001/*
002 * Copyright (C) 2010 The Guava Authors
003 *
004 * Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
005 * in compliance with the License. You may obtain a copy of the License at
006 *
007 * http://www.apache.org/licenses/LICENSE-2.0
008 *
009 * Unless required by applicable law or agreed to in writing, software distributed under the License
010 * is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
011 * or implied. See the License for the specific language governing permissions and limitations under
012 * the License.
013 */
014
015package com.google.common.base;
016
017import static com.google.common.base.Preconditions.checkArgument;
018import static com.google.common.base.Preconditions.checkNotNull;
019import static java.lang.Math.min;
020import static java.util.logging.Level.WARNING;
021
022import com.google.common.annotations.GwtCompatible;
023import com.google.common.annotations.VisibleForTesting;
024import com.google.errorprone.annotations.InlineMe;
025import com.google.errorprone.annotations.InlineMeValidationDisabled;
026import java.util.logging.Logger;
027import javax.annotation.CheckForNull;
028import org.checkerframework.checker.nullness.qual.Nullable;
029
030/**
031 * Static utility methods pertaining to {@code String} or {@code CharSequence} instances.
032 *
033 * @author Kevin Bourrillion
034 * @since 3.0
035 */
036@GwtCompatible
037@ElementTypesAreNonnullByDefault
038public final class Strings {
039  private Strings() {}
040
041  /**
042   * Returns the given string if it is non-null; the empty string otherwise.
043   *
044   * @param string the string to test and possibly return
045   * @return {@code string} itself if it is non-null; {@code ""} if it is null
046   */
047  public static String nullToEmpty(@CheckForNull String string) {
048    return Platform.nullToEmpty(string);
049  }
050
051  /**
052   * Returns the given string if it is nonempty; {@code null} otherwise.
053   *
054   * @param string the string to test and possibly return
055   * @return {@code string} itself if it is nonempty; {@code null} if it is empty or null
056   */
057  @CheckForNull
058  public static String emptyToNull(@CheckForNull String string) {
059    return Platform.emptyToNull(string);
060  }
061
062  /**
063   * Returns {@code true} if the given string is null or is the empty string.
064   *
065   * <p>Consider normalizing your string references with {@link #nullToEmpty}. If you do, you can
066   * use {@link String#isEmpty()} instead of this method, and you won't need special null-safe forms
067   * of methods like {@link String#toUpperCase} either. Or, if you'd like to normalize "in the other
068   * direction," converting empty strings to {@code null}, you can use {@link #emptyToNull}.
069   *
070   * @param string a string reference to check
071   * @return {@code true} if the string is null or is the empty string
072   */
073  public static boolean isNullOrEmpty(@CheckForNull String string) {
074    return Platform.stringIsNullOrEmpty(string);
075  }
076
077  /**
078   * Returns a string, of length at least {@code minLength}, consisting of {@code string} prepended
079   * with as many copies of {@code padChar} as are necessary to reach that length. For example,
080   *
081   * <ul>
082   *   <li>{@code padStart("7", 3, '0')} returns {@code "007"}
083   *   <li>{@code padStart("2010", 3, '0')} returns {@code "2010"}
084   * </ul>
085   *
086   * <p>See {@link java.util.Formatter} for a richer set of formatting capabilities.
087   *
088   * @param string the string which should appear at the end of the result
089   * @param minLength the minimum length the resulting string must have. Can be zero or negative, in
090   *     which case the input string is always returned.
091   * @param padChar the character to insert at the beginning of the result until the minimum length
092   *     is reached
093   * @return the padded string
094   */
095  public static String padStart(String string, int minLength, char padChar) {
096    checkNotNull(string); // eager for GWT.
097    if (string.length() >= minLength) {
098      return string;
099    }
100    StringBuilder sb = new StringBuilder(minLength);
101    for (int i = string.length(); i < minLength; i++) {
102      sb.append(padChar);
103    }
104    sb.append(string);
105    return sb.toString();
106  }
107
108  /**
109   * Returns a string, of length at least {@code minLength}, consisting of {@code string} appended
110   * with as many copies of {@code padChar} as are necessary to reach that length. For example,
111   *
112   * <ul>
113   *   <li>{@code padEnd("4.", 5, '0')} returns {@code "4.000"}
114   *   <li>{@code padEnd("2010", 3, '!')} returns {@code "2010"}
115   * </ul>
116   *
117   * <p>See {@link java.util.Formatter} for a richer set of formatting capabilities.
