Discussion:
[PATCH] sms: add regional support for 2 languages
Nandini Rebello
2018-09-28 08:51:31 UTC
Permalink
Adding support for Bengali and Gujrati for GSM 7bit SMS.
More languages to follow on similar style soon.
---
doc/messagemanager-api.txt | 2 +
src/sms.c | 8 +
src/smsutil.c | 4 +-
src/smsutil.h | 2 +
src/util.c | 454 ++++++++++++++++++++++++++++++++++++++++++++-
src/util.h | 2 +
6 files changed, 468 insertions(+), 4 deletions(-)

diff --git a/doc/messagemanager-api.txt b/doc/messagemanager-api.txt
index 43c4d07..8d85a1b 100644
--- a/doc/messagemanager-api.txt
+++ b/doc/messagemanager-api.txt
@@ -104,6 +104,8 @@ Properties string ServiceCenterAddress
"turkish" - Turkish alphabet
"spanish" - Spanish alphabet
"portuguese" - Portuguese alphabet
+ "bengali" - Bengali alphabet
+ "gujarati" - Gujarati alphabet

The standard, language-specific alphabets are defined
in 3GPP TS23.038, Annex A. By default, oFono uses
diff --git a/src/sms.c b/src/sms.c
index b86158e..a7c005e 100644
--- a/src/sms.c
+++ b/src/sms.c
@@ -170,6 +170,10 @@ static const char *sms_alphabet_to_string(enum sms_alphabet alphabet)
return "spanish";
case SMS_ALPHABET_PORTUGUESE:
return "portuguese";
+ case SMS_ALPHABET_BENGALI:
+ return "bengali";
+ case SMS_ALPHABET_GUJARATI:
+ return "gujarati";
case SMS_ALPHABET_DEFAULT:
return "default";
}
@@ -188,6 +192,10 @@ static gboolean sms_alphabet_from_string(const char *str,
*alphabet = SMS_ALPHABET_SPANISH;
else if (g_str_equal(str, "portuguese"))
*alphabet = SMS_ALPHABET_PORTUGUESE;
+ else if (g_str_equal(str, "bengali"))
+ *alphabet = SMS_ALPHABET_BENGALI;
+ else if (g_str_equal(str, "gujarati"))
+ *alphabet = SMS_ALPHABET_GUJARATI;
else
return FALSE;

diff --git a/src/smsutil.c b/src/smsutil.c
index 24dcfaa..dfa838d 100644
--- a/src/smsutil.c
+++ b/src/smsutil.c
@@ -2280,10 +2280,10 @@ char *sms_decode_text(GSList *sms_list)
* If language is not defined in 3GPP TS 23.038,
* implementations are instructed to ignore it
*/
- if (locking_shift > SMS_ALPHABET_PORTUGUESE)
+ if (locking_shift > SMS_ALPHABET_GUJARATI)
locking_shift = GSM_DIALECT_DEFAULT;

- if (single_shift > SMS_ALPHABET_PORTUGUESE)
+ if (single_shift > SMS_ALPHABET_GUJARATI)
single_shift = GSM_DIALECT_DEFAULT;

converted = convert_gsm_to_utf8_with_lang(buf, written,
