/**
* 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 org.apache.hadoop.util;
import com.google.common.base.Preconditions;
import java.io.PrintWriter;
import java.io.StringWriter;
import java.net.URI;
import java.net.URISyntaxException;
import java.text.DateFormat;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
import java.util.Date;
import java.util.HashSet;
import java.util.Iterator;
import java.util.LinkedHashSet;
import java.util.List;
import java.util.Locale;
import java.util.Map;
import java.util.Set;
import java.util.StringTokenizer;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import org.apache.commons.lang.SystemUtils;
import org.apache.hadoop.classification.InterfaceAudience;
import org.apache.hadoop.classification.InterfaceStability;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.net.NetUtils;
import com.google.common.net.InetAddresses;
/**
* General string utils
*/
@InterfaceAudience.Private
@InterfaceStability.Unstable
public class StringUtils {
/**
* Priority of the StringUtils shutdown hook.
*/
public static final int SHUTDOWN_HOOK_PRIORITY = 0;
/**
* Shell environment variables: $ followed by one letter or _ followed by
* multiple letters, numbers, or underscores. The group captures the
* environment variable name without the leading $.
*/
public static final Pattern SHELL_ENV_VAR_PATTERN =
Pattern.compile("\\$([A-Za-z_]{1}[A-Za-z0-9_]*)");
/**
* Windows environment variables: surrounded by %. The group captures the
* environment variable name without the leading and trailing %.
*/
public static final Pattern WIN_ENV_VAR_PATTERN = Pattern.compile("%(.*?)%");
/**
* Regular expression that matches and captures environment variable names
* according to platform-specific rules.
*/
public static final Pattern ENV_VAR_PATTERN = Shell.WINDOWS ?
WIN_ENV_VAR_PATTERN : SHELL_ENV_VAR_PATTERN;
/**
* Make a string representation of the exception.
* @param e The exception to stringify
* @return A string with exception name and call stack.
*/
public static String stringifyException(Throwable e) {
StringWriter stm = new StringWriter();
PrintWriter wrt = new PrintWriter(stm);
e.printStackTrace(wrt);
wrt.close();
return stm.toString();
}
/**
* Given a full hostname, return the word upto the first dot.
* @param fullHostname the full hostname
* @return the hostname to the first dot
*/
public static String simpleHostname(String fullHostname) {
if (InetAddresses.isInetAddress(fullHostname)) {
return fullHostname;
}
int offset = fullHostname.indexOf('.');
if (offset != -1) {
return fullHostname.substring(0, offset);
}
return fullHostname;
}
/**
* Given an integer, return a string that is in an approximate, but human
* readable format.
* @param number the number to format
* @return a human readable form of the integer
*
* @deprecated use {@link TraditionalBinaryPrefix#long2String(long, String, int)}.
*/
@Deprecated
public static String humanReadableInt(long number) {
return TraditionalBinaryPrefix.long2String(number, "", 1);
}
/** The same as String.format(Locale.ENGLISH, format, objects). */
public static String format(final String format, final Object... objects) {
return String.format(Locale.ENGLISH, format, objects);
}
/**
* Format a percentage for presentation to the user.
* @param fraction the percentage as a fraction, e.g. 0.1 = 10%
* @param decimalPlaces the number of decimal places
* @return a string representation of the percentage
*/
public static String formatPercent(double fraction, int decimalPlaces) {
return format("%." + decimalPlaces + "f%%", fraction*100);
}
/**
* Given an array of strings, return a comma-separated list of its elements.
