am 4d53cb02: Merge "Enable support for SMS national language shift tables." into gingerbread

* commit '4d53cb02168fe35104d7e644dd9b3efd8ca4c91b':
  Enable support for SMS national language shift tables.

3 files changed, 662 insertions, 44 deletions

diff --git a/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmAlphabetTest.java b/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmAlphabetTest.java
index 3a9c511..e83a822 100644
--- a/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmAlphabetTest.java
+++ b/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmAlphabetTest.java
@@ -20,7 +20,6 @@ import junit.framework.TestCase;
 
 import android.test.suitebuilder.annotation.LargeTest;
 import android.test.suitebuilder.annotation.SmallTest;
-import android.test.suitebuilder.annotation.Suppress;
 
 public class GsmAlphabetTest extends TestCase {
 
@@ -38,20 +37,21 @@ public class GsmAlphabetTest extends TestCase {
 
         String message = "aaaaaaaaaabbbbbbbbbbcccccccccc";
         byte[] userData = GsmAlphabet.stringToGsm7BitPackedWithHeader(message,
-                SmsHeader.toByteArray(header));
-        int septetCount = GsmAlphabet.countGsmSeptets(message, false);
+                SmsHeader.toByteArray(header), 0, 0);
+        int septetCount = GsmAlphabet.countGsmSeptetsUsingTables(message, true, 0, 0);
         String parsedMessage = GsmAlphabet.gsm7BitPackedToString(
-                userData, SmsHeader.toByteArray(header).length+2, septetCount, 1);
+                userData, SmsHeader.toByteArray(header).length+2, septetCount, 1, 0, 0);
         assertEquals(message, parsedMessage);
     }
 
     // TODO: This method should *really* be a series of individual test methods.
-    @LargeTest
+    // However, it's a SmallTest because it executes quickly.
+    @SmallTest
     public void testBasic() throws Exception {
         // '@' maps to char 0
         assertEquals(0, GsmAlphabet.charToGsm('@'));
 
-        // `a (a with grave accent) maps to last GSM charater
+        // `a (a with grave accent) maps to last GSM character
         assertEquals(0x7f, GsmAlphabet.charToGsm('\u00e0'));
 
         //
@@ -97,7 +97,7 @@ public class GsmAlphabetTest extends TestCase {
 
         assertEquals('@', GsmAlphabet.gsmToChar(0));
 
-        // `a (a with grave accent) maps to last GSM charater
+        // `a (a with grave accent) maps to last GSM character
         assertEquals('\u00e0', GsmAlphabet.gsmToChar(0x7f));
 
         assertEquals('\uffff',
@@ -116,8 +116,12 @@ public class GsmAlphabetTest extends TestCase {
         assertEquals(' ', GsmAlphabet.gsmExtendedToChar(
                 GsmAlphabet.GSM_EXTENDED_ESCAPE));
 
-        // Unmappable
-        assertEquals(' ', GsmAlphabet.gsmExtendedToChar(0));
+        // Reserved for extension to extension table (mapped to space)
+        assertEquals(' ', GsmAlphabet.gsmExtendedToChar(GsmAlphabet.GSM_EXTENDED_ESCAPE));
+
+        // Unmappable (mapped to character in default or national locking shift table)
+        assertEquals('@', GsmAlphabet.gsmExtendedToChar(0));
+        assertEquals('\u00e0', GsmAlphabet.gsmExtendedToChar(0x7f));
 
         //
         // stringTo7BitPacked, gsm7BitPackedToString
@@ -128,7 +132,7 @@ public class GsmAlphabetTest extends TestCase {
 
         // Check all alignment cases
         for (int i = 0; i < 9; i++, testString.append('@')) {
-            packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString());
+            packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString(), 0, 0);
             assertEquals(testString.toString(),
                     GsmAlphabet.gsm7BitPackedToString(packed, 1, 0xff & packed[0]));
         }
@@ -149,7 +153,7 @@ public class GsmAlphabetTest extends TestCase {
             assertEquals(1, GsmAlphabet.countGsmSeptets(c));
         }
 
-        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString());
+        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString(), 0, 0);
         assertEquals(testString.toString(),
                 GsmAlphabet.gsm7BitPackedToString(packed, 1, 0xff & packed[0]));
 
@@ -164,7 +168,7 @@ public class GsmAlphabetTest extends TestCase {
 
         }
 
-        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString());
+        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString(), 0, 0);
         assertEquals(testString.toString(),
                 GsmAlphabet.gsm7BitPackedToString(packed, 1, 0xff & packed[0]));
 
@@ -175,7 +179,7 @@ public class GsmAlphabetTest extends TestCase {
             testString.append('@');
         }
 
-        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString());
+        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString(), 0, 0);
         assertEquals(testString.toString(),
                 GsmAlphabet.gsm7BitPackedToString(packed, 1, 0xff & packed[0]));
 
@@ -183,7 +187,7 @@ public class GsmAlphabetTest extends TestCase {
         testString.append('@');
 
         try {
-            GsmAlphabet.stringToGsm7BitPacked(testString.toString());
+            GsmAlphabet.stringToGsm7BitPacked(testString.toString(), 0, 0);
             fail("expected exception");
         } catch (EncodeException ex) {
             // exception expected
@@ -196,7 +200,7 @@ public class GsmAlphabetTest extends TestCase {
             testString.append('{');
         }
 
-        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString());
+        packed = GsmAlphabet.stringToGsm7BitPacked(testString.toString(), 0, 0);
         assertEquals(testString.toString(),
                 GsmAlphabet.gsm7BitPackedToString(packed, 1, 0xff & packed[0]));
 