diff --git a/src/smsutil.h b/src/smsutil.h
index f293350..3f585cb 100644
--- a/src/smsutil.h
+++ b/src/smsutil.h
@@ -159,6 +159,8 @@ enum sms_alphabet {
SMS_ALPHABET_TURKISH,
SMS_ALPHABET_SPANISH,
SMS_ALPHABET_PORTUGUESE,
+ SMS_ALPHABET_BENGALI,
+ SMS_ALPHABET_GUJARATI,
};

enum sms_mwi_type {
diff --git a/src/util.c b/src/util.c
index df9d4f0..2f01cec 100644
--- a/src/util.c
+++ b/src/util.c
@@ -288,6 +288,324 @@ static const struct codepoint por_ext_unicode[] = {
{ 0x20AC, 0x1B65 }
};

+/* Appendix A.2.4. in 3GPP TS23.038 V.8.2.0 */
+static const struct codepoint ben_ext_gsm[] = {
+ { 0x00, 0x0040 },
+ { 0x01, 0x00A3 },
+ { 0x02, 0x0024 },
+ { 0x03, 0x00A5 },
+ { 0x04, 0x00BF },
+ { 0x05, 0x0022 },
+ { 0x06, 0x00A4 },
+ { 0x07, 0x0025 },
+ { 0x08, 0x0026 },
+ { 0x09, 0x0027 },
+ { 0x0A, 0x000C }, /* See NOTE 3 */
+ { 0x0B, 0x002A },
+ { 0x0C, 0x002B },
+ { 0x0E, 0x002D },
+ { 0x0F, 0x002F },
+ { 0x10, 0X003C },
+ { 0x11, 0X003D },
+ { 0x12, 0X003E },
+ { 0x13, 0X00A1 },
+ { 0x14, 0X005E },
+ { 0x15, 0X00A1 },
+ { 0x16, 0X005F },
+ { 0x17, 0X0023 },
+ { 0x18, 0X002A },
+ { 0x19, 0X09E6 },
+ { 0x1A, 0X09E7 },
+ { 0x1B, 0X0020 }, /* See NOTE 1 */
+ { 0x1C, 0X09E8 },
+ { 0x1D, 0X09E9 },
+ { 0x1E, 0X09EA },
+ { 0x1F, 0X09EB },
+ { 0x20, 0x09EC },
+ { 0x21, 0x09ED },
+ { 0x22, 0x09EE },
+ { 0x23, 0x09EF },
+ { 0x24, 0x09DF },
+ { 0x25, 0x09E0 },
+ { 0x26, 0x09E1 },
+ { 0x27, 0x09E2 },
+ { 0x28, 0x007B },
+ { 0x29, 0x007D },
+ { 0x2A, 0x09E3 },
+ { 0x2B, 0x09F2 },
+ { 0x2C, 0x09F3 },
+ { 0x2D, 0x09F4 },
+ { 0x2E, 0x09F5 },
+ { 0x2F, 0x005C },
+ { 0x30, 0x09F6 },
+ { 0x31, 0x09F7 },
+ { 0x32, 0x09F8 },
+ { 0x33, 0x09F9 },
+ { 0x34, 0x09FA },
+ { 0x3C, 0x005B },
+ { 0x3D, 0x007E },
+ { 0x3E, 0x005D },
+ { 0x40, 0x007C },
+ { 0x41, 0x0041 },
+ { 0x42, 0x0042 },
+ { 0x43, 0x0043 },
+ { 0x44, 0x0044 },
+ { 0x45, 0x0045 },
+ { 0x46, 0x0046 },
+ { 0x47, 0x0047 },
+ { 0x48, 0x0048 },
+ { 0x49, 0x0049 },
+ { 0x4A, 0x004A },
+ { 0x4B, 0x004B },
+ { 0x4C, 0x004C },
+ { 0x4D, 0x004D },
+ { 0x4E, 0x004E },
+ { 0x4F, 0x004F },
+ { 0x50, 0x0050 },
+ { 0x51, 0x0051 },
+ { 0x52, 0x0052 },
+ { 0x53, 0x0053 },
+ { 0x54, 0x0054 },
+ { 0x55, 0x0055 },
+ { 0x56, 0x0056 },
+ { 0x57, 0x0057 },
+ { 0x58, 0x0058 },
+ { 0x59, 0x0059 },
+ { 0x5A, 0x005A },