118   *
119   * @param string the string which should appear at the beginning of the result
120   * @param minLength the minimum length the resulting string must have. Can be zero or negative, in
121   *     which case the input string is always returned.
122   * @param padChar the character to append to the end of the result until the minimum length is
123   *     reached
124   * @return the padded string
125   */
126  public static String padEnd(String string, int minLength, char padChar) {
127    checkNotNull(string); // eager for GWT.
128    if (string.length() >= minLength) {
129      return string;
130    }
131    StringBuilder sb = new StringBuilder(minLength);
132    sb.append(string);
133    for (int i = string.length(); i < minLength; i++) {
134      sb.append(padChar);
135    }
136    return sb.toString();
137  }
138
139  /**
140   * Returns a string consisting of a specific number of concatenated copies of an input string. For
141   * example, {@code repeat("hey", 3)} returns the string {@code "heyheyhey"}.
142   *
143   * <p><b>Java 11+ users:</b> use {@code string.repeat(count)} instead.
144   *
145   * @param string any non-null string
146   * @param count the number of times to repeat it; a nonnegative integer
147   * @return a string containing {@code string} repeated {@code count} times (the empty string if
148   *     {@code count} is zero)
149   * @throws IllegalArgumentException if {@code count} is negative
150   */
151  @InlineMe(replacement = "string.repeat(count)")
152  @InlineMeValidationDisabled("Java 11+ API only")
153  public static String repeat(String string, int count) {
154    checkNotNull(string); // eager for GWT.
155
156    if (count <= 1) {
157      checkArgument(count >= 0, "invalid count: %s", count);
158      return (count == 0) ? "" : string;
159    }
160
161    // IF YOU MODIFY THE CODE HERE, you must update StringsRepeatBenchmark
162    final int len = string.length();
163    final long longSize = (long) len * (long) count;
164    final int size = (int) longSize;
165    if (size != longSize) {
166      throw new ArrayIndexOutOfBoundsException("Required array size too large: " + longSize);
167    }
168
169    final char[] array = new char[size];
170    string.getChars(0, len, array, 0);
171    int n;
172    for (n = len; n < size - n; n <<= 1) {
173      System.arraycopy(array, 0, array, n, n);
174    }
175    System.arraycopy(array, 0, array, n, size - n);
176    return new String(array);
177  }
178
179  /**
180   * Returns the longest string {@code prefix} such that {@code a.toString().startsWith(prefix) &&
181   * b.toString().startsWith(prefix)}, taking care not to split surrogate pairs. If {@code a} and
182   * {@code b} have no common prefix, returns the empty string.
183   *
184   * @since 11.0
185   */
186  public static String commonPrefix(CharSequence a, CharSequence b) {
187    checkNotNull(a);
188    checkNotNull(b);
189
190    int maxPrefixLength = min(a.length(), b.length());
191    int p = 0;
192    while (p < maxPrefixLength && a.charAt(p) == b.charAt(p)) {
193      p++;
194    }
195    if (validSurrogatePairAt(a, p - 1) || validSurrogatePairAt(b, p - 1)) {
196      p--;
197    }
198    return a.subSequence(0, p).toString();
199  }
200
201  /**
202   * Returns the longest string {@code suffix} such that {@code a.toString().endsWith(suffix) &&
203   * b.toString().endsWith(suffix)}, taking care not to split surrogate pairs. If {@code a} and
204   * {@code b} have no common suffix, returns the empty string.
205   *
206   * @since 11.0
207   */
208  public static String commonSuffix(CharSequence a, CharSequence b) {
209    checkNotNull(a);
210    checkNotNull(b);
211
212    int maxSuffixLength = min(a.length(), b.length());
213    int s = 0;
214    while (s < maxSuffixLength && a.charAt(a.length() - s - 1) == b.charAt(b.length() - s - 1)) {
215      s++;
216    }
217    if (validSurrogatePairAt(a, a.length() - s - 1)
218        || validSurrogatePairAt(b, b.length() - s - 1)) {
219      s--;
220    }
221    return a.subSequence(a.length() - s, a.length()).toString();
222  }
223
224  /**
225   * True when a valid surrogate pair starts at the given {@code index} in the given {@code string}.
226   * Out-of-range indexes return false.
227   */
228  @VisibleForTesting
229  static boolean validSurrogatePairAt(CharSequence string, int index) {
230    return index >= 0
231        && index <= (string.length() - 2)
232        && Character.isHighSurrogate(string.charAt(index))
233        && Character.isLowSurrogate(string.charAt(index + 1));
234  }
235
236  /**
237   * Returns the given {@code template} string with each occurrence of {@code "%s"} replaced with
238   * the corresponding argument value from {@code args}; or, if the placeholder and argument counts
239   * do not match, returns a best-effort form of that string. Will not throw an exception under
240   * normal conditions.
241   *
242   * <p><b>Note:</b> For most string-formatting needs, use {@link String#format String.format},
243   * {@link java.io.PrintWriter#format PrintWriter.format}, and related methods. These support the
244   * full range of <a
245   * href="https://docs.oracle.com/javase/9/docs/api/java/util/Formatter.html#syntax">format
246   * specifiers</a>, and alert you to usage errors by throwing {@link
247   * java.util.IllegalFormatException}.
248   *
249   * <p>In certain cases, such as outputting debugging information or constructing a message to be
250   * used for another unchecked exception, an exception during string formatting would serve little
251   * purpose except to supplant the real information you were trying to provide. These are the cases
252   * this method is made for; it instead generates a best-effort string with all supplied argument
253   * values present. This method is also useful in environments such as GWT where {@code
254   * String.format} is not available. As an example, method implementations of the {@link
255   * Preconditions} class use this formatter, for both of the reasons just discussed.
256   *
257   * <p><b>Warning:</b> Only the exact two-character placeholder sequence {@code "%s"} is
258   * recognized.
259   *
260   * @param template a string containing zero or more {@code "%s"} placeholder sequences. {@code
261   *     null} is treated as the four-character string {@code "null"}.
262   * @param args the arguments to be substituted into the message template. The first argument
263   *     specified is substituted for the first occurrence of {@code "%s"} in the template, and so
264   *     forth. A {@code null} argument is converted to the four-character string {@code "null"};
265   *     non-null values are converted to strings using {@link Object#toString()}.
266   * @since 25.1
267   */
268  // TODO(diamondm) consider using Arrays.toString() for array parameters
269  public static String lenientFormat(
270      @CheckForNull String template, @CheckForNull @Nullable Object... args) {
271    template = String.valueOf(template); // null -> "null"
272
273    if (args == null) {
274      args = new Object[] {"(Object[])null"};
275    } else {
276      for (int i = 0; i < args.length; i++) {
277        args[i] = lenientToString(args[i]);
278      }
279    }
280
281    // start substituting the arguments into the '%s' placeholders
282    StringBuilder builder = new StringBuilder(template.length() + 16 * args.length);
283    int templateStart = 0;
284    int i = 0;
285    while (i < args.length) {
286      int placeholderStart = template.indexOf("%s", templateStart);
287      if (placeholderStart == -1) {
288        break;
289      }
290      builder.append(template, templateStart, placeholderStart);
291      builder.append(args[i++]);
292      templateStart = placeholderStart + 2;
293    }
294    builder.append(template, templateStart, template.length());
295
296    // if we run out of placeholders, append the extra args in square braces
297    if (i < args.length) {
298      builder.append(" [");
299      builder.append(args[i++]);
300      while (i < args.length) {
301        builder.append(", ");
302        builder.append(args[i++]);
303      }
304      builder.append(']');
305    }
306
307    return builder.toString();
308  }
309
310  private static String lenientToString(@CheckForNull Object o) {
311    if (o == null) {
312      return "null";
313    }
314    try {
315      return o.toString();
316    } catch (Exception e) {
317      // Default toString() behavior - see Object.toString()
318      String objectToString =
319          o.getClass().getName() + '@' + Integer.toHexString(System.identityHashCode(o));
320      // Logger is created inline with fixed name to avoid forcing Proguard to create another class.
321      Logger.getLogger("com.google.common.base.Strings")
322          .log(WARNING, "Exception during lenientFormat for " + objectToString, e);
323      return "<" + objectToString + " threw " + e.getClass().getName() + ">";
324    }
325  }
326}