@@ -204,17 +208,29 @@ public class GsmAlphabetTest extends TestCase {
         testString.append('{');
 
         try {
-            GsmAlphabet.stringToGsm7BitPacked(testString.toString());
+            GsmAlphabet.stringToGsm7BitPacked(testString.toString(), 0, 0);
             fail("expected exception");
         } catch (EncodeException ex) {
             // exception expected
         }
 
+        // Reserved for extension to extension table (mapped to space)
+        packed = new byte[]{(byte)(0x1b | 0x80), 0x1b >> 1};
+        assertEquals(" ", GsmAlphabet.gsm7BitPackedToString(packed, 0, 2));
+
+        // Unmappable (mapped to character in default alphabet table)
+        packed[0] = 0x1b;
+        packed[1] = 0x00;
+        assertEquals("@", GsmAlphabet.gsm7BitPackedToString(packed, 0, 2));
+        packed[0] = (byte)(0x1b | 0x80);
+        packed[1] = (byte)(0x7f >> 1);
+        assertEquals("\u00e0", GsmAlphabet.gsm7BitPackedToString(packed, 0, 2));
+
         //
         // 8 bit unpacked format
         //
         // Note: we compare hex strings here
-        // because Assert doesnt have array-comparisons
+        // because Assert doesn't have array comparisons
 
         byte unpacked[];
 
@@ -306,5 +322,16 @@ public class GsmAlphabetTest extends TestCase {
 
         assertEquals("a",
                 GsmAlphabet.gsm8BitUnpackedToString(unpacked, 1, unpacked.length - 1));
+
+        // Reserved for extension to extension table (mapped to space)
+        unpacked[0] = 0x1b;
+        unpacked[1] = 0x1b;
+        assertEquals(" ", GsmAlphabet.gsm8BitUnpackedToString(unpacked, 0, 2));
+
+        // Unmappable (mapped to character in default or national locking shift table)
+        unpacked[1] = 0x00;
+        assertEquals("@", GsmAlphabet.gsm8BitUnpackedToString(unpacked, 0, 2));
+        unpacked[1] = 0x7f;
+        assertEquals("\u00e0", GsmAlphabet.gsm8BitUnpackedToString(unpacked, 0, 2));
     }
 }
diff --git a/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmSmsTest.java b/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmSmsTest.java
index 3103fc1..41a719e 100644
--- a/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmSmsTest.java
+++ b/telephony/tests/telephonytests/src/com/android/internal/telephony/GsmSmsTest.java
@@ -16,16 +16,12 @@
 
 package com.android.internal.telephony;
 
-import com.android.internal.telephony.GsmAlphabet;
-import com.android.internal.telephony.SmsHeader;
 import com.android.internal.telephony.gsm.SmsMessage;
 import com.android.internal.util.HexDump;
 
 import android.test.AndroidTestCase;
 import android.test.suitebuilder.annotation.SmallTest;
 
-import android.util.Log;
-
 public class GsmSmsTest extends AndroidTestCase {
 
     @SmallTest
@@ -211,8 +207,38 @@ public class GsmSmsTest extends AndroidTestCase {
                 sms.getMessageBody());
     }
 
+    // GSM 7 bit tables in String form, Escape (0x1B) replaced with '@'
+    private static final String[] sBasicTables = {
+        // GSM 7 bit default alphabet
+        "@\u00a3$\u00a5\u00e8\u00e9\u00f9\u00ec\u00f2\u00c7\n\u00d8\u00f8\r\u00c5\u00e5\u0394_"
+            + "\u03a6\u0393\u039b\u03a9\u03a0\u03a8\u03a3\u0398\u039e@\u00c6\u00e6\u00df\u00c9"
+            + " !\"#\u00a4%&'()*+,-./0123456789:;<=>?\u00a1ABCDEFGHIJKLMNOPQRSTUVWXYZ\u00c4\u00d6"
+            + "\u00d1\u00dc\u00a7\u00bfabcdefghijklmnopqrstuvwxyz\u00e4\u00f6\u00f1\u00fc\u00e0",
+
+        // Turkish locking shift table
+        "@\u00a3$\u00a5\u20ac\u00e9\u00f9\u0131\u00f2\u00c7\n\u011e\u011f\r\u00c5\u00e5\u0394_"
+            + "\u03a6\u0393\u039b\u03a9\u03a0\u03a8\u03a3\u0398\u039e@\u015e\u015f\u00df\u00c9"
+            + " !\"#\u00a4%&'()*+,-./0123456789:;<=>?\u0130ABCDEFGHIJKLMNOPQRSTUVWXYZ\u00c4\u00d6"
+            + "\u00d1\u00dc\u00a7\u00e7abcdefghijklmnopqrstuvwxyz\u00e4\u00f6\u00f1\u00fc\u00e0",
+
+        // no locking shift table defined for Spanish
+        "",
+
+        // Portuguese locking shift table
+        "@\u00a3$\u00a5\u00ea\u00e9\u00fa\u00ed\u00f3\u00e7\n\u00d4\u00f4\r\u00c1\u00e1\u0394_"
+            + "\u00aa\u00c7\u00c0\u221e^\\\u20ac\u00d3|@\u00c2\u00e2\u00ca\u00c9 !\"#\u00ba%&'()"
+            + "*+,-./0123456789:;<=>?\u00cdABCDEFGHIJKLMNOPQRSTUVWXYZ\u00c3\u00d5\u00da\u00dc"
+            + "\u00a7~abcdefghijklmnopqrstuvwxyz\u00e3\u00f5`\u00fc\u00e0"
+    };
+
     @SmallTest
     public void testDecode() throws Exception {
+        decodeSingle(0);    // default table
+        decodeSingle(1);    // Turkish locking shift table
+        decodeSingle(3);    // Portuguese locking shift table
+    }
+
+    private void decodeSingle(int language) throws Exception {
         byte[] septets = new byte[(7 * 128 + 7) / 8];
 
         int bitOffset = 0;
@@ -238,15 +264,168 @@ public class GsmSmsTest extends AndroidTestCase {
             bitOffset += 7;
         }
 