+ { 0x65, 0x20AC }
+};
+
+static const struct codepoint ben_ext_unicode[] = {
+ { 0x0022, 0x1B05 },
+ { 0X0023, 0x1B17 },
+ { 0x0024, 0x1B02 },
+ { 0x0025, 0x1B07 },
+ { 0x0026, 0x1B08 },
+ { 0x0027, 0x1B09 },
+ { 0x002A, 0x1B0B },
+ { 0X002A, 0x1B18 },
+ { 0x002B, 0x1B0C },
+ { 0x002D, 0x1B0E },
+ { 0x002F, 0x1B0F },
+ { 0X003C, 0x1B10 },
+ { 0X003D, 0x1B11 },
+ { 0X003E, 0x1B12 },
+ { 0x0040, 0x1B00 },
+ { 0x0041, 0x1B41 },
+ { 0x0042, 0x1B42 },
+ { 0x0043, 0x1B43 },
+ { 0x0044, 0x1B44 },
+ { 0x0045, 0x1B45 },
+ { 0x0046, 0x1B46 },
+ { 0x0047, 0x1B47 },
+ { 0x0048, 0x1B48 },
+ { 0x0049, 0x1B49 },
+ { 0x004A, 0x1B4A },
+ { 0x004B, 0x1B4B },
+ { 0x004C, 0x1B4C },
+ { 0x004D, 0x1B4D },
+ { 0x004E, 0x1B4E },
+ { 0x004F, 0x1B4F },
+ { 0x0050, 0x1B50 },
+ { 0x0051, 0x1B51 },
+ { 0x0052, 0x1B52 },
+ { 0x0053, 0x1B53 },
+ { 0x0054, 0x1B54 },
+ { 0x0055, 0x1B55 },
+ { 0x0056, 0x1B56 },
+ { 0x0057, 0x1B57 },
+ { 0x0058, 0x1B58 },
+ { 0x0059, 0x1B59 },
+ { 0x005A, 0x1B5A },
+ { 0x005B, 0x1B3C },
+ { 0x005C, 0x1B2F },
+ { 0x005D, 0x1B3E },
+ { 0X005E, 0x1B14 },
+ { 0X005F, 0x1B16 },
+ { 0x007B, 0x1B28 },
+ { 0x007C, 0x1B40 },
+ { 0x007D, 0x1B29 },
+ { 0x007E, 0x1B3D },
+ { 0X00A1, 0x1B13 },
+ { 0X00A1, 0x1B15 },
+ { 0x00A3, 0x1B01 },
+ { 0x00A4, 0x1B06 },
+ { 0x00A5, 0x1B03 },
+ { 0x00BF, 0x1B04 },
+ { 0x09DF, 0x1B24 },
+ { 0x09E0, 0x1B25 },
+ { 0x09E1, 0x1B26 },
+ { 0x09E2, 0x1B27 },
+ { 0x09E3, 0x1B2A },
+ { 0X09E6, 0x1B19 },
+ { 0X09E7, 0x1B1A },
+ { 0X09E8, 0x1B1C },
+ { 0X09E9, 0x1B1D },
+ { 0X09EA, 0x1B1E },
+ { 0X09EB, 0x1B1F },
+ { 0x09EC, 0x1B20 },
+ { 0x09ED, 0x1B21 },
+ { 0x09EE, 0x1B22 },
+ { 0x09EF, 0x1B23 },
+ { 0x09F2, 0x1B2B },
+ { 0x09F3, 0x1B2C },
+ { 0x09F4, 0x1B2D },
+ { 0x09F5, 0x1B2E },
+ { 0x09F6, 0x1B30 },
+ { 0x09F7, 0x1B31 },
+ { 0x09F8, 0x1B32 },
+ { 0x09F9, 0x1B33 },
+ { 0x09FA, 0x1B34 },
+ { 0x20AC, 0x1B65 }
+};
+
+/* Appendix A.2.5. in 3GPP TS23.038 V.8.2.0 */
+static const struct codepoint guj_ext_gsm[] = {
+ { 0x00, 0x0040 },
+ { 0x01, 0x00A3 },
+ { 0x02, 0x0024 },
+ { 0x03, 0x00A5 },
+ { 0x04, 0x00BF },
+ { 0x05, 0x0022 },
+ { 0x06, 0x00A4 },
+ { 0x07, 0x0025 },
+ { 0x08, 0x0026 },
+ { 0x09, 0x0027 },
+ { 0x0A, 0x000C }, /* See NOTE 3 */
+ { 0x0B, 0x002A },
+ { 0x0C, 0x002B },
+ { 0x0E, 0x002D },
