[3.14] gh-88091: Fix unicodedata.decomposition() for Hangul Syllables (GH-144993...

(cherry picked from commit 56c4f10d6e474604a162521228b5f3b5ff79236c)

diff --git a/Lib/test/test_unicodedata.py b/Lib/test/test_unicodedata.py
index abbcffbe3fcee9f6743096e59f3ce5e9993a93b8..93d573996407590982168d2a8a2e7d58ace99fa5 100644 (file)
--- a/Lib/test/test_unicodedata.py
+++ b/Lib/test/test_unicodedata.py
@@ -89,9 +89,9 @@ class UnicodeFunctionsTest(unittest.TestCase):
 
     # Update this if the database changes. Make sure to do a full rebuild
     # (e.g. 'make distclean && make') to get the correct checksum.
-    expectedchecksum = ('35e842600fa7ae2db93739db08ef201b726a2374'
+    expectedchecksum = ('1ba453ec456896f1190d849b6e9b7c2e1a4128e0'
                         if quicktest else
-                        '23ab09ed4abdf93db23b97359108ed630dd8311d')
+                        '46ca89d9fe34881d0be3a4a4b29f5aa8c019640c')
 
     def test_function_checksum(self):
         db = self.db
@@ -346,6 +346,12 @@ class UnicodeFunctionsTest(unittest.TestCase):
         # New in 16.0.0
         self.assertEqual(self.db.decomposition('\U0001CCD6'), '' if self.old else '<font> 0041')
 
+        # Hangul characters
+        self.assertEqual(self.db.decomposition('\uAC00'), '1100 1161')
+        self.assertEqual(self.db.decomposition('\uD4DB'), '1111 1171 11B6')
+        self.assertEqual(self.db.decomposition('\uC2F8'), '110A 1161')
+        self.assertEqual(self.db.decomposition('\uD7A3'), '1112 1175 11C2')
+
         self.assertRaises(TypeError, self.db.decomposition)
         self.assertRaises(TypeError, self.db.decomposition, 'xx')
 
@@ -649,9 +655,9 @@ class UnicodeFunctionsTest(unittest.TestCase):
 class Unicode_3_2_0_FunctionsTest(UnicodeFunctionsTest):
     db = unicodedata.ucd_3_2_0
     old = True
-    expectedchecksum = ('4154d8d1232837e255edf3cdcbb5ab184d71f4a4'
+    expectedchecksum = ('883824cb6c0ccf994e4451ebf281e2d6d479af47'
                         if quicktest else
-                        'b0a8df4ce8cf910def4e75f2d03c93defcc9bb09')
+                        'caf1a7f2f380f927461837f1901ef20683f98683')
 
 
 class UnicodeMiscTest(unittest.TestCase):

diff --git a/Misc/NEWS.d/next/Library/2026-02-19-10-57-40.gh-issue-88091.N7qGV-.rst b/Misc/NEWS.d/next/Library/2026-02-19-10-57-40.gh-issue-88091.N7qGV-.rst
new file mode 100644 (file)
index 0000000..15cf250
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2026-02-19-10-57-40.gh-issue-88091.N7qGV-.rst
@@ -0,0 +1 @@
+Fix :func:`unicodedata.decomposition` for Hangul characters.

diff --git a/Modules/unicodedata.c b/Modules/unicodedata.c
index 0547046394485451e5fcee8ca95840d8d5a0eae0..83de1be56a7fafc9ee83171c13cd3ee3c2f929df 100644 (file)
--- a/Modules/unicodedata.c
+++ b/Modules/unicodedata.c
@@ -388,6 +388,17 @@ unicodedata_UCD_east_asian_width_impl(PyObject *self, int chr)
     return PyUnicode_FromString(_PyUnicode_EastAsianWidthNames[index]);
 }
 
+// For Hangul decomposition
+#define SBase   0xAC00
+#define LBase   0x1100
+#define VBase   0x1161
+#define TBase   0x11A7
+#define LCount  19
+#define VCount  21
+#define TCount  28
+#define NCount  (VCount*TCount)
+#define SCount  (LCount*NCount)
+
 /*[clinic input]
 unicodedata.UCD.decomposition
 
@@ -418,6 +429,25 @@ unicodedata_UCD_decomposition_impl(PyObject *self, int chr)
             return Py_GetConstant(Py_CONSTANT_EMPTY_STR); /* unassigned */
     }
 
+    // Hangul Decomposition.
+    // See section 3.12.2, "Hangul Syllable Decomposition"
+    // https://www.unicode.org/versions/latest/core-spec/chapter-3/#G56669
+    if (SBase <= code && code < (SBase + SCount)) {
+        int SIndex = code - SBase;
+        int L = LBase + SIndex / NCount;
+        int V = VBase + (SIndex % NCount) / TCount;
+        int T = TBase + SIndex % TCount;
+        if (T != TBase) {
+            PyOS_snprintf(decomp, sizeof(decomp),
+                          "%04X %04X %04X", L, V, T);
+        }
+        else {
+            PyOS_snprintf(decomp, sizeof(decomp),
+                          "%04X %04X", L, V);
+        }
+        return PyUnicode_FromString(decomp);
+    }
+
     if (code < 0 || code >= 0x110000)
         index = 0;
     else {
@@ -480,16 +510,6 @@ get_decomp_record(PyObject *self, Py_UCS4 code,
     (*index)++;
 }
 
-#define SBase   0xAC00
-#define LBase   0x1100
-#define VBase   0x1161
-#define TBase   0x11A7
-#define LCount  19
-#define VCount  21
-#define TCount  28
-#define NCount  (VCount*TCount)
-#define SCount  (LCount*NCount)
-
 static PyObject*
 nfd_nfkd(PyObject *self, PyObject *input, int k)
 {
@@ -543,7 +563,9 @@ nfd_nfkd(PyObject *self, PyObject *input, int k)
                 }
                 output = new_output;
             }
-            /* Hangul Decomposition. */
+            // Hangul Decomposition.
+            // See section 3.12.2, "Hangul Syllable Decomposition"
+            // https://www.unicode.org/versions/latest/core-spec/chapter-3/#G56669
             if (SBase <= code && code < (SBase+SCount)) {
                 int SIndex = code - SBase;
                 int L = LBase + SIndex / NCount;