-        String decoded = GsmAlphabet.gsm7BitPackedToString(septets, 0, 128);
-        byte[] reEncoded = GsmAlphabet.stringToGsm7BitPacked(decoded);
+        String decoded = GsmAlphabet.gsm7BitPackedToString(septets, 0, 128, 0, language, 0);
+        byte[] reEncoded = GsmAlphabet.stringToGsm7BitPacked(decoded, language, 0);
+
+        assertEquals(sBasicTables[language], decoded);
 
         // reEncoded has the count septets byte at the front
-        assertEquals(reEncoded.length, septets.length + 1);
+        assertEquals(septets.length + 1, reEncoded.length);
 
         for (int i = 0; i < septets.length; i++) {
-            assertEquals(reEncoded[i + 1], septets[i]);
+            assertEquals(septets[i], reEncoded[i + 1]);
+        }
+    }
+
+    private static final int GSM_ESCAPE_CHARACTER = 0x1b;
+
+    private static final String[] sExtendedTables = {
+        // GSM 7 bit default alphabet extension table
+        "\f^{}\\[~]|\u20ac",
+
+        // Turkish single shift extension table
+        "\f^{}\\[~]|\u011e\u0130\u015e\u00e7\u20ac\u011f\u0131\u015f",
+
+        // Spanish single shift extension table
+        "\u00e7\f^{}\\[~]|\u00c1\u00cd\u00d3\u00da\u00e1\u20ac\u00ed\u00f3\u00fa",
+
+        // Portuguese single shift extension table
+        "\u00ea\u00e7\f\u00d4\u00f4\u00c1\u00e1\u03a6\u0393^\u03a9\u03a0\u03a8\u03a3\u0398\u00ca"
+            + "{}\\[~]|\u00c0\u00cd\u00d3\u00da\u00c3\u00d5\u00c2\u20ac\u00ed\u00f3\u00fa\u00e3"
+            + "\u00f5\u00e2"
+    };
+
+    private static final int[][] sExtendedTableIndexes = {
+        {0x0a, 0x14, 0x28, 0x29, 0x2f, 0x3c, 0x3d, 0x3e, 0x40, 0x65},
+        {0x0a, 0x14, 0x28, 0x29, 0x2f, 0x3c, 0x3d, 0x3e, 0x40, 0x47, 0x49, 0x53, 0x63,
+                0x65, 0x67, 0x69, 0x73},
+        {0x09, 0x0a, 0x14, 0x28, 0x29, 0x2f, 0x3c, 0x3d, 0x3e, 0x40, 0x41, 0x49, 0x4f,
+                0x55, 0x61, 0x65, 0x69, 0x6f, 0x75},
+        {0x05, 0x09, 0x0a, 0x0b, 0x0c, 0x0e, 0x0f, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
+                0x18, 0x19, 0x1f, 0x28, 0x29, 0x2f, 0x3c, 0x3d, 0x3e, 0x40, 0x41, 0x49,
+                0x4f, 0x55, 0x5b, 0x5c, 0x61, 0x65, 0x69, 0x6f, 0x75, 0x7b, 0x7c, 0x7f}
+    };
+
+    @SmallTest
+    public void testDecodeExtended() throws Exception {
+        for (int language = 0; language < 3; language++) {
+            int[] tableIndex = sExtendedTableIndexes[language];
+            int numSeptets = tableIndex.length * 2;  // two septets per extended char
+            byte[] septets = new byte[(7 * numSeptets + 7) / 8];
+
+            int bitOffset = 0;
+
+            for (int v : tableIndex) {
+                // escape character
+                int byteOffset = bitOffset / 8;
+                int shift = bitOffset % 8;
+
+                septets[byteOffset] |= GSM_ESCAPE_CHARACTER << shift;
+
+                if (shift > 1) {
+                    septets[byteOffset + 1] = (byte) (GSM_ESCAPE_CHARACTER >> (8 - shift));
+                }
+
+                bitOffset += 7;
+
+                // extended table index
+                byteOffset = bitOffset / 8;
+                shift = bitOffset % 8;
+
+                septets[byteOffset] |= v << shift;
+
+                if (shift > 1) {
+                    septets[byteOffset + 1] = (byte) (v >> (8 - shift));
+                }
+
+                bitOffset += 7;
+            }
+
+            String decoded = GsmAlphabet.gsm7BitPackedToString(septets, 0, numSeptets, 0,
+                    0, language);
+            byte[] reEncoded = GsmAlphabet.stringToGsm7BitPacked(decoded, 0, language);
+
+            assertEquals(sExtendedTables[language], decoded);
+
+            // reEncoded has the count septets byte at the front
+            assertEquals(septets.length + 1, reEncoded.length);
+
+            for (int i = 0; i < septets.length; i++) {
+                assertEquals(septets[i], reEncoded[i + 1]);
+            }
         }
     }
 
