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