+ { 0x0F, 0x002F },
+ { 0x10, 0x003C },
+ { 0x11, 0x003D },
+ { 0x12, 0x003E },
+ { 0x13, 0x00A1 },
+ { 0x14, 0x005E },
+ { 0x15, 0x00A1 },
+ { 0x16, 0x005F },
+ { 0x17, 0x0023 },
+ { 0x18, 0x002A },
+ { 0x19, 0x0964 },
+ { 0x1A, 0x0965 },
+ { 0x1B, 0x0020 }, /* See NOTE 1 */
+ { 0x1C, 0x0AE6 },
+ { 0x1D, 0x0AE7 },
+ { 0x1E, 0x0AE8 },
+ { 0x1F, 0x0AE9 },
+ { 0x20, 0x0AEA },
+ { 0x21, 0x0AEB },
+ { 0x22, 0x0AEC },
+ { 0x23, 0x0AED },
+ { 0x24, 0x0AEE },
+ { 0x25, 0x0AEF },
+ { 0x28, 0x007B },
+ { 0x29, 0x007D },
+ { 0x2F, 0x005C },
+ { 0x3C, 0x005B },
+ { 0x3D, 0x007E },
+ { 0x3E, 0x005D },
+ { 0x40, 0x007C },
+ { 0x41, 0x0041 },
+ { 0x42, 0x0042 },
+ { 0x43, 0x0043 },
+ { 0x44, 0x0044 },
+ { 0x45, 0x0045 },
+ { 0x46, 0x0046 },
+ { 0x47, 0x0047 },
+ { 0x48, 0x0048 },
+ { 0x49, 0x0049 },
+ { 0x4A, 0x004A },
+ { 0x4B, 0x004B },
+ { 0x4C, 0x004C },
+ { 0x4D, 0x004D },
+ { 0x4E, 0x004E },
+ { 0x4F, 0x004F },
+ { 0x50, 0x0050 },
+ { 0x51, 0x0051 },
+ { 0x52, 0x0052 },
+ { 0x53, 0x0053 },
+ { 0x54, 0x0054 },
+ { 0x55, 0x0055 },
+ { 0x56, 0x0056 },
+ { 0x57, 0x0057 },
+ { 0x58, 0x0058 },
+ { 0x59, 0x0059 },
+ { 0x5A, 0x005A },
+ { 0x65, 0x20AC }
+};
+
+static const struct codepoint guj_ext_unicode[] = {
+ { 0x0022, 0x1B05 },
+ { 0x0023, 0x1B17 },
+ { 0x0024, 0x1B02 },
+ { 0x0025, 0x1B07 },
+ { 0x0026, 0x1B08 },
+ { 0x0027, 0x1B09 },
+ { 0x002A, 0x1B0B },
+ { 0x002A, 0x1B18 },
+ { 0x002B, 0x1B0C },
+ { 0x002D, 0x1B0E },
+ { 0x002F, 0x1B0F },
+ { 0x003C, 0x1B10 },
+ { 0x003D, 0x1B11 },
+ { 0x003E, 0x1B12 },
+ { 0x0040, 0x1B00 },
+ { 0x0041, 0x1B41 },
+ { 0x0042, 0x1B42 },
+ { 0x0043, 0x1B43 },
+ { 0x0044, 0x1B44 },
+ { 0x0045, 0x1B45 },
+ { 0x0046, 0x1B46 },
+ { 0x0047, 0x1B47 },
+ { 0x0048, 0x1B48 },
+ { 0x0049, 0x1B49 },
+ { 0x004A, 0x1B4A },
+ { 0x004B, 0x1B4B },
+ { 0x004C, 0x1B4C },
+ { 0x004D, 0x1B4D },
+ { 0x004E, 0x1B4E },
+ { 0x004F, 0x1B4F },
+ { 0x0050, 0x1B50 },
+ { 0x0051, 0x1B51 },
+ { 0x0052, 0x1B52 },
+ { 0x0053, 0x1B53 },
+ { 0x0054, 0x1B54 },
+ { 0x0055, 0x1B55 },
+ { 0x0056, 0x1B56 },
+ { 0x0057, 0x1B57 },
+ { 0x0058, 0x1B58 },
+ { 0x0059, 0x1B59 },
+ { 0x005A, 0x1B5A },
+ { 0x005B, 0x1B3C },
+ { 0x005C, 0x1B2F },
+ { 0x005D, 0x1B3E },
+ { 0x005E, 0x1B14 },
+ { 0x005F, 0x1B16 },
+ { 0x007B, 0x1B28 },
+ { 0x007C, 0x1B40 },
