ICU-20316 case mapping append unchanged text before calling low-level functions that may immediately append the mapping for a code point
diff --git a/icu4j/main/classes/core/src/com/ibm/icu/impl/CaseMapImpl.java b/icu4j/main/classes/core/src/com/ibm/icu/impl/CaseMapImpl.java
index 2b30c02..bdd3073 100644
--- a/icu4j/main/classes/core/src/com/ibm/icu/impl/CaseMapImpl.java
+++ b/icu4j/main/classes/core/src/com/ibm/icu/impl/CaseMapImpl.java
@@ -486,6 +486,11 @@
} else {
c = lead;
}
+ // We need to append unchanged text before calling the UCaseProps.toFullXyz() methods
+ // because they will sometimes append their mapping to dest,
+ // and that must be after copying the previous text.
+ appendUnchanged(src, prev, cpStart - prev, dest, options, edits);
+ prev = cpStart;
if (caseLocale >= 0) {
if (iter == null) {
iter = new StringContextIterator(src, cpStart, srcIndex);
@@ -497,7 +502,6 @@
c = UCaseProps.INSTANCE.toFullFolding(c, dest, options);
}
if (c >= 0) {
- appendUnchanged(src, prev, cpStart - prev, dest, options, edits);
appendResult(c, dest, srcIndex - cpStart, options, edits);
prev = srcIndex;
}
@@ -568,9 +572,13 @@
} else {
iter.setCPStartAndLimit(cpStart, srcIndex);
}
+ // We need to append unchanged text before calling UCaseProps.toFullUpper()
+ // because it will sometimes append its mapping to dest,
+ // and that must be after copying the previous text.
+ appendUnchanged(src, prev, cpStart - prev, dest, options, edits);
+ prev = cpStart;
c = UCaseProps.INSTANCE.toFullUpper(c, iter, dest, caseLocale);
if (c >= 0) {
- appendUnchanged(src, prev, cpStart - prev, dest, options, edits);
appendResult(c, dest, srcIndex - cpStart, options, edits);
prev = srcIndex;
}
diff --git a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/lang/UCharacterCaseTest.java b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/lang/UCharacterCaseTest.java
index 1649a71..7d2948d 100644
--- a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/lang/UCharacterCaseTest.java
+++ b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/lang/UCharacterCaseTest.java
@@ -960,8 +960,8 @@
// Iterate once forward, once backward, to cover more runtime conditions.
int srcLength = expSrcIndex;
int destLength = expDestIndex;
- List<Integer> srcIndexes = new ArrayList<Integer>();
- List<Integer> destIndexes = new ArrayList<Integer>();
+ List<Integer> srcIndexes = new ArrayList<>();
+ List<Integer> destIndexes = new ArrayList<>();
srcIndexes.add(-1);
destIndexes.add(-1);
int srcIndex = 0;
@@ -1503,6 +1503,18 @@
assertEquals("title", "\u1F88\u1F80\u1FF3", result);
}
+ @Test
+ public void TestFoldBug20316() {
+ String s = "廬ᾒ뻪ᣃइ垚Ⴡₓ렞체ꖲ갹ݖ䕷꾬쯎㊅ᦘᰄ㸜䡏遁럢豑黾奯㸀⊻줮끎蒹衤劔뽳趧熶撒쫃窩겨ཇ脌쵐嫑⟑겭㋋濜隣ᳰ봢ℼ櫩靛㉃炔鋳" +
+ "оे⳨ᦧྃ깢粣ᑤꇪ찃̹鵄ዤꛛᰙ⡝捣쯋톐蕩栭쥀뎊ᄯ৻恳〬昴껤룩列潱ᑮ煃鶖안꽊鹭宪帐❖ा쥈잔";
+ String result = CaseMap.fold().apply(s);
+ assertTrue("廬ᾒ...->廬ἢι...", result.startsWith("廬ἢι"));
+ s = "儊ẖ깸ᝓ恷ᇁ䜄쌼ꇸჃ䗑䘬䒥㈴槁蛚紆洔㖣믏亝醣黹Ά嶨䖕篕舀ꖧ₭ଯᒗ✧ԗ墖쁳㽎苊澎긁⾆⒞蠻왃囨ᡠ邏꾭⪐턣搤穳≠톲絋砖ሷ⠆" +
+ "瞏惢鵶剕듘ᅤ♟Ԡⴠ⊡鹔ጙ갑⣚堟ᣗ✸㕇絮䠎瘗⟡놥擢ꉭ佱ྪ飹痵⿑⨴츿璿僖㯷넴鋰膄釚겼ナ黪差";
+ result = CaseMap.fold().apply(s);
+ assertTrue("儊ẖ...->儊h\u0331...", result.startsWith("儊h\u0331"));
+ }
+
// private data members - test data --------------------------------------
private static final Locale TURKISH_LOCALE_ = new Locale("tr", "TR");
@@ -1747,7 +1759,7 @@
*/
private String[] getUnicodeStrings(String str)
{
- List<String> v = new ArrayList<String>(10);
+ List<String> v = new ArrayList<>(10);
int start = 0;
for (int casecount = 4; casecount > 0; casecount --) {
int end = str.indexOf("; ", start);