* @param strs Array of strings
* @return Empty string if strs.length is 0, comma separated list of strings
* otherwise
*/
public static String arrayToString(String[] strs) {
if (strs.length == 0) { return ""; }
StringBuilder sbuf = new StringBuilder();
sbuf.append(strs[0]);
for (int idx = 1; idx < strs.length; idx++) {
sbuf.append(",");
sbuf.append(strs[idx]);
}
return sbuf.toString();
}
/**
* Given an array of bytes it will convert the bytes to a hex string
* representation of the bytes
* @param bytes
* @param start start index, inclusively
* @param end end index, exclusively
* @return hex string representation of the byte array
*/
public static String byteToHexString(byte[] bytes, int start, int end) {
if (bytes == null) {
throw new IllegalArgumentException("bytes == null");
}
StringBuilder s = new StringBuilder();
for(int i = start; i < end; i++) {
s.append(format("%02x", bytes[i]));
}
return s.toString();
}
/** Same as byteToHexString(bytes, 0, bytes.length). */
public static String byteToHexString(byte bytes[]) {
return byteToHexString(bytes, 0, bytes.length);
}
/**
* Given a hexstring this will return the byte array corresponding to the
* string
* @param hex the hex String array
* @return a byte array that is a hex string representation of the given
* string. The size of the byte array is therefore hex.length/2
*/
public static byte[] hexStringToByte(String hex) {
byte[] bts = new byte[hex.length() / 2];
for (int i = 0; i < bts.length; i++) {
bts[i] = (byte) Integer.parseInt(hex.substring(2 * i, 2 * i + 2), 16);
}
return bts;
}
/**
*
* @param uris
*/
public static String uriToString(URI[] uris){
if (uris == null) {
return null;
}
StringBuilder ret = new StringBuilder(uris[0].toString());
for(int i = 1; i < uris.length;i++){
ret.append(",");
ret.append(uris[i].toString());
}
return ret.toString();
}
/**
* @param str
* The string array to be parsed into an URI array.
* @return <tt>null</tt> if str is <tt>null</tt>, else the URI array
* equivalent to str.
* @throws IllegalArgumentException
* If any string in str violates RFC 2396.
*/
public static URI[] stringToURI(String[] str){
if (str == null)
return null;
URI[] uris = new URI[str.length];
for (int i = 0; i < str.length;i++){
try{
uris[i] = new URI(str[i]);
}catch(URISyntaxException ur){
throw new IllegalArgumentException(
"Failed to create uri for " + str[i], ur);
}
}
return uris;
}
/**
*
* @param str
*/
public static Path[] stringToPath(String[] str){
if (str == null) {
return null;
}
Path[] p = new Path[str.length];
for (int i = 0; i < str.length;i++){
p[i] = new Path(str[i]);
}
return p;
}
/**
*
* Given a finish and start time in long milliseconds, returns a
* String in the format Xhrs, Ymins, Z sec, for the time difference between two times.
* If finish time comes before start time then negative valeus of X, Y and Z wil return.
*
* @param finishTime finish time
* @param startTime start time
*/
public static String formatTimeDiff(long finishTime, long startTime){
long timeDiff = finishTime - startTime;
return formatTime(timeDiff);
}
/**
*
* Given the time in long milliseconds, returns a
* String in the format Xhrs, Ymins, Z sec.
*
* @param timeDiff The time difference to format
*/
public static String formatTime(long timeDiff){
StringBuilder buf = new StringBuilder();
long hours = timeDiff / (60*60*1000);
long rem = (timeDiff % (60*60*1000));
long minutes = rem / (60*1000);
rem = rem % (60*1000);
long seconds = rem / 1000;
if (hours != 0){
buf.append(hours);
buf.append("hrs, ");
}
if (minutes != 0){
buf.append(minutes);
buf.append("mins, ");
}
// return "0sec if no difference
buf.append(seconds);
buf.append("sec");
return buf.toString();
}
/**
* Formats time in ms and appends difference (finishTime - startTime)
* as returned by formatTimeDiff().
* If finish time is 0, empty string is returned, if start time is 0
* then difference is not appended to return value.
* @param dateFormat date format to use
* @param finishTime fnish time
* @param startTime start time
* @return formatted value.
*/
public static String getFormattedTimeWithDiff(DateFormat dateFormat,
long finishTime, long startTime){
StringBuilder buf = new StringBuilder();
if (0 != finishTime) {
buf.append(dateFormat.format(new Date(finishTime)));
if (0 != startTime){
buf.append(" (" + formatTimeDiff(finishTime , startTime) + ")");
}
}
return buf.toString();
}
/**
* Returns an arraylist of strings.