+ { 0x007D, 0x1B29 },
+ { 0x007E, 0x1B3D },
+ { 0x00A1, 0x1B13 },
+ { 0x00A1, 0x1B15 },
+ { 0x00A3, 0x1B01 },
+ { 0x00A4, 0x1B06 },
+ { 0x00A5, 0x1B03 },
+ { 0x00BF, 0x1B04 },
+ { 0x0964, 0x1B19 },
+ { 0x0965, 0x1B1A },
+ { 0x0AE6, 0x1B1C },
+ { 0x0AE7, 0x1B1D },
+ { 0x0AE8, 0x1B1E },
+ { 0x0AE9, 0x1B1F },
+ { 0x0AEA, 0x1B20 },
+ { 0x0AEB, 0x1B21 },
+ { 0x0AEC, 0x1B22 },
+ { 0x0AED, 0x1B23 },
+ { 0x0AEE, 0x1B24 },
+ { 0x0AEF, 0x1B25 },
+ { 0x20AC, 0x1B65 }
+};
+
/* Used for conversion of GSM to Unicode */
static const unsigned short def_gsm[] = {
0x0040, 0x00A3, 0x0024, 0x00A5, 0x00E8, 0x00E9, 0x00F9, 0x00EC,
@@ -454,6 +772,112 @@ static const struct codepoint por_unicode[] = {
{ 0x00FC, 0x7E }, { 0x0394, 0x10 }, { 0x20AC, 0x18 }, { 0x221E, 0x15 }
};

+/* Appendix A.3.4 in 3GPP TS23.038 */
+static const unsigned short ben_gsm[] = {
+ 0x0981, 0x0982, 0x0983, 0x0985, 0x0986, 0x0987, 0x0988, 0x0989,
+ 0x098A, 0x098B, 0x000A, 0x098C, 0x0020, 0x000D, 0x0020, 0x098F,
+ 0x0990, 0x0020, 0x0020, 0x0993, 0x0994, 0x0995, 0x0996, 0x0997,
+ 0x0998, 0x0999, 0x099A, 0x00A0, 0x099B, 0x099C, 0x099D, 0x099E,
+ 0x0020, 0x0021, 0x099F, 0x09A0, 0x09A1, 0x09A2, 0x09A3, 0x09A4,
+ 0x0029, 0x0028, 0x09A5, 0x09A6, 0x002C, 0x09A7, 0x002E, 0x09A8,
+ 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037,
+ 0x0038, 0x0039, 0x003A, 0x003B, 0x0020, 0x09AA, 0x09AB, 0x003F,
+ 0x09AC, 0x09AD, 0x09AE, 0x09AF, 0x09B0, 0x0020, 0x09B2, 0x0020,
+ 0x0020, 0x0020, 0x09B6, 0x09B7, 0x09B8, 0x09B9, 0x09BC, 0x09BD,
+ 0x09BE, 0x09BF, 0x09C0, 0x09C1, 0x09C2, 0x09C3, 0x09C4, 0x0020,
+ 0x0020, 0x09C7, 0x09C8, 0x0020, 0x0020, 0x09CB, 0x09CC, 0x09CD,
+ 0x09CE, 0x0061, 0x0062, 0x0063, 0x0064, 0x0065, 0x0066, 0x0067,
+ 0x0068, 0x0069, 0x006A, 0x006B, 0x006C, 0x006D, 0x006E, 0x006F,
+ 0x0070, 0x0071, 0x0072, 0x0073, 0x0074, 0x0075, 0x0076, 0x0077,
+ 0x0078, 0x0079, 0x007A, 0x09D7, 0x09DC, 0x09DD, 0x09F0, 0x09F1
+};
+
+static const struct codepoint ben_unicode[] = {
+ { 0x000A, 0x0A }, { 0x000D, 0x0D }, { 0x0020, 0x20 }, { 0x0021, 0x21 },
+ { 0x0028, 0x29 }, { 0x0029, 0x28 }, { 0x002C, 0x2C }, { 0x002E, 0x2E },
+ { 0x0030, 0x30 }, { 0x0031, 0x31 }, { 0x0032, 0x32 }, { 0x0033, 0x33 },
+ { 0x0034, 0x34 }, { 0x0035, 0x35 }, { 0x0036, 0x36 }, { 0x0037, 0x37 },
+ { 0x0038, 0x38 }, { 0x0039, 0x39 }, { 0x003A, 0x3A }, { 0x003B, 0x3B },
+ { 0x003F, 0x3F }, { 0x0061, 0x61 }, { 0x0062, 0x62 }, { 0x0063, 0x63 },
+ { 0x0064, 0x64 }, { 0x0065, 0x65 }, { 0x0066, 0x66 }, { 0x0067, 0x67 },
+ { 0x0068, 0x68 }, { 0x0069, 0x69 }, { 0x006A, 0x6A }, { 0x006B, 0x6B },
+ { 0x006C, 0x6C }, { 0x006D, 0x6D }, { 0x006E, 0x6E }, { 0x006F, 0x6F },
+ { 0x0070, 0x70 }, { 0x0071, 0x71 }, { 0x0072, 0x72 }, { 0x0073, 0x73 },
+ { 0x0074, 0x74 }, { 0x0075, 0x75 }, { 0x0076, 0x76 }, { 0x0077, 0x77 },
+ { 0x0078, 0x78 }, { 0x0079, 0x79 }, { 0x007A, 0x7A }, { 0x00A0, 0x20 },
+ { 0x0981, 0x00 }, { 0x0982, 0x01 }, { 0x0983, 0x02 }, { 0x0985, 0x03 },
+ { 0x0986, 0x04 }, { 0x0987, 0x05 }, { 0x0988, 0x06 }, { 0x0989, 0x07 },
+ { 0x098A, 0x08 }, { 0x098B, 0x09 }, { 0x098C, 0x0B }, { 0x098F, 0x0F },
+ { 0x0990, 0x10 }, { 0x0993, 0x13 }, { 0x0994, 0x14 }, { 0x0995, 0x15 },
+ { 0x0996, 0x16 }, { 0x0997, 0x17 }, { 0x0998, 0x18 }, { 0x0999, 0x19 },
+ { 0x099A, 0x1A }, { 0x099B, 0x1C }, { 0x099C, 0x1D }, { 0x099D, 0x1E },
+ { 0x099E, 0x1F }, { 0x099F, 0x22 }, { 0x09A0, 0x23 }, { 0x09A1, 0x24 },
+ { 0x09A2, 0x25 }, { 0x09A3, 0x26 }, { 0x09A4, 0x27 }, { 0x09A5, 0x2A },
+ { 0x09A6, 0x2B }, { 0x09A7, 0x2D }, { 0x09A8, 0x2F }, { 0x09AA, 0x3D },
+ { 0x09AB, 0x3E }, { 0x09AC, 0x40 }, { 0x09AD, 0x41 }, { 0x09AE, 0x42 },
+ { 0x09AF, 0x43 }, { 0x09B0, 0x44 }, { 0x09B2, 0x46 }, { 0x09B6, 0x4A },
+ { 0x09B7, 0x4B }, { 0x09B8, 0x4C }, { 0x09B9, 0x4D }, { 0x09BC, 0x4E },
+ { 0x09BD, 0x4F }, { 0x09BE, 0x50 }, { 0x09BF, 0x51 }, { 0x09C0, 0x52 },
+ { 0x09C1, 0x53 }, { 0x09C2, 0x54 }, { 0x09C3, 0x55 }, { 0x09C4, 0x56 },
+ { 0x09C7, 0x59 }, { 0x09C8, 0x5A }, { 0x09CB, 0x5D }, { 0x09CC, 0x5E },
+ { 0x09CD, 0x5F }, { 0x09CE, 0x60 }, { 0x09D7, 0x7B }, { 0x09DC, 0x7C },
+ { 0x09DD, 0x7D }, { 0x09F0, 0x7E }, { 0x09F1, 0x7F }
+};
+
+/* Appendix A.3.5 in 3GPP TS23.038 */
+static const unsigned short guj_gsm[] = {
+ 0x0A81, 0x0A82, 0x0A83, 0x0A85, 0x0A86, 0x0A87, 0x0A88, 0x0A89,
+ 0x0A8A, 0x0A8B, 0x000A, 0x0A8C, 0x0A8D, 0x000D, 0x0020, 0x0A8F,
+ 0x0A90, 0x0A91, 0x0020, 0x0A93, 0x0A94, 0x0A95, 0x0A96, 0x0A97,
+ 0x0A98, 0x0A99, 0x0A9A, 0x00A0, 0x0A9B, 0x0A9C, 0x0A9D, 0x0A9E,
+ 0x0020, 0x0021, 0x0A9F, 0x0AA0, 0x0AA1, 0x0AA2, 0x0AA3, 0x0AA4,
+ 0x0029, 0x0028, 0x0AA5, 0x0AA6, 0x002C, 0x0AA7, 0x002E, 0x0AA8,
