or substitution that
// could become an object).
if (flavor == ConfigSyntax.JSON) {
throw parseError("JSON does not allow duplicate fields: '"
+ key
+ "' was already seen at "
+ existing.origin().description());
} else {
newValue = newValue.withFallback(existing);
}
}
values.put(key, newValue);
} else {
if (flavor == ConfigSyntax.JSON) {
throw new ConfigException.BugOrBroken(
"somehow got multi-element path in JSON mode");
}
AbstractConfigObject obj = createValueUnderPath(
remaining, newValue);
final AbstractConfigValue existing = values.get(key);
if (existing != null) {
obj = obj.withFallback(existing);
}
values.put(key, obj);
}
afterComma = false;
}
if (checkElementSeparator()) {
// continue looping
afterComma = true;
} else {
t = nextTokenIgnoringNewline();
if (t.token == Tokens.CLOSE_CURLY) {
if (!hadOpenCurly) {
throw parseError(addQuoteSuggestion(lastPath, lastInsideEquals,
t.toString(), "unbalanced close brace '}' with no open brace"));
}
break;
} else if (hadOpenCurly) {
throw parseError(addQuoteSuggestion(lastPath, lastInsideEquals,
t.toString(), "Expecting close brace } or a comma, got " + t));
} else {
if (t.token == Tokens.END) {
putBack(t);
break;
} else {
throw parseError(addQuoteSuggestion(lastPath, lastInsideEquals,
t.toString(), "Expecting end of input or a comma, got " + t));
}
}
}
}
return new SimpleConfigObject(objectOrigin, values);
}
private SimpleConfigList parseArray() {
// invoked just after the OPEN_SQUARE
final ConfigOrigin arrayOrigin = lineOrigin();
final List<AbstractConfigValue> values = new ArrayList<AbstractConfigValue>();
consolidateValueTokens();
TokenWithComments t = nextTokenIgnoringNewline();
// special-case the first element
if (t.token == Tokens.CLOSE_SQUARE) {
return new SimpleConfigList(arrayOrigin,
Collections.<AbstractConfigValue> emptyList());
} else if (Tokens.isValue(t.token) || t.token == Tokens.OPEN_CURLY
|| t.token == Tokens.OPEN_SQUARE) {
values.add(parseValue(t));
} else {
throw parseError(addKeyName("List should have ] or a first element after the open [, instead had token: "
+ t
+ " (if you want "
+ t
+ " to be part of a string value, then double-quote it)"));
}
// now remaining elements
while (true) {
// just after a value
if (checkElementSeparator()) {
// comma (or newline equivalent) consumed
} else {
t = nextTokenIgnoringNewline();
if (t.token == Tokens.CLOSE_SQUARE) {
return new SimpleConfigList(arrayOrigin, values);
} else {
throw parseError(addKeyName("List should have ended with ] or had a comma, instead had token: "
+ t
+ " (if you want "
+ t
+ " to be part of a string value, then double-quote it)"));
}
}
// now just after a comma
consolidateValueTokens();
t = nextTokenIgnoringNewline();
if (Tokens.isValue(t.token) || t.token == Tokens.OPEN_CURLY
|| t.token == Tokens.OPEN_SQUARE) {
values.add(parseValue(t));
} else if (flavor != ConfigSyntax.JSON && t.token == Tokens.CLOSE_SQUARE) {
// we allow one trailing comma
putBack(t);
} else {
throw parseError(addKeyName("List should have had new element after a comma, instead had token: "
+ t
+ " (if you want the comma or "
+ t
+ " to be part of a string value, then double-quote it)"));
}
}
}
AbstractConfigValue parse() {
TokenWithComments t = nextTokenIgnoringNewline();
if (t.token == Tokens.START) {
// OK
} else {
throw new ConfigException.BugOrBroken(
"token stream did not begin with START, had " + t);
}
t = nextTokenIgnoringNewline();
AbstractConfigValue result = null;
if (t.token == Tokens.OPEN_CURLY || t.token == Tokens.OPEN_SQUARE) {
result = parseValue(t);
} else {
if (flavor == ConfigSyntax.JSON) {
if (t.token == Tokens.END) {
throw parseError("Empty document");
} else {
throw parseError("Document must have an object or array at root, unexpected token: "
+ t);
}
} else {
// the root object can omit the surrounding braces.
// this token should be the first field's key, or part
// of it, so put it back.
putBack(t);
result = parseObject(false);
// in this case we don't try to use commentsStack comments
// since they would all presumably apply to fields not the
// root object
}
}
t = nextTokenIgnoringNewline();
if (t.token == Tokens.END) {
return result;
} else {
throw parseError("Document has trailing tokens after first object or array: "
+ t);
}
}
}
static class Element {
StringBuilder sb;
// an element can be empty if it has a quoted empty string "" in it
boolean canBeEmpty;
Element(final String initial, final boolean canBeEmpty) {
this.canBeEmpty = canBeEmpty;
this.sb = new StringBuilder(initial);
}
@Override
public String toString() {
return "Element(" + sb.toString() + "," + canBeEmpty + ")";
}
}
private static void addPathText(final List<Element> buf, final boolean wasQuoted,
final String newText) {
final int i = wasQuoted ? -1 : newText.indexOf('.');
final Element current = buf.get(buf.size() - 1);
if (i < 0) {
// add to current path element
current.sb.append(newText);
// any empty quoted string means this element can
// now be empty.
if (wasQuoted && current.sb.length() == 0)
current.canBeEmpty = true;
} else {
// "buf" plus up to the period is an element
current.sb.append(newText.substring(0, i));
// then start a new element
buf.add(new Element("", false));
// recurse to consume remainder of newText
addPathText(buf, false, newText.substring(i + 1));
}
}
private static Path parsePathExpression(final Iterator<Token> expression,
final ConfigOrigin origin) {
return parsePathExpression(expression, origin, null);
}
// originalText may be null if not available
private static Path parsePathExpression(final Iterator<Token> expression,
final ConfigOrigin origin, final String originalText) {
// each builder in "buf" is an element in the path.
final List<Element> buf = new ArrayList<Element>();
buf.add(new Element("", false));
if (!expression.hasNext()) {
throw new ConfigException.BadPath(origin, originalText,
"Expecting a field name or path here, but got nothing");
}
while (expression.hasNext()) {
final Token t = expression.next();
if (Tokens.isValueWithType(t, ConfigValueType.STRING)) {
final AbstractConfigValue v = Tokens.getValue(t);
// this is a quoted string; so any periods
// in here don't count as path separators
final String s = v.transformToString();
addPathText(buf, true, s);
} else if (t == Tokens.END) {
// ignore this; when parsing a file, it should not happen
// since we're parsing a token list rather than the main
// token iterator, and when parsing a path expression from the
// API, it's expected to have an END.
} else {
// any periods outside of a quoted string count as
// separators
final String text;
if (Tokens.isValue(t)) {
// appending a number here may add
// a period, but we _do_ count those as path
// separators, because we basically want
// "foo 3.0bar" to parse as a string even
// though there's a number in it. The fact that
// we tokenize non-string values is largely an
// implementation detail.
final AbstractConfigValue v = Tokens.getValue(t);
text = v.transformToString();
} else if (Tokens.isUnquotedText(t)) {
text = Tokens.getUnquotedText(t);
} else {
throw new ConfigException.BadPath(
origin,
originalText,
"Token not allowed in path expression: "
+ t
+ " (you can double-quote this token if you really want it here)");
}
addPathText(buf, false, text);
}
}
final PathBuilder pb = new PathBuilder();
for (final Element e : buf) {
if (e.sb.length() == 0 && !e.canBeEmpty) {
throw new ConfigException.BadPath(
origin,
originalText,
"path has a leading, trailing, or two adjacent period '.' (use quoted \"\" empty string if you want an empty element)");
} else {
pb.appendKey(e.sb.toString());
}
}
return pb.result();
}
static ConfigOrigin apiOrigin = SimpleConfigOrigin.newSimple("path parameter");
static Path parsePath(final String path) {
final Path speculated = speculativeFastParsePath(path);
if (speculated != null)
return speculated;
final StringReader reader = new StringReader(path);
try {
final Iterator<Token> tokens = Tokenizer.tokenize(apiOrigin, reader,
ConfigSyntax.CONF);
tokens.next(); // drop START
return parsePathExpression(tokens, apiOrigin, path);
} finally {
reader.close();
}
}
// the idea is to see if the string has any chars that might require the
// full parser to deal with.
private static boolean hasUnsafeChars(final String s) {
for (int i = 0; i < s.length(); ++i) {
final char c = s.charAt(i);
if (Character.isLetter(c) || c == '.')
continue;
else
return true;
}
return false;
}
private static void appendPathString(final PathBuilder pb, final String s) {
final int splitAt = s.indexOf('.');
if (splitAt < 0) {
pb.appendKey(s);
} else {
pb.appendKey(s.substring(0, splitAt));
appendPathString(pb, s.substring(splitAt + 1));
}
}
// do something much faster than the full parser if
// we just have something like "foo" or "foo.bar"
private static Path speculativeFastParsePath(final String path) {
final String s = ConfigImplUtil.unicodeTrim(path);
if (s.isEmpty())
return null;
if (hasUnsafeChars(s))
return null;
if (s.startsWith(".") || s.endsWith(".") || s.contains(".."))
return null; // let the full parser throw the error
final PathBuilder pb = new PathBuilder();
appendPathString(pb, s);
return pb.result();
}
}