+    @SmallTest
+    public void testDecodeExtendedFallback() throws Exception {
+        // verify that unmapped characters in extension table fall back to locking shift table
+        for (int language = 0; language < 3; language++) {
+            int[] tableIndex = sExtendedTableIndexes[language];
+            int numChars = 128 - tableIndex.length;
+            int numSeptets = numChars * 2;  // two septets per extended char
+            byte[] septets = new byte[(7 * numSeptets + 7) / 8];
+
+            int tableOffset = 0;
+            int bitOffset = 0;
+
+            StringBuilder defaultTable = new StringBuilder(128);
+            StringBuilder turkishTable = new StringBuilder(128);
+            StringBuilder portugueseTable = new StringBuilder(128);
+
+            for (char c = 0; c < 128; c++) {
+                // skip characters that are present in the current extension table
+                if (tableOffset < tableIndex.length && tableIndex[tableOffset] == c) {
+                    tableOffset++;
+                    continue;
+                }
+
+                // escape character
+                int byteOffset = bitOffset / 8;
+                int shift = bitOffset % 8;
+
+                septets[byteOffset] |= GSM_ESCAPE_CHARACTER << shift;
+
+                if (shift > 1) {
+                    septets[byteOffset + 1] = (byte) (GSM_ESCAPE_CHARACTER >> (8 - shift));
+                }
+
+                bitOffset += 7;
+
+                // extended table index
+                byteOffset = bitOffset / 8;
+                shift = bitOffset % 8;
+
+                septets[byteOffset] |= c << shift;
+
+                if (shift > 1) {
+                    septets[byteOffset + 1] = (byte) (c >> (8 - shift));
+                }
+
+                bitOffset += 7;
+
+                if (c == GsmAlphabet.GSM_EXTENDED_ESCAPE) {
+                    // double Escape maps to space character
+                    defaultTable.append(' ');
+                    turkishTable.append(' ');
+                    portugueseTable.append(' ');
+                } else {
+                    // other unmapped chars map to the default or locking shift table
+                    defaultTable.append(sBasicTables[0].charAt(c));
+                    turkishTable.append(sBasicTables[1].charAt(c));
+                    portugueseTable.append(sBasicTables[3].charAt(c));
+                }
+            }
+
+            String decoded = GsmAlphabet.gsm7BitPackedToString(septets, 0, numSeptets, 0,
+                    0, language);
+
+            assertEquals(defaultTable.toString(), decoded);
+
+            decoded = GsmAlphabet.gsm7BitPackedToString(septets, 0, numSeptets, 0, 1, language);
+            assertEquals(turkishTable.toString(), decoded);
+
+            decoded = GsmAlphabet.gsm7BitPackedToString(septets, 0, numSeptets, 0, 3, language);
+            assertEquals(portugueseTable.toString(), decoded);
+        }
+    }
 }
diff --git a/telephony/tests/telephonytests/src/com/android/internal/telephony/SmsMessageBodyTest.java b/telephony/tests/telephonytests/src/com/android/internal/telephony/SmsMessageBodyTest.java
index b214887..170bd9b 100644
--- a/telephony/tests/telephonytests/src/com/android/internal/telephony/SmsMessageBodyTest.java
+++ b/telephony/tests/telephonytests/src/com/android/internal/telephony/SmsMessageBodyTest.java
@@ -19,21 +19,57 @@ package com.android.internal.telephony;
 import android.telephony.SmsMessage;
 import android.telephony.TelephonyManager;
 import android.test.AndroidTestCase;
+import android.test.suitebuilder.annotation.LargeTest;
+import android.test.suitebuilder.annotation.MediumTest;
 import android.test.suitebuilder.annotation.SmallTest;
+import android.util.Log;
 
+import java.util.Random;
+
+import static android.telephony.SmsMessage.MAX_USER_DATA_BYTES;
+import static android.telephony.SmsMessage.MAX_USER_DATA_BYTES_WITH_HEADER;
 import static android.telephony.SmsMessage.MAX_USER_DATA_SEPTETS;
+import static android.telephony.SmsMessage.MAX_USER_DATA_SEPTETS_WITH_HEADER;
 
+/**
+ * Test cases to verify selection of the optimal 7 bit encoding tables
+ * (for all combinations of enabled national language tables) for messages
+ * containing Turkish, Spanish, Portuguese, Greek, and other symbols
+ * present in the GSM default and national language tables defined in
+ * 3GPP TS 23.038. Also verifies correct SMS encoding for CDMA, which only
+ * supports the GSM 7 bit default alphabet, ASCII 8 bit, and UCS-2.
+ * Tests both encoding variations: unsupported characters mapped to space,
+ * and unsupported characters force entire message to UCS-2.
+ */
 public class SmsMessageBodyTest extends AndroidTestCase {
+    private static final String TAG = "SmsMessageBodyTest";
 
