cbm-part: Handle additional coding schemes and parse languages

We store the language as iso639 as this is what the what TS 23.038
references.

Signed-off-by: Guido Günther <agx@sigxcpu.org>

3 files changed, 109 insertions, 2 deletions

diff --git a/src/mm-cbm-part.c b/src/mm-cbm-part.c
index 192bab72..18df7c17 100644
--- a/src/mm-cbm-part.c
+++ b/src/mm-cbm-part.c
@@ -27,9 +27,17 @@
 #include "mm-sms-part-3gpp.h"
 
 #define CBS_DATA_CODING_GROUP_MASK           0b11110000
+#define CBS_DATA_CODING_LANG_MASK            0b00001111
 #define CBS_DATA_CODING_LANG_GSM7            0b00000000
 #define CBS_DATA_CODING_GSM7                 0b00010000
 #define CBS_DATA_CODING_UCS2                 0b00010001
+#define CBS_DATA_CODING_OTHER                0b00100000
+#define CBS_DATA_CODING_OTHER_CZECH          0b00100000
+#define CBS_DATA_CODING_OTHER_HEBREW         0b00100001
+#define CBS_DATA_CODING_OTHER_ARABIC         0b00100010
+#define CBS_DATA_CODING_OTHER_RUSSIAN        0b00100011
+#define CBS_DATA_CODING_OTHER_ICELANDIC      0b00100100
+#define CBS_DATA_CODING_UNSPECIFIED          0b00110000
 #define CBS_DATA_CODING_GENERAL_NO_CLASS     0b01000000
 #define CBS_DATA_CODING_GENERAL_CLASS        0b01010000
 #define CBS_DATA_CODING_GENERAL_CHARSET_MASK 0b00001100
@@ -46,10 +54,48 @@ struct _MMCbmPart {
     guint8 part_num;
 
     gchar *text;
+    gchar *language;
     MMSmsEncoding encoding;
 };
 
 
+static gchar *
+mm_cbm_part_coding_group_to_language (guint8 group)
+{
+    switch (group & CBS_DATA_CODING_LANG_MASK) {
+    case 0b0000:
+        return g_strdup ("de");
+    case 0b0001:
+        return g_strdup ("en");
+    case 0b0010:
+        return g_strdup ("it");
+    case 0b0011:
+        return g_strdup ("fr");
+    case 0b0100:
+        return g_strdup ("es");
+    case 0b0101:
+        return g_strdup ("nl");
+    case 0b0110:
+        return g_strdup ("sv");
+    case 0b0111:
+        return g_strdup ("dk");
+    case 0b1000:
+        return g_strdup ("pt");
+    case 0b1001:
+        return g_strdup ("fi");
+    case 0b1010:
+        return g_strdup ("no");
+    case 0b1011:
+        return g_strdup ("el");
+    case 0b1110:
+        return g_strdup ("pl");
+    case 0b1111:
+    default:
+        return NULL;
+    };
+}
+
+
 MMCbmPart *
 mm_cbm_part_new_from_pdu (const gchar  *hexpdu,
                           gpointer      log_object,
@@ -136,12 +182,40 @@ mm_cbm_part_new_from_binary_pdu (const guint8  *pdu,
     /* Order matches 3GPP TS 23.038 Chapter 5 */
     if (group == CBS_DATA_CODING_LANG_GSM7) {
         cbm_part->encoding = MM_SMS_ENCODING_GSM7;
+        cbm_part->language = mm_cbm_part_coding_group_to_language (pdu[offset]);
     } else if (pdu[offset] == CBS_DATA_CODING_GSM7) {
         has_lang = TRUE;
         cbm_part->encoding = MM_SMS_ENCODING_GSM7;
     } else if (pdu[offset] == CBS_DATA_CODING_UCS2) {
         has_7bit_lang = TRUE;
         cbm_part->encoding = MM_SMS_ENCODING_UCS2;
+    } else if (group == CBS_DATA_CODING_OTHER) {
+        switch (group) {
+        case CBS_DATA_CODING_OTHER_CZECH:
+            cbm_part->encoding = MM_SMS_ENCODING_GSM7;
+            cbm_part->language = g_strdup ("cz");
+            break;
+        case CBS_DATA_CODING_OTHER_HEBREW:
+            cbm_part->encoding = MM_SMS_ENCODING_UCS2;
+            cbm_part->language = g_strdup ("he");
+            break;
+        case CBS_DATA_CODING_OTHER_ARABIC:
+            cbm_part->encoding = MM_SMS_ENCODING_UCS2;
+            cbm_part->language = g_strdup ("ar");
+            break;
+        case CBS_DATA_CODING_OTHER_RUSSIAN:
+            cbm_part->encoding = MM_SMS_ENCODING_UCS2;
+            cbm_part->language = g_strdup ("ru");
+            break;
+        case CBS_DATA_CODING_OTHER_ICELANDIC:
+            cbm_part->encoding = MM_SMS_ENCODING_GSM7;
+            cbm_part->language = g_strdup ("is");
+            break;
+        default:
+            cbm_part->encoding = MM_SMS_ENCODING_GSM7;
+        }
+    } else if (group == CBS_DATA_CODING_UNSPECIFIED) {
+        cbm_part->encoding = MM_SMS_ENCODING_GSM7;
     } else if ((group == CBS_DATA_CODING_GENERAL_CLASS) ||
                (group == CBS_DATA_CODING_GENERAL_NO_CLASS) ||
                (group == CBS_DATA_CODING_UDH)) {
@@ -167,10 +241,31 @@ mm_cbm_part_new_from_binary_pdu (const guint8  *pdu,
     offset++;
 