* @param str the comma seperated string values
* @return the arraylist of the comma seperated string values
*/
public static String[] getStrings(String str){
Collection<String> values = getStringCollection(str);
if(values.size() == 0) {
return null;
}
return values.toArray(new String[values.size()]);
}
/**
* Returns a collection of strings.
* @param str comma seperated string values
* @return an <code>ArrayList</code> of string values
*/
public static Collection<String> getStringCollection(String str){
String delim = ",";
return getStringCollection(str, delim);
}
/**
* Returns a collection of strings.
*
* @param str
* String to parse
* @param delim
* delimiter to separate the values
* @return Collection of parsed elements.
*/
public static Collection<String> getStringCollection(String str, String delim) {
List<String> values = new ArrayList<String>();
if (str == null)
return values;
StringTokenizer tokenizer = new StringTokenizer(str, delim);
while (tokenizer.hasMoreTokens()) {
values.add(tokenizer.nextToken());
}
return values;
}
/**
* Splits a comma separated value <code>String</code>, trimming leading and trailing whitespace on each value.
* Duplicate and empty values are removed.
* @param str a comma separated <String> with values
* @return a <code>Collection</code> of <code>String</code> values
*/
public static Collection<String> getTrimmedStringCollection(String str){
Set<String> set = new LinkedHashSet<String>(
Arrays.asList(getTrimmedStrings(str)));
set.remove("");
return set;
}
/**
* Splits a comma separated value <code>String</code>, trimming leading and trailing whitespace on each value.
* @param str a comma separated <String> with values
* @return an array of <code>String</code> values
*/
public static String[] getTrimmedStrings(String str){
if (null == str || str.trim().isEmpty()) {
return emptyStringArray;
}
return str.trim().split("\\s*,\\s*");
}
/**
* Trims all the strings in a Collection<String> and returns a Set<String>.
* @param strings
* @return
*/
public static Set<String> getTrimmedStrings(Collection<String> strings) {
Set<String> trimmedStrings = new HashSet<String>();
for (String string: strings) {
trimmedStrings.add(string.trim());
}
return trimmedStrings;
}
final public static String[] emptyStringArray = {};
final public static char COMMA = ',';
final public static String COMMA_STR = ",";
final public static char ESCAPE_CHAR = '\\';
/**
* Split a string using the default separator
* @param str a string that may have escaped separator
* @return an array of strings
*/
public static String[] split(String str) {
return split(str, ESCAPE_CHAR, COMMA);
}
/**
* Split a string using the given separator
* @param str a string that may have escaped separator
* @param escapeChar a char that be used to escape the separator
* @param separator a separator char
* @return an array of strings
*/
public static String[] split(
String str, char escapeChar, char separator) {
if (str==null) {
return null;
}
ArrayList<String> strList = new ArrayList<String>();
StringBuilder split = new StringBuilder();
int index = 0;
while ((index = findNext(str, separator, escapeChar, index, split)) >= 0) {
++index; // move over the separator for next search
strList.add(split.toString());
split.setLength(0); // reset the buffer
}
strList.add(split.toString());
// remove trailing empty split(s)
int last = strList.size(); // last split
while (--last>=0 && "".equals(strList.get(last))) {
strList.remove(last);
}
return strList.toArray(new String[strList.size()]);
}
/**
* Split a string using the given separator, with no escaping performed.
* @param str a string to be split. Note that this may not be null.
* @param separator a separator char
* @return an array of strings
*/
public static String[] split(
String str, char separator) {
// String.split returns a single empty result for splitting the empty
// string.
if (str.isEmpty()) {
return new String[]{""};
}
ArrayList<String> strList = new ArrayList<String>();
int startIndex = 0;
int nextIndex = 0;
while ((nextIndex = str.indexOf(separator, startIndex)) != -1) {
strList.add(str.substring(startIndex, nextIndex));
startIndex = nextIndex + 1;
}
strList.add(str.substring(startIndex));
// remove trailing empty split(s)
int last = strList.size(); // last split
while (--last>=0 && "".equals(strList.get(last))) {
strList.remove(last);
}
return strList.toArray(new String[strList.size()]);
}
/**
* Finds the first occurrence of the separator character ignoring the escaped
* separators starting from the index. Note the substring between the index
* and the position of the separator is passed.
* @param str the source string
* @param separator the character to find
* @param escapeChar character used to escape
* @param start from where to search
* @param split used to pass back the extracted string
*/
public static int findNext(String str, char separator, char escapeChar,
int start, StringBuilder split) {
int numPreEscapes = 0;
for (int i = start; i < str.length(); i++) {
char curChar = str.charAt(i);
if (numPreEscapes == 0 && curChar == separator) { // separator
return i;
} else {
split.append(curChar);
numPreEscapes = (curChar == escapeChar)
? (++numPreEscapes) % 2
: 0;
}
}
return -1;
}
/**
* Escape commas in the string using the default escape char
* @param str a string
* @return an escaped string
*/
public static String escapeString(String str) {
return escapeString(str, ESCAPE_CHAR, COMMA);
}
/**
* Escape <code>charToEscape</code> in the string
* with the escape char <code>escapeChar</code>
*
* @param str string
* @param escapeChar escape char
* @param charToEscape the char to be escaped
* @return an escaped string
*/
public static String escapeString(
String str, char escapeChar, char charToEscape) {
return escapeString(str, escapeChar, new char[] {charToEscape});
}
// check if the character array has the character
private static boolean hasChar(char[] chars, char character) {
for (char target : chars) {
if (character == target) {
return true;
}
}
return false;
}
/**
* @param charsToEscape array of characters to be escaped
*/
public static String escapeString(String str, char escapeChar,
char[] charsToEscape) {
if (str == null) {
return null;
}
StringBuilder result = new StringBuilder();
for (int i=0; i<str.length(); i++) {
char curChar = str.charAt(i);
if (curChar == escapeChar || hasChar(charsToEscape, curChar)) {
// special char
result.append(escapeChar);
}
result.append(curChar);
}
return result.toString();
}
/**
* Unescape commas in the string using the default escape char
* @param str a string
* @return an unescaped string
*/
public static String unEscapeString(String str) {
return unEscapeString(str, ESCAPE_CHAR, COMMA);
}
/**
* Unescape <code>charToEscape</code> in the string
* with the escape char <code>escapeChar</code>
*
* @param str string
* @param escapeChar escape char
* @param charToEscape the escaped char
* @return an unescaped string
*/
public static String unEscapeString(
String str, char escapeChar, char charToEscape) {
return unEscapeString(str, escapeChar, new char[] {charToEscape});
}
/**
* @param charsToEscape array of characters to unescape
*/
public static String unEscapeString(String str, char escapeChar,
char[] charsToEscape) {
if (str == null) {
return null;
}
StringBuilder result = new StringBuilder(str.length());
boolean hasPreEscape = false;
for (int i=0; i<str.length(); i++) {
char curChar = str.charAt(i);
if (hasPreEscape) {
if (curChar != escapeChar && !hasChar(charsToEscape, curChar)) {
// no special char
throw new IllegalArgumentException("Illegal escaped string " + str +
" unescaped " + escapeChar + " at " + (i-1));
}
// otherwise discard the escape char
result.append(curChar);
hasPreEscape = false;
} else {
if (hasChar(charsToEscape, curChar)) {
throw new IllegalArgumentException("Illegal escaped string " + str +
" unescaped " + curChar + " at " + i);
} else if (curChar == escapeChar) {
hasPreEscape = true;
} else {
result.append(curChar);
}
}
}
if (hasPreEscape ) {
throw new IllegalArgumentException("Illegal escaped string " + str +
", not expecting " + escapeChar + " in the end." );
}
return result.toString();
}
/**
* Return a message for logging.