+    // ASCII chars in the GSM 7 bit default alphabet
     private static final String sAsciiChars = "@$_ !\"#%&'()*+,-./0123456789" +
             ":;<=>?ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\n\r";
-    private static final String sGsmBasicChars = "\u00a3\u00a5\u00e8\u00e9" +
-            "\u00f9\u00ec\u00f2\u00c7\u00d8\u00f8\u00c5\u00e5\u0394\u03a6" +
-            "\u0393\u039b\u03a9\u03a0\u03a8\u03a3\u0398\u00c6\u00e6" +
-            "\u00df\u00c9\u00a4\u00a1\u00c4\u00d6\u00d1\u00dc\u00a7\u00bf" +
-            "\u00e4\u00f6\u00f1\u00fc\u00e0";
-    private static final String sGsmExtendedAsciiChars = "{|}\\[~]^\f";
+
+    // Unicode chars in the GSM 7 bit default alphabet and both locking shift tables
+    private static final String sGsmDefaultChars = "\u00a3\u00a5\u00e9\u00c7\u0394\u00c9" +
+            "\u00dc\u00a7\u00fc\u00e0";
+
+    // Unicode chars in the GSM 7 bit default table and Turkish locking shift tables
+    private static final String sGsmDefaultAndTurkishTables = "\u00f9\u00f2\u00c5\u00e5\u00df" +
+            "\u00a4\u00c4\u00d6\u00d1\u00e4\u00f6\u00f1";
+
+    // Unicode chars in the GSM 7 bit default table but not the locking shift tables
+    private static final String sGsmDefaultTableOnly = "\u00e8\u00ec\u00d8\u00f8\u00c6\u00e6" +
+            "\u00a1\u00bf";
+
+    // ASCII chars in the GSM default extension table
+    private static final String sGsmExtendedAsciiChars = "{}[]\f";
+
+    // chars in GSM default extension table and Portuguese locking shift table
+    private static final String sGsmExtendedPortugueseLocking = "^\\|~";
+
+    // Euro currency symbol
     private static final String sGsmExtendedEuroSymbol = "\u20ac";
+
+    // CJK ideographs, Hiragana, Katakana, full width letters, Cyrillic, etc.
     private static final String sUnicodeChars = "\u4e00\u4e01\u4e02\u4e03" +
             "\u4e04\u4e05\u4e06\u4e07\u4e08\u4e09\u4e0a\u4e0b\u4e0c\u4e0d" +
             "\u4e0e\u4e0f\u3041\u3042\u3043\u3044\u3045\u3046\u3047\u3048" +
@@ -43,6 +79,86 @@ public class SmsMessageBodyTest extends AndroidTestCase {
             "\u0400\u0401\u0402\u0403\u0404\u0405\u0406\u0407\u0408" +
             "\u00a2\u00a9\u00ae\u2122";
 
+    // chars in Turkish single shift and locking shift tables
+    private static final String sTurkishChars = "\u0131\u011e\u011f\u015e\u015f\u0130";
+
+    // chars in Spanish single shift table and Portuguese single and locking shift tables
+    private static final String sPortugueseAndSpanishChars = "\u00c1\u00e1\u00cd\u00ed"
+            + "\u00d3\u00f3\u00da\u00fa";
+
+    // chars in all national language tables but not in the standard GSM alphabets
+    private static final String sNationalLanguageTablesOnly = "\u00e7";
+
+    // chars in Portuguese single shift and locking shift tables
+    private static final String sPortugueseChars = "\u00ea\u00d4\u00f4\u00c0\u00c2\u00e2"
+            + "\u00ca\u00c3\u00d5\u00e3\u00f5";
+
+    // chars in Portuguese locking shift table only
+    private static final String sPortugueseLockingShiftChars = "\u00aa\u221e\u00ba`";
+
+    // Greek letters in GSM alphabet missing from Portuguese locking and single shift tables
+    private static final String sGreekLettersNotInPortugueseTables = "\u039b\u039e";
+
+    // Greek letters in GSM alphabet and Portuguese single shift (but not locking shift) table
+    private static final String sGreekLettersInPortugueseShiftTable =
+            "\u03a6\u0393\u03a9\u03a0\u03a8\u03a3\u0398";
+
+    // List of classes of characters in SMS tables
+    private static final String[] sCharacterClasses = {
+            sGsmExtendedAsciiChars,
+            sGsmExtendedPortugueseLocking,
+            sGsmDefaultChars,
+            sGsmDefaultAndTurkishTables,
+            sGsmDefaultTableOnly,
+            sGsmExtendedEuroSymbol,
+            sUnicodeChars,
+            sTurkishChars,
+            sPortugueseChars,
+            sPortugueseLockingShiftChars,
+            sPortugueseAndSpanishChars,
+            sGreekLettersNotInPortugueseTables,
+            sGreekLettersInPortugueseShiftTable,
+            sNationalLanguageTablesOnly,
+            sAsciiChars
+    };
+
+    private static final int sNumCharacterClasses = sCharacterClasses.length;
+
+    // For each character class, whether it is present in a particular char table.
+    // First three entries are locking shift tables, followed by four single shift tables
+    private static final boolean[][] sCharClassPresenceInTables = {
+            // ASCII chars in all GSM extension tables
+            {false, false, false, true, true, true, true},
+            // ASCII chars in all GSM extension tables and Portuguese locking shift table
+            {false, false, true, true, true, true, true},
+            // non-ASCII chars in GSM default alphabet and all locking tables
+            {true, true, true, false, false, false, false},
+            // non-ASCII chars in GSM default alphabet and Turkish locking shift table
+            {true, true, false, false, false, false, false},
+            // non-ASCII chars in GSM default alphabet table only
+            {true, false, false, false, false, false, false},
+            // Euro symbol is present in several tables
+            {false, true, true, true, true, true, true},
+            // Unicode characters not present in any 7 bit tables
+            {false, false, false, false, false, false, false},
+            // Characters specific to Turkish language
+            {false, true, false, false, true, false, false},
+            // Characters in Portuguese single shift and locking shift tables
+            {false, false, true, false, false, false, true},
+            // Characters in Portuguese locking shift table only
+            {false, false, true, false, false, false, false},
+            // Chars in Spanish single shift and Portuguese single and locking shift tables
+            {false, false, true, false, false, true, true},
+            // Greek letters in GSM default alphabet missing from Portuguese tables
+            {true, true, false, false, false, false, false},
+            // Greek letters in GSM alphabet and Portuguese single shift table
+            {true, true, false, false, false, false, true},
+            // Chars in all national language tables but not the standard GSM tables
+            {false, true, true, false, true, true, true},
+            // ASCII chars in GSM default alphabet
+            {true, true, true, false, false, false, false}
+    };
+
     private static final int sTestLengthCount = 12;
 
     private static final int[] sSeptetTestLengths =
@@ -60,11 +176,92 @@ public class SmsMessageBodyTest extends AndroidTestCase {
     private static final int[] sUnicodeUnitsRemaining =
             { 70,  69,  68, 35,   1,   0,  63,  34,   1,   0,  66,  41};
 
