Implement names for CJK unified ideographs. Add name to KeyError output.

Verify that the lookup for an existing name succeeds.

4 files changed, 59 insertions, 8 deletions

diff --git a/Lib/test/output/test_ucn b/Lib/test/output/test_ucn
index 1006c07..c41017b 100644
--- a/Lib/test/output/test_ucn
+++ b/Lib/test/output/test_ucn
@@ -2,7 +2,8 @@ test_ucn
 Testing General Unicode Character Name, and case insensitivity... done.
 Testing name to code mapping.... done.
 Testing hangul syllable names.... done.
-Testing code to name mapping for all characters.... done.
-Found 22728 characters in the unicode name database
+Testing names of CJK unified ideographs.... done.
+Testing code to name mapping for all BMP characters.... done.
+Found 50212 characters in the unicode name database
 Testing misc. symbols for unicode character name expansion.... done.
 Testing unicode character name expansion strict error handling.... done.
diff --git a/Lib/test/test_ucn.py b/Lib/test/test_ucn.py
index 6f2b022..e7b8bbd 100644
--- a/Lib/test/test_ucn.py
+++ b/Lib/test/test_ucn.py
@@ -80,16 +80,28 @@ else:
     raise AssertionError, "Found name for U+D7A4"
 print "done."
 
-print "Testing code to name mapping for all characters....",
+print "Testing names of CJK unified ideographs....",
+exec r"""
+verify(u"\N{CJK UNIFIED IDEOGRAPH-3400}" == u"\u3400")
+verify(u"\N{CJK UNIFIED IDEOGRAPH-4DB5}" == u"\u4db5")
+verify(u"\N{CJK UNIFIED IDEOGRAPH-4E00}" == u"\u4e00")
+verify(u"\N{CJK UNIFIED IDEOGRAPH-9FA5}" == u"\u9fa5")
+verify(u"\N{CJK UNIFIED IDEOGRAPH-20000}" == u"\U00020000")
+verify(u"\N{CJK UNIFIED IDEOGRAPH-2A6D6}" == u"\U0002a6d6")
+"""
+print "done."
+
+print "Testing code to name mapping for all BMP characters....",
 count = 0
-for code in range(65536):
+for code in range(0x10000):
     try:
         char = unichr(code)
         name = unicodedata.name(char)
-        verify(unicodedata.lookup(name) == char)
-        count += 1
     except (KeyError, ValueError):
         pass
+    else:
+        verify(unicodedata.lookup(name) == char)
+        count += 1
 print "done."
 
 print "Found", count, "characters in the unicode name database"
diff --git a/Misc/NEWS b/Misc/NEWS
index 71da82d..8fdd1f6 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -318,7 +318,7 @@ Extension modules
   is now named bsddb185.
 
 - unicodedata was updated to Unicode 3.2. In now also supports names
-  for Hangul syllables.
+  for Hangul syllables and CJK unified ideographs.
 
 - resource.getrlimit() now returns longs instead of ints.
 
diff --git a/Modules/unicodedata.c b/Modules/unicodedata.c
index 330b376..3620936 100644
--- a/Modules/unicodedata.c
+++ b/Modules/unicodedata.c
@@ -348,6 +348,16 @@ _getucname(Py_UCS4 code, char* buffer, int buflen)
 	return 1;
     }
 
+    if ((0x3400 <= code && code <= 0x4DB5) ||  /* CJK Ideograph Extension A */
+        (0x4E00 <= code && code <= 0x9FA5) ||  /* CJK Ideograph */
+        (0x20000 <= code && code <= 0x2A6D6)) {/* CJK Ideograph Extension B */
+        if (buflen < 28)
+            /* Worst case: CJK UNIFIED IDEOGRAPH-20000 */
+            return 0;
+        sprintf(buffer, "CJK UNIFIED IDEOGRAPH-%X", code);
+        return 1;
+    }
+
     if (code >= 0x110000)
         return 0;
 
@@ -449,6 +459,30 @@ _getcode(const char* name, int namelen, Py_UCS4* code)
 	    *code = SBase + (L*VCount+V)*TCount + T;
 	    return 1;
 	}
+        /* Otherwise, it's an illegal syllable name. */
+        return 0;
+    }
+
+    /* Check for unified ideographs. */
+    if (strncmp(name, "CJK UNIFIED IDEOGRAPH-", 22) == 0) {
+        /* Four or five hexdigits must follow. */
+        v = 0;
+        name += 22;
+        namelen -= 22;
+        if (namelen != 4 && namelen != 5)
+            return 0;
+        while (namelen--) {
+            v *= 16;
+            if (*name >= '0' && *name <= '9')
+                v += *name - '0';
+            else if (*name >= 'A' && *name <= 'F')
+                v += *name - 'A' + 10;
+            else
+                return 0;
+            name++;
+        }
+        *code = v;
+        return 1;
     }
 
     /* the following is the same as python's dictionary lookup, with
@@ -535,7 +569,11 @@ unicodedata_lookup(PyObject* self, PyObject* args)
         return NULL;
 
     if (!_getcode(name, namelen, &code)) {
-        PyErr_SetString(PyExc_KeyError, "undefined character name");
+        char fmt[] = "undefined character name '%s'";
+        char *buf = PyMem_MALLOC(sizeof(fmt) + namelen);
+        sprintf(buf, fmt, name);
+        PyErr_SetString(PyExc_KeyError, buf);
+        PyMem_FREE(buf);
         return NULL;
     }