* @param prefix prefix keyword for the message
* @param msg content of the message
* @return a message for logging
*/
private static String toStartupShutdownString(String prefix, String [] msg) {
StringBuilder b = new StringBuilder(prefix);
b.append("\n/************************************************************");
for(String s : msg)
b.append("\n" + prefix + s);
b.append("\n************************************************************/");
return b.toString();
}
/**
* Print a log message for starting up and shutting down
* @param clazz the class of the server
* @param args arguments
* @param LOG the target log object
*/
public static void startupShutdownMessage(Class<?> clazz, String[] args,
final org.apache.commons.logging.Log LOG) {
startupShutdownMessage(clazz, args, LogAdapter.create(LOG));
}
/**
* Print a log message for starting up and shutting down
* @param clazz the class of the server
* @param args arguments
* @param LOG the target log object
*/
public static void startupShutdownMessage(Class<?> clazz, String[] args,
final org.slf4j.Logger LOG) {
startupShutdownMessage(clazz, args, LogAdapter.create(LOG));
}
static void startupShutdownMessage(Class<?> clazz, String[] args,
final LogAdapter LOG) {
final String hostname = NetUtils.getHostname();
final String classname = clazz.getSimpleName();
LOG.info(
toStartupShutdownString("STARTUP_MSG: ", new String[] {
"Starting " + classname,
" host = " + hostname,
" args = " + Arrays.asList(args),
" version = " + VersionInfo.getVersion(),
" classpath = " + System.getProperty("java.class.path"),
" build = " + VersionInfo.getUrl() + " -r "
+ VersionInfo.getRevision()
+ "; compiled by '" + VersionInfo.getUser()
+ "' on " + VersionInfo.getDate(),
" java = " + System.getProperty("java.version") }
)
);
if (SystemUtils.IS_OS_UNIX) {
try {
SignalLogger.INSTANCE.register(LOG);
} catch (Throwable t) {
LOG.warn("failed to register any UNIX signal loggers: ", t);
}
}
ShutdownHookManager.get().addShutdownHook(
new Runnable() {
@Override
public void run() {
LOG.info(toStartupShutdownString("SHUTDOWN_MSG: ", new String[]{
"Shutting down " + classname + " at " + hostname}));
}
}, SHUTDOWN_HOOK_PRIORITY);
}
/**
* The traditional binary prefixes, kilo, mega, ..., exa,
* which can be represented by a 64-bit integer.
* TraditionalBinaryPrefix symbol are case insensitive.
*/
public static enum TraditionalBinaryPrefix {
KILO(10),
MEGA(KILO.bitShift + 10),
GIGA(MEGA.bitShift + 10),
TERA(GIGA.bitShift + 10),
PETA(TERA.bitShift + 10),
EXA (PETA.bitShift + 10);
public final long value;
public final char symbol;
public final int bitShift;
public final long bitMask;
private TraditionalBinaryPrefix(int bitShift) {
this.bitShift = bitShift;
this.value = 1L << bitShift;
this.bitMask = this.value - 1L;
this.symbol = toString().charAt(0);
}
/**
* @return The TraditionalBinaryPrefix object corresponding to the symbol.
*/
public static TraditionalBinaryPrefix valueOf(char symbol) {
symbol = Character.toUpperCase(symbol);
for(TraditionalBinaryPrefix prefix : TraditionalBinaryPrefix.values()) {
if (symbol == prefix.symbol) {
return prefix;
}
}
throw new IllegalArgumentException("Unknown symbol '" + symbol + "'");
}
/**
* Convert a string to long.
* The input string is first be trimmed
* and then it is parsed with traditional binary prefix.
*
* For example,
* "-1230k" will be converted to -1230 * 1024 = -1259520;
* "891g" will be converted to 891 * 1024^3 = 956703965184;
*
* @param s input string
* @return a long value represented by the input string.
*/
public static long string2long(String s) {
s = s.trim();
final int lastpos = s.length() - 1;
final char lastchar = s.charAt(lastpos);
if (Character.isDigit(lastchar))
return Long.parseLong(s);
else {
long prefix;
try {
prefix = TraditionalBinaryPrefix.valueOf(lastchar).value;
} catch (IllegalArgumentException e) {
throw new IllegalArgumentException("Invalid size prefix '" + lastchar
+ "' in '" + s
+ "'. Allowed prefixes are k, m, g, t, p, e(case insensitive)");
}
long num = Long.parseLong(s.substring(0, lastpos));
if (num > (Long.MAX_VALUE/prefix) || num < (Long.MIN_VALUE/prefix)) {
throw new IllegalArgumentException(s + " does not fit in a Long");
}
return num * prefix;
}
}
/**
* Convert a long integer to a string with traditional binary prefix.