+    // Combinations of enabled GSM national language single shift tables
+    private static final int[][] sEnabledSingleShiftTables = {
+            {},         // GSM default alphabet only
+            {1},        // Turkish (single shift only)
+            {1},        // Turkish (single and locking shift)
+            {2},        // Spanish
+            {3},        // Portuguese (single shift only)
+            {3},        // Portuguese (single and locking shift)
+            {1, 2},     // Turkish + Spanish (single shift only)
+            {1, 2},     // Turkish + Spanish (single and locking shift)
+            {1, 3},     // Turkish + Portuguese (single shift only)
+            {1, 3},     // Turkish + Portuguese (single and locking shift)
+            {2, 3},     // Spanish + Portuguese (single shift only)
+            {2, 3},     // Spanish + Portuguese (single and locking shift)
+            {1, 2, 3},  // Turkish, Spanish, Portuguese (single shift only)
+            {1, 2, 3},  // Turkish, Spanish, Portuguese (single and locking shift)
+            {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13} // all language tables
+    };
+
+    // Combinations of enabled GSM national language locking shift tables
+    private static final int[][] sEnabledLockingShiftTables = {
+            {},         // GSM default alphabet only
+            {},         // Turkish (single shift only)
+            {1},        // Turkish (single and locking shift)
+            {},         // Spanish (no locking shift table)
+            {},         // Portuguese (single shift only)
+            {3},        // Portuguese (single and locking shift)
+            {},         // Turkish + Spanish (single shift only)
+            {1},        // Turkish + Spanish (single and locking shift)
+            {},         // Turkish + Portuguese (single shift only)
+            {1, 3},     // Turkish + Portuguese (single and locking shift)
+            {},         // Spanish + Portuguese (single shift only)
+            {3},        // Spanish + Portuguese (single and locking shift)
+            {},         // Turkish, Spanish, Portuguese (single shift only)
+            {1, 3},     // Turkish, Spanish, Portuguese (single and locking shift)
+            {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13} // all language tables
+    };
+
+    // LanguagePair counter indexes to check for each entry above
+    private static final int[][] sLanguagePairIndexesByEnabledIndex = {
+            {0},                            // default tables only
+            {0, 1},                         // Turkish (single shift only)
+            {0, 1, 4, 5},                   // Turkish (single and locking shift)
+            {0, 2},                         // Spanish
+            {0, 3},                         // Portuguese (single shift only)
+            {0, 3, 8, 11},                  // Portuguese (single and locking shift)
+            {0, 1, 2},                      // Turkish + Spanish (single shift only)
+            {0, 1, 2, 4, 5, 6},             // Turkish + Spanish (single and locking shift)
+            {0, 1, 3},                      // Turkish + Portuguese (single shift only)
+            {0, 1, 3, 4, 5, 7, 8, 9, 11},   // Turkish + Portuguese (single and locking shift)
+            {0, 2, 3},                      // Spanish + Portuguese (single shift only)
+            {0, 2, 3, 8, 10, 11},           // Spanish + Portuguese (single and locking shift)
+            {0, 1, 2, 3},                   // all languages (single shift only)
+            {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}, // all languages (single and locking shift)
+            {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}  // all languages (no Indic chars in test)
+    };
+
+    /**
+     * User data header requires one octet for length. Count as one septet, because
+     * all combinations of header elements below will have at least one free bit
+     * when padding to the nearest septet boundary.
+     */
+    private static final int UDH_SEPTET_COST_LENGTH = 1;
+
+    /**
+     * Using a non-default language locking shift table OR single shift table
+     * requires a user data header of 3 octets, or 4 septets, plus UDH length.
+     */
+    private static final int UDH_SEPTET_COST_ONE_SHIFT_TABLE = 4;
+
+    /**
+     * Using a non-default language locking shift table AND single shift table
+     * requires a user data header of 6 octets, or 7 septets, plus UDH length.
+     */
+    private static final int UDH_SEPTET_COST_TWO_SHIFT_TABLES = 7;
+
+    /**
+     * Multi-part messages require a user data header of 5 octets, or 6 septets,
+     * plus UDH length.
+     */
+    private static final int UDH_SEPTET_COST_CONCATENATED_MESSAGE = 6;
 
     @SmallTest
     public void testCalcLengthAscii() throws Exception {
         StringBuilder sb = new StringBuilder(320);
-        int[] values = {0, 0, 0, SmsMessage.ENCODING_7BIT};
+        int[] values = {0, 0, 0, SmsMessage.ENCODING_7BIT, 0, 0};
         int startPos = 0;
         int asciiCharsLen = sAsciiChars.length();
 
@@ -94,20 +291,10 @@ public class SmsMessageBodyTest extends AndroidTestCase {
     }
 
     @SmallTest
-    public void testCalcLength7bitGsm() throws Exception {
-        // TODO
-    }
-
-    @SmallTest
-    public void testCalcLength7bitGsmExtended() throws Exception {
-        // TODO
-    }
-
-    @SmallTest
     public void testCalcLengthUnicode() throws Exception {
         StringBuilder sb = new StringBuilder(160);
-        int[] values = {0, 0, 0, SmsMessage.ENCODING_16BIT};
-        int[] values7bit = {1, 0, 0, SmsMessage.ENCODING_7BIT};
+        int[] values = {0, 0, 0, SmsMessage.ENCODING_16BIT, 0, 0};
+        int[] values7bit = {1, 0, 0, SmsMessage.ENCODING_7BIT, 0, 0};
         int startPos = 0;
         int unicodeCharsLen = sUnicodeChars.length();
 