+ 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037,
+ 0x0038, 0x0039, 0x003A, 0x003B, 0x0020, 0x0AAA, 0x0AAB, 0x003F,
+ 0x0AAC, 0x0AAD, 0x0AAE, 0x0AAF, 0x0AB0, 0x0020, 0x0AB2, 0x0AB3,
+ 0x0020, 0x0AB5, 0x0AB6, 0x0AB7, 0x0AB8, 0x0AB9, 0x0ABC, 0x0ABD,
+ 0x0ABE, 0x0ABF, 0x0AC0, 0x0AC1, 0x0AC2, 0x0AC3, 0x0AC4, 0x0AC5,
+ 0x0020, 0x0AC7, 0x0AC8, 0x0AC9, 0x0020, 0x0ACB, 0x0ACC, 0x0ACD,
+ 0x0AD0, 0x0061, 0x0062, 0x0063, 0x0064, 0x0065, 0x0066, 0x0067,
+ 0x0068, 0x0069, 0x006A, 0x006B, 0x006C, 0x006D, 0x006E, 0x006F,
+ 0x0070, 0x0071, 0x0072, 0x0073, 0x0074, 0x0075, 0x0076, 0x0077,
+ 0x0078, 0x0079, 0x007A, 0x0AE0, 0x0AE1, 0x0AE2, 0x0AE3, 0x0AF1
+};
+
+static const struct codepoint guj_unicode[] = {
+ { 0x000A, 0x0A }, { 0x000D, 0x0D }, { 0x0020, 0x20 }, { 0x0021, 0x21 },
+ { 0x0028, 0x29 }, { 0x0029, 0x28 }, { 0x002C, 0x2C }, { 0x002E, 0x2E },
+ { 0x0030, 0x30 }, { 0x0031, 0x31 }, { 0x0032, 0x32 }, { 0x0033, 0x33 },
+ { 0x0034, 0x34 }, { 0x0035, 0x35 }, { 0x0036, 0x36 }, { 0x0037, 0x37 },
+ { 0x0038, 0x38 }, { 0x0039, 0x39 }, { 0x003A, 0x3A }, { 0x003B, 0x3B },
+ { 0x003F, 0x3F }, { 0x0061, 0x61 }, { 0x0062, 0x62 }, { 0x0063, 0x63 },
+ { 0x0064, 0x64 }, { 0x0065, 0x65 }, { 0x0066, 0x66 }, { 0x0067, 0x67 },
+ { 0x0068, 0x68 }, { 0x0069, 0x69 }, { 0x006A, 0x6A }, { 0x006B, 0x6B },
+ { 0x006C, 0x6C }, { 0x006D, 0x6D }, { 0x006E, 0x6E }, { 0x006F, 0x6F },
+ { 0x0070, 0x70 }, { 0x0071, 0x71 }, { 0x0072, 0x72 }, { 0x0073, 0x73 },
+ { 0x0074, 0x74 }, { 0x0075, 0x75 }, { 0x0076, 0x76 }, { 0x0077, 0x77 },
+ { 0x0078, 0x78 }, { 0x0079, 0x79 }, { 0x007A, 0x7A }, { 0x00A0, 0x20 },
+ { 0x0A81, 0x00 }, { 0x0A82, 0x01 }, { 0x0A83, 0x02 }, { 0x0A85, 0x03 },
+ { 0x0A86, 0x04 }, { 0x0A87, 0x05 }, { 0x0A88, 0x06 }, { 0x0A89, 0x07 },
+ { 0x0A8A, 0x08 }, { 0x0A8B, 0x09 }, { 0x0A8C, 0x0B }, { 0x0A8D, 0x0C },
+ { 0x0A8F, 0x0F }, { 0x0A90, 0x10 }, { 0x0A91, 0x11 }, { 0x0A93, 0x13 },
+ { 0x0A94, 0x14 }, { 0x0A95, 0x15 }, { 0x0A96, 0x16 }, { 0x0A97, 0x17 },
+ { 0x0A98, 0x18 }, { 0x0A99, 0x19 }, { 0x0A9A, 0x1A }, { 0x0A9B, 0x1C },
+ { 0x0A9C, 0x1D }, { 0x0A9D, 0x1E }, { 0x0A9E, 0x1F }, { 0x0A9F, 0x22 },
+ { 0x0AA0, 0x23 }, { 0x0AA1, 0x24 }, { 0x0AA2, 0x25 }, { 0x0AA3, 0x26 },
+ { 0x0AA4, 0x27 }, { 0x0AA5, 0x2A }, { 0x0AA6, 0x2B }, { 0x0AA7, 0x2D },
