ICU-11802 add 4 Emoji properties from emoji-data.txt 2.0

X-SVN-Rev: 38183

icu4c/source/common/ubidi_props_data.h

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 1999-2015, International Business Machines
+ * Copyright (C) 1999-2016, International Business Machines
  * Corporation and others.  All Rights Reserved.
  *
  * file name: ubidi_props_data.h

icu4c/source/common/ucase_props_data.h

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 1999-2015, International Business Machines
+ * Copyright (C) 1999-2016, International Business Machines
  * Corporation and others.  All Rights Reserved.
  *
  * file name: ucase_props_data.h

icu4c/source/common/unicode/uchar.h

@@ -1,6 +1,6 @@
 /*
 **********************************************************************
-*   Copyright (C) 1997-2015, International Business Machines
+*   Copyright (C) 1997-2016, International Business Machines
 *   Corporation and others.  All Rights Reserved.
 **********************************************************************
 *
@@ -397,8 +397,36 @@ typedef enum UProperty {
     UCHAR_CHANGES_WHEN_CASEMAPPED=55,
     /** Binary property Changes_When_NFKC_Casefolded. @stable ICU 4.4 */
     UCHAR_CHANGES_WHEN_NFKC_CASEFOLDED=56,
+    /**
+     * Binary property Emoji.
+     * See http://www.unicode.org/reports/tr51/#Emoji_Properties
+     *
+     * @draft ICU 57
+     */
+    UCHAR_EMOJI=57,
+    /**
+     * Binary property Emoji_Presentation.
+     * See http://www.unicode.org/reports/tr51/#Emoji_Properties
+     *
+     * @draft ICU 57
+     */
+    UCHAR_EMOJI_PRESENTATION=58,
+    /**
+     * Binary property Emoji_Modifier.
+     * See http://www.unicode.org/reports/tr51/#Emoji_Properties
+     *
+     * @draft ICU 57
+     */
+    UCHAR_EMOJI_MODIFIER=59,
+    /**
+     * Binary property Emoji_Modifier_Base.
+     * See http://www.unicode.org/reports/tr51/#Emoji_Properties
+     *
+     * @draft ICU 57
+     */
+    UCHAR_EMOJI_MODIFIER_BASE=60,
     /** One more than the last constant for binary Unicode properties. @stable ICU 2.1 */
-    UCHAR_BINARY_LIMIT=57,
+    UCHAR_BINARY_LIMIT=61,
 
     /** Enumerated property Bidi_Class.
         Same as u_charDirection, returns UCharDirection values. @stable ICU 2.2 */

icu4c/source/common/uprops.cpp

@@ -1,7 +1,7 @@
 /*
 *******************************************************************************
 *
-*   Copyright (C) 2002-2014, International Business Machines
+*   Copyright (C) 2002-2016, International Business Machines
 *   Corporation and others.  All Rights Reserved.
 *
 *******************************************************************************
@@ -270,7 +270,11 @@ static const BinaryProperty binProps[UCHAR_BINARY_LIMIT]={
     { UPROPS_SRC_CASE,  0, caseBinaryPropertyContains },  // UCHAR_CHANGES_WHEN_TITLECASED
     { UPROPS_SRC_CASE_AND_NORM,  0, changesWhenCasefolded },
     { UPROPS_SRC_CASE,  0, caseBinaryPropertyContains },  // UCHAR_CHANGES_WHEN_CASEMAPPED
-    { UPROPS_SRC_NFKC_CF, 0, changesWhenNFKC_Casefolded }
+    { UPROPS_SRC_NFKC_CF, 0, changesWhenNFKC_Casefolded },
+    { 2,                U_MASK(UPROPS_2_EMOJI), defaultContains },
+    { 2,                U_MASK(UPROPS_2_EMOJI_PRESENTATION), defaultContains },
+    { 2,                U_MASK(UPROPS_2_EMOJI_MODIFIER), defaultContains },
+    { 2,                U_MASK(UPROPS_2_EMOJI_MODIFIER_BASE), defaultContains },
 };
 
 U_CAPI UBool U_EXPORT2

icu4c/source/common/uprops.h

@@ -1,7 +1,7 @@
 /*
 *******************************************************************************
 *
-*   Copyright (C) 2002-2012, International Business Machines
+*   Copyright (C) 2002-2016, International Business Machines
 *   Corporation and others.  All Rights Reserved.
 *
 *******************************************************************************
@@ -187,13 +187,21 @@ enum {
 /*
  * Properties in vector word 2
  * Bits
- * 31..26   reserved
+ * 31..28   http://www.unicode.org/reports/tr51/#Emoji_Properties
+ * 27..26   reserved
  * 25..20   Line Break
  * 19..15   Sentence Break
  * 14..10   Word Break
  *  9.. 5   Grapheme Cluster Break
  *  4.. 0   Decomposition Type
  */
+enum {
+    UPROPS_2_EMOJI=28,
+    UPROPS_2_EMOJI_PRESENTATION,
+    UPROPS_2_EMOJI_MODIFIER,
+    UPROPS_2_EMOJI_MODIFIER_BASE,
+};
+
 #define UPROPS_LB_MASK          0x03f00000
 #define UPROPS_LB_SHIFT         20
 

icu4c/source/data/unidata/changes.txt

@@ -1,4 +1,4 @@
-* Copyright (C) 2004-2015, International Business Machines
+* Copyright (C) 2004-2016, International Business Machines
 * Corporation and others.  All Rights Reserved.
 *
 *   file name:  changes.txt
@@ -45,6 +45,35 @@ Khitan scripts will be encoded later.
 