@@ -139,6 +326,229 @@ public class SmsMessageBodyTest extends AndroidTestCase {
         }
     }
 
+    private static class LanguagePair {
+        // index is 2 for Portuguese locking shift because there is no Spanish locking shift table
+        private final int langTableIndex;
+        private final int langShiftTableIndex;
+        int length;
+        int missingChars7bit;
+
+        LanguagePair(int langTable, int langShiftTable) {
+            langTableIndex = langTable;
+            langShiftTableIndex = langShiftTable;
+        }
+
+        void clear() {
+            length = 0;
+            missingChars7bit = 0;
+        }
+
+        void addChar(boolean[] charClassTableRow) {
+            if (charClassTableRow[langTableIndex]) {
+                length++;
+            } else if (charClassTableRow[3 + langShiftTableIndex]) {
+                length += 2;
+            } else {
+                length++;    // use ' ' for unmapped char in 7 bit only mode
+                missingChars7bit++;
+            }
+        }
+    }
+
+    private static class CounterHelper {
+        LanguagePair[] mCounters;
+        int[] mStatsCounters;
+        int mUnicodeCounter;
+
+        CounterHelper() {
+            mCounters = new LanguagePair[12];
+            mStatsCounters = new int[12];
+            for (int i = 0; i < 12; i++) {
+                mCounters[i] = new LanguagePair(i/4, i%4);
+            }
+        }
+
+        void clear() {
+            // Note: don't clear stats counters
+            for (int i = 0; i < 12; i++) {
+                mCounters[i].clear();
+            }
+        }
+
+        void addChar(int charClass) {
+            boolean[] charClassTableRow = sCharClassPresenceInTables[charClass];
+            for (int i = 0; i < 12; i++) {
+                mCounters[i].addChar(charClassTableRow);
+            }
+        }
+
+        void fillData(int enabledLangsIndex, boolean use7bitOnly, int[] values, int length) {
+            int[] languagePairs = sLanguagePairIndexesByEnabledIndex[enabledLangsIndex];
+            int minNumSeptets = Integer.MAX_VALUE;
+            int minNumSeptetsWithHeader = Integer.MAX_VALUE;
+            int minNumMissingChars = Integer.MAX_VALUE;
+            int langIndex = -1;
+            int langShiftIndex = -1;
+            for (int i : languagePairs) {
+                LanguagePair pair = mCounters[i];
+                int udhLength = 0;
+                if (i != 0) {
+                    udhLength = UDH_SEPTET_COST_LENGTH;
+                    if (i < 4 || i % 4 == 0) {
+                        udhLength += UDH_SEPTET_COST_ONE_SHIFT_TABLE;
+                    } else {
+                        udhLength += UDH_SEPTET_COST_TWO_SHIFT_TABLES;
+                    }
+                }
+                int numSeptetsWithHeader;
+                if (pair.length > (MAX_USER_DATA_SEPTETS - udhLength)) {
+                    if (udhLength == 0) {
+                        udhLength = UDH_SEPTET_COST_LENGTH;
+                    }
+                    udhLength += UDH_SEPTET_COST_CONCATENATED_MESSAGE;
+                    int septetsPerPart = MAX_USER_DATA_SEPTETS - udhLength;
+                    int msgCount = (pair.length + septetsPerPart - 1) / septetsPerPart;
+                    numSeptetsWithHeader = udhLength * msgCount + pair.length;
+                } else {
+                    numSeptetsWithHeader = udhLength + pair.length;
+                }
+
+                if (use7bitOnly) {
+                    if (pair.missingChars7bit < minNumMissingChars || (pair.missingChars7bit ==
+                            minNumMissingChars && numSeptetsWithHeader < minNumSeptetsWithHeader)) {
+                        minNumSeptets = pair.length;
+                        minNumSeptetsWithHeader = numSeptetsWithHeader;
+                        minNumMissingChars = pair.missingChars7bit;
+                        langIndex = pair.langTableIndex;
+                        langShiftIndex = pair.langShiftTableIndex;
+                    }
+                } else {
+                    if (pair.missingChars7bit == 0 && numSeptetsWithHeader < minNumSeptetsWithHeader) {
+                        minNumSeptets = pair.length;
+                        minNumSeptetsWithHeader = numSeptetsWithHeader;
+                        langIndex = pair.langTableIndex;
+                        langShiftIndex = pair.langShiftTableIndex;
+                    }
+                }
+            }
+            if (langIndex == -1) {
+                // nothing matches, use values for Unicode
+                int byteCount = length * 2;
+                if (byteCount > MAX_USER_DATA_BYTES) {
+                    values[0] = (byteCount + MAX_USER_DATA_BYTES_WITH_HEADER - 1) /
+                            MAX_USER_DATA_BYTES_WITH_HEADER;
+                    values[2] = ((values[0] * MAX_USER_DATA_BYTES_WITH_HEADER) - byteCount) / 2;
+                } else {
+                    values[0] = 1;
+                    values[2] = (MAX_USER_DATA_BYTES - byteCount) / 2;
+                }
+                values[1] = length;
+                values[3] = SmsMessage.ENCODING_16BIT;
+                values[4] = 0;
+                values[5] = 0;
+                mUnicodeCounter++;
+            } else {
+                int udhLength = 0;
+                if (langIndex != 0 || langShiftIndex != 0) {
+                    udhLength = UDH_SEPTET_COST_LENGTH;
+                    if (langIndex == 0 || langShiftIndex == 0) {
+                        udhLength += UDH_SEPTET_COST_ONE_SHIFT_TABLE;
+                    } else {
+                        udhLength += UDH_SEPTET_COST_TWO_SHIFT_TABLES;
+                    }
+                }
+                int msgCount;
+                if (minNumSeptets > (MAX_USER_DATA_SEPTETS - udhLength)) {
+                    if (udhLength == 0) {
+                        udhLength = UDH_SEPTET_COST_LENGTH;
+                    }
+                    udhLength += UDH_SEPTET_COST_CONCATENATED_MESSAGE;
+                    int septetsPerPart = MAX_USER_DATA_SEPTETS - udhLength;
+                    msgCount = (minNumSeptets + septetsPerPart - 1) / septetsPerPart;
+                } else {
+                    msgCount = 1;
+                }
+                values[0] = msgCount;
+                values[1] = minNumSeptets;
+                values[2] = (values[0] * (MAX_USER_DATA_SEPTETS - udhLength)) - minNumSeptets;
+                values[3] = SmsMessage.ENCODING_7BIT;
+                values[4] = (langIndex == 2 ? 3 : langIndex); // Portuguese is code 3, index 2
+                values[5] = langShiftIndex;
+                assertEquals("minNumSeptetsWithHeader", minNumSeptetsWithHeader,
+                        udhLength * msgCount + minNumSeptets);
+                mStatsCounters[langIndex * 4 + langShiftIndex]++;
+            }
+        }
+
+        void printStats() {
+            Log.d(TAG, "Unicode selection count: " + mUnicodeCounter);
+            for (int i = 0; i < 12; i++) {
+                Log.d(TAG, "Language pair index " + i + " count: " + mStatsCounters[i]);
+            }
+        }
+    }
+
+    @LargeTest
+    public void testCalcLengthMixed7bit() throws Exception {
+        StringBuilder sb = new StringBuilder(320);
+        CounterHelper ch = new CounterHelper();
+        Random r = new Random(0x4321);  // use the same seed for reproducibility
+        int[] expectedValues = new int[6];
+        int[] origLockingShiftTables = GsmAlphabet.getEnabledLockingShiftTables();
+        int[] origSingleShiftTables = GsmAlphabet.getEnabledSingleShiftTables();
+        int enabledLanguagesTestCases = sEnabledSingleShiftTables.length;
+        long startTime = System.currentTimeMillis();
+
+        // Repeat for 10 test runs
+        for (int run = 0; run < 10; run++) {
+            sb.setLength(0);
+            ch.clear();
+            int unicodeOnlyCount = 0;
+
+            // Test incrementally from 1 to 320 character random messages
+            for (int i = 1; i < 320; i++) {
+                // 1% chance to add from each special character class, else add an ASCII char
+                int charClass = r.nextInt(100);
+                if (charClass >= sNumCharacterClasses) {
+                    charClass = sNumCharacterClasses - 1;   // last class is ASCII
+                }
+                int classLength = sCharacterClasses[charClass].length();
+                char nextChar = sCharacterClasses[charClass].charAt(r.nextInt(classLength));
+                sb.append(nextChar);
+                ch.addChar(charClass);
+
+//                if (i % 20 == 0) {
+//                    Log.d(TAG, "test string: " + sb);
+//                }
+
+                // Test string against all combinations of enabled languages
+                boolean unicodeOnly = true;
+                for (int j = 0; j < enabledLanguagesTestCases; j++) {
+                    GsmAlphabet.setEnabledSingleShiftTables(sEnabledSingleShiftTables[j]);
+                    GsmAlphabet.setEnabledLockingShiftTables(sEnabledLockingShiftTables[j]);
+                    ch.fillData(j, false, expectedValues, i);
+                    if (expectedValues[3] == SmsMessage.ENCODING_7BIT) {
+                        unicodeOnly = false;
+                    }
+                    callGsmLengthMethods(sb, false, expectedValues);
+                    // test 7 bit only mode
+                    ch.fillData(j, true, expectedValues, i);
+                    callGsmLengthMethods(sb, true, expectedValues);
+                }
+                // after 10 iterations with a Unicode-only string, skip to next test string
+                // so we can spend more time testing strings that do encode into 7 bits.
+                if (unicodeOnly && ++unicodeOnlyCount == 10) {
+//                    Log.d(TAG, "Unicode only: skipping to next test string");
+                    break;
+                }
+            }
+        }
+        ch.printStats();
+        Log.d(TAG, "Completed in " + (System.currentTimeMillis() - startTime) + " ms");
+        GsmAlphabet.setEnabledLockingShiftTables(origLockingShiftTables);
+        GsmAlphabet.setEnabledSingleShiftTables(origSingleShiftTables);
+    }
+
     private void callGsmLengthMethods(CharSequence msgBody, boolean use7bitOnly,
             int[] expectedValues)
     {
@@ -164,6 +574,8 @@ public class SmsMessageBodyTest extends AndroidTestCase {
         assertEquals("codeUnitCount",      expectedValues[1], ted.codeUnitCount);
         assertEquals("codeUnitsRemaining", expectedValues[2], ted.codeUnitsRemaining);
         assertEquals("codeUnitSize",       expectedValues[3], ted.codeUnitSize);
+        assertEquals("languageTable",      expectedValues[4], ted.languageTable);
+        assertEquals("languageShiftTable", expectedValues[5], ted.languageShiftTable);
     }
 
     private void callCdmaLengthMethods(CharSequence msgBody, boolean use7bitOnly,