*
* @param n the value to be converted
* @param unit The unit, e.g. "B" for bytes.
* @param decimalPlaces The number of decimal places.
* @return a string with traditional binary prefix.
*/
public static String long2String(long n, String unit, int decimalPlaces) {
if (unit == null) {
unit = "";
}
//take care a special case
if (n == Long.MIN_VALUE) {
return "-8 " + EXA.symbol + unit;
}
final StringBuilder b = new StringBuilder();
//take care negative numbers
if (n < 0) {
b.append('-');
n = -n;
}
if (n < KILO.value) {
//no prefix
b.append(n);
return (unit.isEmpty()? b: b.append(" ").append(unit)).toString();
} else {
//find traditional binary prefix
int i = 0;
for(; i < values().length && n >= values()[i].value; i++);
TraditionalBinaryPrefix prefix = values()[i - 1];
if ((n & prefix.bitMask) == 0) {
//exact division
b.append(n >> prefix.bitShift);
} else {
final String format = "%." + decimalPlaces + "f";
String s = format(format, n/(double)prefix.value);
//check a special rounding up case
if (s.startsWith("1024")) {
prefix = values()[i];
s = format(format, n/(double)prefix.value);
}
b.append(s);
}
return b.append(' ').append(prefix.symbol).append(unit).toString();
}
}
}
/**
* Escapes HTML Special characters present in the string.
* @param string
* @return HTML Escaped String representation
*/
public static String escapeHTML(String string) {
if(string == null) {
return null;
}
StringBuilder sb = new StringBuilder();
boolean lastCharacterWasSpace = false;
char[] chars = string.toCharArray();
for(char c : chars) {
if(c == ' ') {
if(lastCharacterWasSpace){
lastCharacterWasSpace = false;
sb.append(" ");
}else {
lastCharacterWasSpace=true;
sb.append(" ");
}
}else {
lastCharacterWasSpace = false;
switch(c) {
case '<': sb.append("<"); break;
case '>': sb.append(">"); break;
case '&': sb.append("&"); break;
case '"': sb.append("""); break;
default : sb.append(c);break;
}
}
}
return sb.toString();
}
/**
* @return a byte description of the given long interger value.
*/
public static String byteDesc(long len) {
return TraditionalBinaryPrefix.long2String(len, "B", 2);
}
/** @deprecated use StringUtils.format("%.2f", d). */
@Deprecated
public static String limitDecimalTo2(double d) {
return format("%.2f", d);
}
/**
* Concatenates strings, using a separator.
*
* @param separator Separator to join with.
* @param strings Strings to join.
*/
public static String join(CharSequence separator, Iterable<?> strings) {
Iterator<?> i = strings.iterator();
if (!i.hasNext()) {
return "";
}
StringBuilder sb = new StringBuilder(i.next().toString());
while (i.hasNext()) {
sb.append(separator);
sb.append(i.next().toString());
}
return sb.toString();
}
/**
* Concatenates strings, using a separator.
*
* @param separator to join with
* @param strings to join
* @return the joined string
*/
public static String join(CharSequence separator, String[] strings) {
// Ideally we don't have to duplicate the code here if array is iterable.
StringBuilder sb = new StringBuilder();
boolean first = true;
for (String s : strings) {
if (first) {
first = false;
} else {
sb.append(separator);
}
sb.append(s);
}
return sb.toString();
}
/**
* Convert SOME_STUFF to SomeStuff
*
* @param s input string
* @return camelized string
*/
public static String camelize(String s) {
StringBuilder sb = new StringBuilder();
String[] words = split(StringUtils.toLowerCase(s), ESCAPE_CHAR, '_');
for (String word : words)
sb.append(org.apache.commons.lang.StringUtils.capitalize(word));
return sb.toString();
}
/**
* Matches a template string against a pattern, replaces matched tokens with
* the supplied replacements, and returns the result. The regular expression
* must use a capturing group. The value of the first capturing group is used
* to look up the replacement. If no replacement is found for the token, then
* it is replaced with the empty string.
*
* For example, assume template is "%foo%_%bar%_%baz%", pattern is "%(.*?)%",
* and replacements contains 2 entries, mapping "foo" to "zoo" and "baz" to
* "zaz". The result returned would be "zoo__zaz".
*
* @param template String template to receive replacements
* @param pattern Pattern to match for identifying tokens, must use a capturing
* group
* @param replacements Map<String, String> mapping tokens identified by the
* capturing group to their replacement values
* @return String template with replacements
*/
public static String replaceTokens(String template, Pattern pattern,
Map<String, String> replacements) {
StringBuffer sb = new StringBuffer();
Matcher matcher = pattern.matcher(template);
while (matcher.find()) {
String replacement = replacements.get(matcher.group(1));
if (replacement == null) {
replacement = "";
}
matcher.appendReplacement(sb, Matcher.quoteReplacement(replacement));
}
matcher.appendTail(sb);
return sb.toString();
}
/**
* Get stack trace for a given thread.
*/
public static String getStackTrace(Thread t) {
final StackTraceElement[] stackTrace = t.getStackTrace();
StringBuilder str = new StringBuilder();
for (StackTraceElement e : stackTrace) {
str.append(e.toString() + "\n");
}
return str.toString();
}
/**
* From a list of command-line arguments, remove both an option and the
* next argument.
*
* @param name Name of the option to remove. Example: -foo.
* @param args List of arguments.
* @return null if the option was not found; the value of the
* option otherwise.
* @throws IllegalArgumentException if the option's argument is not present
*/
public static String popOptionWithArgument(String name, List<String> args)
throws IllegalArgumentException {
String val = null;
for (Iterator<String> iter = args.iterator(); iter.hasNext(); ) {
String cur = iter.next();
if (cur.equals("--")) {
// stop parsing arguments when you see --
break;
} else if (cur.equals(name)) {
iter.remove();
if (!iter.hasNext()) {
throw new IllegalArgumentException("option " + name + " requires 1 " +
"argument.");
}
val = iter.next();
iter.remove();
break;
}
}
return val;
}
/**
* From a list of command-line arguments, remove an option.
*
* @param name Name of the option to remove. Example: -foo.
* @param args List of arguments.
* @return true if the option was found and removed; false otherwise.
*/
public static boolean popOption(String name, List<String> args) {
for (Iterator<String> iter = args.iterator(); iter.hasNext(); ) {
String cur = iter.next();
if (cur.equals("--")) {
// stop parsing arguments when you see --
break;
} else if (cur.equals(name)) {
iter.remove();
return true;
}
}
return false;
}
/**
* From a list of command-line arguments, return the first non-option
* argument. Non-option arguments are those which either come after
* a double dash (--) or do not start with a dash.
*
* @param args List of arguments.
* @return The first non-option argument, or null if there were none.
*/
public static String popFirstNonOption(List<String> args) {
for (Iterator<String> iter = args.iterator(); iter.hasNext(); ) {
String cur = iter.next();
if (cur.equals("--")) {
if (!iter.hasNext()) {
return null;
}
cur = iter.next();
iter.remove();
return cur;
} else if (!cur.startsWith("-")) {
iter.remove();
return cur;
}
}
return null;
}
/**
* Converts all of the characters in this String to lower case with
* Locale.ENGLISH.
*
* @param str string to be converted
* @return the str, converted to lowercase.
*/
public static String toLowerCase(String str) {
return str.toLowerCase(Locale.ENGLISH);
}
/**
* Converts all of the characters in this String to upper case with
* Locale.ENGLISH.
*
* @param str string to be converted
* @return the str, converted to uppercase.
*/
public static String toUpperCase(String str) {
return str.toUpperCase(Locale.ENGLISH);
}
/**
* Compare strings locale-freely by using String#equalsIgnoreCase.
*
* @param s1 Non-null string to be converted
* @param s2 string to be converted
* @return the str, converted to uppercase.
*/
public static boolean equalsIgnoreCase(String s1, String s2) {
Preconditions.checkNotNull(s1);
// don't check non-null against s2 to make the semantics same as
// s1.equals(s2)
return s1.equalsIgnoreCase(s2);
}
}