Enable StringUtils to split CharSequence like String#split
This is a groundwork for enabling LocaleSpan. To enable LocaleSpan everywhere, we need to switch to CharSequence from String so that Span infromation can be preserved end-to-end. To achieve this, we need to have CharSequence version of String#split. BUG: 16029304 Change-Id: I0dd103185dcf62fb1e25054a374340790e6a4678
Yohei Yukawa
parent
e645715b25
commit
eac8670830
|
|
@ -18,6 +18,7 @@ package com.android.inputmethod.latin.utils;
|
|||
|
||||
import static com.android.inputmethod.latin.Constants.CODE_UNSPECIFIED;
|
||||
|
||||
import android.text.Spanned;
|
||||
import android.text.TextUtils;
|
||||
|
||||
import com.android.inputmethod.annotations.UsedForTesting;
|
||||
|
|
@ -26,6 +27,8 @@ import com.android.inputmethod.latin.Constants;
|
|||
import java.util.ArrayList;
|
||||
import java.util.Arrays;
|
||||
import java.util.Locale;
|
||||
import java.util.regex.Matcher;
|
||||
import java.util.regex.Pattern;
|
||||
|
||||
public final class StringUtils {
|
||||
public static final int CAPITALIZE_NONE = 0; // No caps, or mixed case
|
||||
|
|
@ -503,6 +506,55 @@ public final class StringUtils {
|
|||
return lastIndex - i;
|
||||
}
|
||||
|
||||
/**
|
||||
* Splits the given {@code charSequence} with at occurrences of the given {@code regex}.
|
||||
* <p>
|
||||
* This is equivalent to
|
||||
* {@code charSequence.toString().split(regex, preserveTrailingEmptySegments ? -1 : 0)}
|
||||
* except that the spans are preserved in the result array.
|
||||
* </p>
|
||||
* @param input the character sequence to be split.
|
||||
* @param regex the regex pattern to be used as the separator.
|
||||
* @param preserveTrailingEmptySegments {@code true} to preserve the trailing empty
|
||||
* segments. Otherwise, trailing empty segments will be removed before being returned.
|
||||
* @return the array which contains the result. All the spans in the {@param input} is
|
||||
* preserved.
|
||||
*/
|
||||
@UsedForTesting
|
||||
public static CharSequence[] split(final CharSequence charSequence, final String regex,
|
||||
final boolean preserveTrailingEmptySegments) {
|
||||
// A short-cut for non-spanned strings.
|
||||
if (!(charSequence instanceof Spanned)) {
|
||||
// -1 means that trailing empty segments will be preserved.
|
||||
return charSequence.toString().split(regex, preserveTrailingEmptySegments ? -1 : 0);
|
||||
}
|
||||
|
||||
// Hereafter, emulate String.split for CharSequence.
|
||||
final ArrayList<CharSequence> sequences = new ArrayList<>();
|
||||
final Matcher matcher = Pattern.compile(regex).matcher(charSequence);
|
||||
int nextStart = 0;
|
||||
boolean matched = false;
|
||||
while (matcher.find()) {
|
||||
sequences.add(charSequence.subSequence(nextStart, matcher.start()));
|
||||
nextStart = matcher.end();
|
||||
matched = true;
|
||||
}
|
||||
if (!matched) {
|
||||
// never matched. preserveTrailingEmptySegments is ignored in this case.
|
||||
return new CharSequence[] { charSequence };
|
||||
}
|
||||
sequences.add(charSequence.subSequence(nextStart, charSequence.length()));
|
||||
if (!preserveTrailingEmptySegments) {
|
||||
for (int i = sequences.size() - 1; i >= 0; --i) {
|
||||
if (!TextUtils.isEmpty(sequences.get(i))) {
|
||||
break;
|
||||
}
|
||||
sequences.remove(i);
|
||||
}
|
||||
}
|
||||
return sequences.toArray(new CharSequence[sequences.size()]);
|
||||
}
|
||||
|
||||
@UsedForTesting
|
||||
public static class Stringizer<E> {
|
||||
public String stringize(final E element) {
|
||||
|
|
|
|||
|
|
@ -18,6 +18,9 @@ package com.android.inputmethod.latin.utils;
|
|||
|
||||
import android.test.AndroidTestCase;
|
||||
import android.test.suitebuilder.annotation.SmallTest;
|
||||
import android.text.SpannableString;
|
||||
import android.text.Spanned;
|
||||
import android.text.SpannedString;
|
||||
|
||||
import com.android.inputmethod.latin.Constants;
|
||||
|
||||
|
|
@ -326,4 +329,171 @@ public class StringAndJsonUtilsTests extends AndroidTestCase {
|
|||
assertEquals(1, StringUtils.getTrailingSingleQuotesCount("'word'"));
|
||||
assertEquals(0, StringUtils.getTrailingSingleQuotesCount("I'm"));
|
||||
}
|
||||
|
||||
private static void assertSpanCount(final int expectedCount, final CharSequence cs) {
|
||||
final int actualCount;
|
||||
if (cs instanceof Spanned) {
|
||||
final Spanned spanned = (Spanned) cs;
|
||||
actualCount = spanned.getSpans(0, spanned.length(), Object.class).length;
|
||||
} else {
|
||||
actualCount = 0;
|
||||
}
|
||||
assertEquals(expectedCount, actualCount);
|
||||
}
|
||||
|
||||
private static void assertSpan(final CharSequence cs, final Object expectedSpan,
|
||||
final int expectedStart, final int expectedEnd, final int expectedFlags) {
|
||||
assertTrue(cs instanceof Spanned);
|
||||
final Spanned spanned = (Spanned) cs;
|
||||
final Object[] actualSpans = spanned.getSpans(0, spanned.length(), Object.class);
|
||||
for (Object actualSpan : actualSpans) {
|
||||
if (actualSpan == expectedSpan) {
|
||||
final int actualStart = spanned.getSpanStart(actualSpan);
|
||||
final int actualEnd = spanned.getSpanEnd(actualSpan);
|
||||
final int actualFlags = spanned.getSpanFlags(actualSpan);
|
||||
assertEquals(expectedStart, actualStart);
|
||||
assertEquals(expectedEnd, actualEnd);
|
||||
assertEquals(expectedFlags, actualFlags);
|
||||
return;
|
||||
}
|
||||
}
|
||||
assertTrue(false);
|
||||
}
|
||||
|
||||
public void testSplitCharSequenceWithSpan() {
|
||||
// text: " a bcd efg hij "
|
||||
// span1: ^^^^^^^
|
||||
// span2: ^^^^^
|
||||
// span3: ^
|
||||
final SpannableString spannableString = new SpannableString(" a bcd efg hij ");
|
||||
final Object span1 = new Object();
|
||||
final Object span2 = new Object();
|
||||
final Object span3 = new Object();
|
||||
final int SPAN1_FLAGS = Spanned.SPAN_EXCLUSIVE_EXCLUSIVE;
|
||||
final int SPAN2_FLAGS = Spanned.SPAN_EXCLUSIVE_INCLUSIVE;
|
||||
final int SPAN3_FLAGS = Spanned.SPAN_INCLUSIVE_INCLUSIVE;
|
||||
spannableString.setSpan(span1, 0, 7, SPAN1_FLAGS);
|
||||
spannableString.setSpan(span2, 0, 5, SPAN2_FLAGS);
|
||||
spannableString.setSpan(span3, 12, 13, SPAN3_FLAGS);
|
||||
final CharSequence[] charSequencesFromSpanned = StringUtils.split(
|
||||
spannableString, " ", true /* preserveTrailingEmptySegmengs */);
|
||||
final CharSequence[] charSequencesFromString = StringUtils.split(
|
||||
spannableString.toString(), " ", true /* preserveTrailingEmptySegmengs */);
|
||||
|
||||
|
||||
assertEquals(7, charSequencesFromString.length);
|
||||
assertEquals(7, charSequencesFromSpanned.length);
|
||||
|
||||
// text: ""
|
||||
// span1: ^
|
||||
// span2: ^
|
||||
// span3:
|
||||
assertEquals("", charSequencesFromString[0].toString());
|
||||
assertSpanCount(0, charSequencesFromString[0]);
|
||||
assertEquals("", charSequencesFromSpanned[0].toString());
|
||||
assertSpanCount(2, charSequencesFromSpanned[0]);
|
||||
assertSpan(charSequencesFromSpanned[0], span1, 0, 0, SPAN1_FLAGS);
|
||||
assertSpan(charSequencesFromSpanned[0], span2, 0, 0, SPAN2_FLAGS);
|
||||
|
||||
// text: "a"
|
||||
// span1: ^
|
||||
// span2: ^
|
||||
// span3:
|
||||
assertEquals("a", charSequencesFromString[1].toString());
|
||||
assertSpanCount(0, charSequencesFromString[1]);
|
||||
assertEquals("a", charSequencesFromSpanned[1].toString());
|
||||
assertSpanCount(2, charSequencesFromSpanned[1]);
|
||||
assertSpan(charSequencesFromSpanned[1], span1, 0, 1, SPAN1_FLAGS);
|
||||
assertSpan(charSequencesFromSpanned[1], span2, 0, 1, SPAN2_FLAGS);
|
||||
|
||||
// text: "bcd"
|
||||
// span1: ^^^
|
||||
// span2: ^^
|
||||
// span3:
|
||||
assertEquals("bcd", charSequencesFromString[2].toString());
|
||||
assertSpanCount(0, charSequencesFromString[2]);
|
||||
assertEquals("bcd", charSequencesFromSpanned[2].toString());
|
||||
assertSpanCount(2, charSequencesFromSpanned[2]);
|
||||
assertSpan(charSequencesFromSpanned[2], span1, 0, 3, SPAN1_FLAGS);
|
||||
assertSpan(charSequencesFromSpanned[2], span2, 0, 2, SPAN2_FLAGS);
|
||||
|
||||
// text: "efg"
|
||||
// span1:
|
||||
// span2:
|
||||
// span3:
|
||||
assertEquals("efg", charSequencesFromString[3].toString());
|
||||
assertSpanCount(0, charSequencesFromString[3]);
|
||||
assertEquals("efg", charSequencesFromSpanned[3].toString());
|
||||
assertSpanCount(0, charSequencesFromSpanned[3]);
|
||||
|
||||
// text: "hij"
|
||||
// span1:
|
||||
// span2:
|
||||
// span3: ^
|
||||
assertEquals("hij", charSequencesFromString[4].toString());
|
||||
assertSpanCount(0, charSequencesFromString[4]);
|
||||
assertEquals("hij", charSequencesFromSpanned[4].toString());
|
||||
assertSpanCount(1, charSequencesFromSpanned[4]);
|
||||
assertSpan(charSequencesFromSpanned[4], span3, 1, 2, SPAN3_FLAGS);
|
||||
|
||||
// text: ""
|
||||
// span1:
|
||||
// span2:
|
||||
// span3:
|
||||
assertEquals("", charSequencesFromString[5].toString());
|
||||
assertSpanCount(0, charSequencesFromString[5]);
|
||||
assertEquals("", charSequencesFromSpanned[5].toString());
|
||||
assertSpanCount(0, charSequencesFromSpanned[5]);
|
||||
|
||||
// text: ""
|
||||
// span1:
|
||||
// span2:
|
||||
// span3:
|
||||
assertEquals("", charSequencesFromString[6].toString());
|
||||
assertSpanCount(0, charSequencesFromString[6]);
|
||||
assertEquals("", charSequencesFromSpanned[6].toString());
|
||||
assertSpanCount(0, charSequencesFromSpanned[6]);
|
||||
}
|
||||
|
||||
public void testSplitCharSequencePreserveTrailingEmptySegmengs() {
|
||||
assertEquals(1, StringUtils.split("", " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(1, StringUtils.split(new SpannedString(""), " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
|
||||
assertEquals(1, StringUtils.split("", " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(1, StringUtils.split(new SpannedString(""), " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
|
||||
assertEquals(0, StringUtils.split(" ", " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(0, StringUtils.split(new SpannedString(" "), " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
|
||||
assertEquals(2, StringUtils.split(" ", " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(2, StringUtils.split(new SpannedString(" "), " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
|
||||
assertEquals(3, StringUtils.split("a b c ", " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(3, StringUtils.split(new SpannedString("a b c "), " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
|
||||
assertEquals(5, StringUtils.split("a b c ", " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(5, StringUtils.split(new SpannedString("a b c "), " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
|
||||
assertEquals(6, StringUtils.split("a b ", " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(6, StringUtils.split(new SpannedString("a b "), " ",
|
||||
false /* preserveTrailingEmptySegmengs */).length);
|
||||
|
||||
assertEquals(7, StringUtils.split("a b ", " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
assertEquals(7, StringUtils.split(new SpannedString("a b "), " ",
|
||||
true /* preserveTrailingEmptySegmengs */).length);
|
||||
}
|
||||
}
|
||||
|
|
|
|||
Loading…
Reference in New Issue