+ { 0x0AA8, 0x2F }, { 0x0AAA, 0x3D }, { 0x0AAB, 0x3E }, { 0x0AAC, 0x40 },
+ { 0x0AAD, 0x41 }, { 0x0AAE, 0x42 }, { 0x0AAF, 0x43 }, { 0x0AB0, 0x44 },
+ { 0x0AB2, 0x46 }, { 0x0AB3, 0x47 }, { 0x0AB5, 0x49 }, { 0x0AB6, 0x4A },
+ { 0x0AB7, 0x4B }, { 0x0AB8, 0x4C }, { 0x0AB9, 0x4D }, { 0x0ABC, 0x4E },
+ { 0x0ABD, 0x4F }, { 0x0ABE, 0x50 }, { 0x0ABF, 0x51 }, { 0x0AC0, 0x52 },
+ { 0x0AC1, 0x53 }, { 0x0AC2, 0x54 }, { 0x0AC3, 0x55 }, { 0x0AC4, 0x56 },
+ { 0x0AC5, 0x57 }, { 0x0AC7, 0x59 }, { 0x0AC8, 0x5A }, { 0x0AC9, 0x5B },
+ { 0x0ACB, 0x5D }, { 0x0ACC, 0x5E }, { 0x0ACD, 0x5F }, { 0x0AD0, 0x60 },
+ { 0x0AE0, 0x7B }, { 0x0AE1, 0x7C }, { 0x0AE2, 0x7D }, { 0x0AE3, 0x7E },
+ { 0x0AF1, 0x7F }
+};
+
static int compare_codepoints(const void *a, const void *b)
{
const struct codepoint *ca = (const struct codepoint *) a;
@@ -523,7 +947,19 @@ static gboolean populate_locking_shift(struct conversion_table *t,
t->locking_u = por_unicode;
t->locking_len_u = TABLE_SIZE(por_unicode);
return TRUE;
- }
+
+ case GSM_DIALECT_BENGALI:
+ t->locking_g = ben_gsm;
+ t->locking_u = ben_unicode;
+ t->locking_len_u = TABLE_SIZE(ben_unicode);
+ return TRUE;
+
+ case GSM_DIALECT_GUJARATI:
+ t->locking_g = guj_gsm;
+ t->locking_u = guj_unicode;
+ t->locking_len_u = TABLE_SIZE(guj_unicode);
+ return TRUE;
+ }

return FALSE;
}
@@ -559,7 +995,21 @@ static gboolean populate_single_shift(struct conversion_table *t,
t->single_u = por_ext_unicode;
t->single_len_u = TABLE_SIZE(por_ext_unicode);
return TRUE;
- }
+
+ case GSM_DIALECT_BENGALI:
+ t->single_g = ben_ext_gsm;
+ t->single_len_g = TABLE_SIZE(ben_ext_gsm);
+ t->single_u = ben_ext_unicode;
+ t->single_len_u = TABLE_SIZE(ben_ext_unicode);
+ return TRUE;
+
+ case GSM_DIALECT_GUJARATI:
+ t->single_g = guj_ext_gsm;
+ t->single_len_g = TABLE_SIZE(guj_ext_gsm);
+ t->single_u = guj_ext_unicode;
+ t->single_len_u = TABLE_SIZE(guj_ext_unicode);
+ return TRUE;
+ }

return FALSE;
}
diff --git a/src/util.h b/src/util.h
index 092b4b5..193eb8b 100644
--- a/src/util.h
+++ b/src/util.h
@@ -24,6 +24,8 @@ enum gsm_dialect {
GSM_DIALECT_TURKISH,
GSM_DIALECT_SPANISH,
GSM_DIALECT_PORTUGUESE,
+ GSM_DIALECT_BENGALI,
+ GSM_DIALECT_GUJARATI,
};

char *convert_gsm_to_utf8(const unsigned char *text, long len, long *items_read,
--
2.7.4
Loading...