     if (has_lang) {
-        PDU_SIZE_CHECK (offset + 4, "cannot skip lang");
+        PDU_SIZE_CHECK (offset + 4, "cannot read lang");
+        if (pdu[offset+2] != '\r') {
+            g_set_error_literal (error,
+                                 MM_CORE_ERROR,
+                                 MM_CORE_ERROR_FAILED,
+                                 "Failed to parse lang");
+            return NULL;
+        }
+        cbm_part->language = g_strdup_printf ("%c%c", pdu[offset], pdu[offset+1]);
         offset += 3;
     } else if (has_7bit_lang) {
-        PDU_SIZE_CHECK (offset + 3, "cannot skip 7bit lang");
+        PDU_SIZE_CHECK (offset + 3, "cannot read 7bit lang");
+        cbm_part->language = mm_sms_decode_text (&pdu[offset],
+                                                 2,
+                                                 MM_SMS_ENCODING_GSM7,
+                                                 0,
+                                                 log_object,
+                                                 NULL);
+        if (!cbm_part->language) {
+            g_set_error_literal (error,
+                                 MM_CORE_ERROR,
+                                 MM_CORE_ERROR_FAILED,
+                                 "Failed to parse lang");
+            return NULL;
+        }
         offset += 2;
     }
 
@@ -210,6 +305,7 @@ mm_cbm_part_new (void)
 void
 mm_cbm_part_free (MMCbmPart *part)
 {
+  g_clear_pointer (&part->language, g_free);
   g_clear_pointer (&part->text, g_free);
   g_slice_free (MMCbmPart, part);
 }
@@ -226,3 +322,4 @@ PART_GET_FUNC (guint, num_parts)
 PART_GET_FUNC (const char *, text)
 PART_GET_FUNC (guint16, channel)
 PART_GET_FUNC (guint16, serial)
+PART_GET_FUNC (const char *, language)
diff --git a/src/mm-cbm-part.h b/src/mm-cbm-part.h
index 0cb9d108..3e0dabf9 100644
--- a/src/mm-cbm-part.h
+++ b/src/mm-cbm-part.h
@@ -67,5 +67,6 @@ const char *mm_cbm_part_get_text (MMCbmPart *part);
 
 guint16     mm_cbm_part_get_serial (MMCbmPart *part);
 guint16     mm_cbm_part_get_channel (MMCbmPart *part);
+const char *mm_cbm_part_get_language (MMCbmPart *part);
 
 #endif
diff --git a/src/tests/test-cbm-part.c b/src/tests/test-cbm-part.c
index 6bdcebb8..89fb4768 100644
--- a/src/tests/test-cbm-part.c
+++ b/src/tests/test-cbm-part.c
@@ -68,6 +68,8 @@ test_cbm_ca (void)
 
     g_assert_cmpstr (mm_cbm_part_get_text (part), ==,
                      "This is a test of the Ontario Alert Ready System. There is no danger to your health or safety" );
+
+    g_assert_null (mm_cbm_part_get_language (part));
 }
 
 static void
@@ -109,6 +111,8 @@ test_cbm_ucs2 (void)
 
     g_assert_cmpstr (mm_cbm_part_get_text (part), ==,
                      " Протягом дня є висока імовірність ракетнᘁ");
+
+    g_assert_null (mm_cbm_part_get_language (part));
 }
 
 
@@ -153,6 +157,8 @@ test_cbm_eu (void)
 
     g_assert_cmpstr (mm_cbm_part_get_text (part), ==,
                      "TEST ALERT, NATIONWIDE ALERT DAY 2022 Thu 2022/12/08 - 10:59 am - Test alert - for Deutschlan");
+
+    g_assert_cmpstr (mm_cbm_part_get_language (part), ==, "en");
 }
 
 static void
@@ -194,6 +200,7 @@ test_cbm_nl_2023 (void)
     g_assert_cmpstr (mm_cbm_part_get_text (part), ==,
                      "NL-Alert 04-12-2023 12:00: TESTBERICHT. De overheid waarschuwt je tijdens noodsituaties via N");
     mm_cbm_part_free (part);
+    g_assert_null (mm_cbm_part_get_language (part));
 
     parse_cbm ("\r\n+CBM: 88\r\n46A011130523CC56905D96D35D206519C42E97E7741039EC06DDC37490BA0C6ABFCB7410F95D7683CA6ED03D1C9683D46550BB5C9683D26EF35BDE0ED3D365D03AEC06D9D36E72D9ED02A9542A10B538A5829AC5E9347804\r\n", &part);
     serial = mm_cbm_part_get_serial (part);
@@ -208,6 +215,7 @@ test_cbm_nl_2023 (void)
     g_assert_cmpstr (mm_cbm_part_get_text (part), ==,
                      "L-Alert. Je leest dan wat je moet doen en waar je meer informatie kan vinden. *** TEST MESSAG");
     mm_cbm_part_free (part);
+    g_assert_null (mm_cbm_part_get_language (part));
 
     parse_cbm ("\r\n+CBM: 65\r\n46A0111305334590B34C4797E5ECB09B3C071DDFF6B2DCDD2EBBE920685DCC4E8F41D7B0DC9D769F41D3FC9C5E6EBB40CE37283CA6A7DF6E90BC1CAFA7E565B2AB\r\n", &part);
     serial = mm_cbm_part_get_serial (part);
@@ -222,6 +230,7 @@ test_cbm_nl_2023 (void)
     g_assert_cmpstr (mm_cbm_part_get_text (part), ==,
                      "E Netherlands Government Public Warning System. No action required." );
     mm_cbm_part_free (part);
+    g_assert_null (mm_cbm_part_get_language (part));
 }