 ---------------------------------------------------------------------------- ***
 
+Emoji properties added in ICU 57: http://bugs.icu-project.org/trac/ticket/11802
+
+Edit preparseucd.py to add & parse new properties.
+They share the UCD property namespace but are not listed in PropertyAliases.txt.
+
+Add emoji-data.txt to the input files, from http://www.unicode.org/Public/emoji/
+Initial data from emoji/2.0/
+
+ICU_ROOT=~/svn.icu/trunk
+ICU_SRC_DIR=$ICU_ROOT/src
+ICUDT=icudt56b
+export LD_LIBRARY_PATH=$ICU_ROOT/dbg/lib
+SRC_DATA_IN=$ICU_SRC_DIR/source/data/in
+UNIDATA=$ICU_SRC_DIR/source/data/unidata
+
+Add binary-property constants to uchar.h enum UProperty & UProperty.java.
+
+~/svn.icutools/trunk/src/unicode$ py/preparseucd.py ~/unidata/uni80/20151217 $ICU_SRC_DIR ~/svn.icutools/trunk/src
+(Needs to be run after uchar.h additions, so that the new properties can be picked up by genprops.)
+
+Data structure: uprops.h/.cpp, corepropsbuilder.cpp, UCharacterProperty.java
+
+make install, then icutools cmake & make, then
+~/svn.icutools/trunk/dbg/unicode/c$ make && genprops/genprops $ICU_SRC_DIR
+
+Generate Java data as usual, only update pnames.icu & uprops.icu.
+
+---------------------------------------------------------------------------- ***
+
 Unicode 8.0 update for ICU 56
 
 * Command-line environment setup

icu4c/source/data/unidata/norm2/nfc.txt

@@ -1,4 +1,4 @@
-# Copyright (C) 1999-2015, International Business Machines
+# Copyright (C) 1999-2016, International Business Machines
 # Corporation and others.  All Rights Reserved.
 #
 # file name: nfc.txt

icu4c/source/data/unidata/norm2/nfkc.txt

@@ -1,4 +1,4 @@
-# Copyright (C) 1999-2015, International Business Machines
+# Copyright (C) 1999-2016, International Business Machines
 # Corporation and others.  All Rights Reserved.
 #
 # file name: nfkc.txt

icu4c/source/data/unidata/norm2/nfkc_cf.txt

@@ -1,5 +1,5 @@
 # Unicode Character Database
-# Copyright (c) 1991-2015 Unicode, Inc.
+# Copyright (c) 1991-2016 Unicode, Inc.
 # For terms of use, see http://www.unicode.org/terms_of_use.html
 # For documentation, see http://www.unicode.org/reports/tr44/
 #

icu4c/source/test/intltest/ucdtest.cpp

@@ -1,6 +1,5 @@
 /********************************************************************
- * COPYRIGHT: 
- * Copyright (c) 1997-2014, International Business Machines Corporation and
+ * Copyright (c) 1997-2016, International Business Machines Corporation and
  * others. All Rights Reserved.
  ********************************************************************/
 
@@ -60,6 +59,7 @@ void UnicodeTest::runIndexedTest( int32_t index, UBool exec, const char* &name,
     TESTCASE_AUTO(TestPatternProperties);
     TESTCASE_AUTO(TestScriptMetadata);
     TESTCASE_AUTO(TestBidiPairedBracketType);
+    TESTCASE_AUTO(TestEmojiProperties);
     TESTCASE_AUTO_END;
 }
 
@@ -507,3 +507,18 @@ void UnicodeTest::TestBidiPairedBracketType() {
     assertTrue("bpt=Open is a subset of Ps", ps.containsAll(bpt_open));
     assertTrue("bpt=Close is a subset of Pe", pe.containsAll(bpt_close));
 }
+
+void UnicodeTest::TestEmojiProperties() {
+    assertFalse("space is not Emoji", u_hasBinaryProperty(0x20, UCHAR_EMOJI));
+    assertTrue("shooting star is Emoji", u_hasBinaryProperty(0x1F320, UCHAR_EMOJI));
+    IcuTestErrorCode errorCode(*this, "TestEmojiProperties()");
+    UnicodeSet emoji("[:Emoji:]", errorCode);
+    assertTrue("lots of Emoji", emoji.size() > 700);
+
+    assertTrue("shooting star is Emoji_Presentation",
+               u_hasBinaryProperty(0x1F320, UCHAR_EMOJI_PRESENTATION));
+    assertTrue("Fitzpatrick 6 is Emoji_Modifier",
+               u_hasBinaryProperty(0x1F3FF, UCHAR_EMOJI_MODIFIER));
+    assertTrue("happy person is Emoji_Modifier_Base",
+               u_hasBinaryProperty(0x1F64B, UCHAR_EMOJI_MODIFIER_BASE));
+}

icu4c/source/test/intltest/ucdtest.h

@@ -1,6 +1,5 @@
 /********************************************************************
- * COPYRIGHT: 
- * Copyright (c) 1997-2013, International Business Machines Corporation and
+ * Copyright (c) 1997-2016, International Business Machines Corporation and
  * others. All Rights Reserved.
  ********************************************************************/
 
@@ -39,6 +38,7 @@ public:
     void TestPatternProperties();
     void TestScriptMetadata();
     void TestBidiPairedBracketType();
+    void TestEmojiProperties();
 
 private: