charsets: make translit optional in utf8_to_unpacked_gsm()

If the conversion is not fully compatible, the user of the method
needs to request transliteration enabled explicitly in order to avoid
returning errors in this method.

plugins/huawei/mm-broadband-modem-huawei.c

@@ -2305,8 +2305,11 @@ encode (MMIfaceModem3gppUssd *self,
     guint8 *gsm, *packed;
     guint32 len = 0, packed_len = 0;
 
+    gsm = mm_charset_utf8_to_unpacked_gsm (command, FALSE, &len, error);
+    if (!gsm)
+        return NULL;
+
     *scheme = MM_MODEM_GSM_USSD_SCHEME_7BIT;
-    gsm = mm_charset_utf8_to_unpacked_gsm (command, &len);
 
     /* If command is a multiple of 7 characters long, Huawei firmwares
      * apparently want that padded.  Maybe all modems?

src/mm-broadband-modem-mbim.c

@@ -4777,10 +4777,9 @@ ussd_encode (const gchar  *command,
         guint32  packed_len = 0;
 
         *scheme = MM_MODEM_GSM_USSD_SCHEME_7BIT;
-        gsm = mm_charset_utf8_to_unpacked_gsm (command, &len);
+        gsm = mm_charset_utf8_to_unpacked_gsm (command, FALSE, &len, error);
         if (!gsm) {
-            g_set_error (error, MM_CORE_ERROR, MM_CORE_ERROR_UNSUPPORTED,
+            g_prefix_error (error, "Failed to encode USSD command in GSM7 charset: ");
-                         "Failed to encode USSD command in GSM7 charset");
             return NULL;
         }
         packed = mm_charset_gsm_pack (gsm, len, 0, &packed_len);

src/mm-charsets.c

@@ -427,16 +427,21 @@ mm_charset_gsm_unpacked_to_utf8 (const guint8  *gsm,
 }
 
 guint8 *
-mm_charset_utf8_to_unpacked_gsm (const gchar *utf8,
+mm_charset_utf8_to_unpacked_gsm (const gchar  *utf8,
-                                 guint32     *out_len)
+                                 gboolean      translit,
+                                 guint32      *out_len,
+                                 GError      **error)
 {
-    GByteArray          *gsm;
+    g_autoptr(GByteArray)  gsm = NULL;
-    const gchar         *c;
+    const gchar           *c;
-    const gchar         *next;
+    const gchar           *next;
-    static const guint8  gesc = GSM_ESCAPE_CHAR;
+    static const guint8    gesc = GSM_ESCAPE_CHAR;
 
-    g_return_val_if_fail (utf8 != NULL, NULL);
+    if (!utf8 || !g_utf8_validate (utf8, -1, NULL)) {
-    g_return_val_if_fail (g_utf8_validate (utf8, -1, NULL), NULL);
+        g_set_error (error, MM_CORE_ERROR, MM_CORE_ERROR_INVALID_ARGS,
+                     "Couldn't convert UTF-8 to GSM: input UTF-8 validation failed");
+        return NULL;
+    }
 
     /* worst case initial length */
     gsm = g_byte_array_sized_new (g_utf8_strlen (utf8, -1) * 2 + 1);
@@ -446,7 +451,7 @@ mm_charset_utf8_to_unpacked_gsm (const gchar *utf8,
         g_byte_array_append (gsm, (guint8 *) "\0", 1);
         if (out_len)
             *out_len = 0;
-        return g_byte_array_free (gsm, FALSE);
+        return g_byte_array_free (g_steal_pointer (&gsm), FALSE);
     }
 
     next = utf8;
@@ -461,8 +466,16 @@ mm_charset_utf8_to_unpacked_gsm (const gchar *utf8,
             /* Add the escape char */
             g_byte_array_append (gsm, &gesc, 1);
             g_byte_array_append (gsm, &gch, 1);
-        } else if (utf8_to_gsm_def_char (c, next - c, &gch))
+        } else if (utf8_to_gsm_def_char (c, next - c, &gch)) {
             g_byte_array_append (gsm, &gch, 1);
+        } else if (translit) {
+            /* add ? */
+            g_byte_array_append (gsm, &gch, 1);
+        } else {
+            g_set_error (error, MM_CORE_ERROR, MM_CORE_ERROR_INVALID_ARGS,
+                         "Couldn't convert UTF-8 char to GSM");
+            return NULL;
+        }
 
         c = next;
     }
@@ -473,7 +486,7 @@ mm_charset_utf8_to_unpacked_gsm (const gchar *utf8,
 
     /* Always make sure returned string is NUL terminated */
     g_byte_array_append (gsm, (guint8 *) "\0", 1);
-    return g_byte_array_free (gsm, FALSE);
+    return g_byte_array_free (g_steal_pointer (&gsm), FALSE);
 }
 
 static gboolean
@@ -872,7 +885,7 @@ mm_utf8_take_and_convert_to_charset (gchar          *str,
         break;
 
     case MM_MODEM_CHARSET_GSM:
-        encoded = (gchar *) mm_charset_utf8_to_unpacked_gsm (str, NULL);
+        encoded = (gchar *) mm_charset_utf8_to_unpacked_gsm (str, FALSE, NULL, NULL);
         g_free (str);
         break;
 

src/mm-charsets.h

@@ -54,7 +54,9 @@ gchar *mm_modem_charset_hex_to_utf8 (const gchar    *src,
                                     MMModemCharset  charset);
 
 guint8 *mm_charset_utf8_to_unpacked_gsm (const gchar   *utf8,
-                                         guint32       *out_len);
+                                         gboolean       translit,
+                                         guint32       *out_len,
+                                         GError       **error);
 guint8 *mm_charset_gsm_unpacked_to_utf8 (const guint8  *gsm,
                                          guint32        len,
                                          gboolean       translit,

src/mm-sms-part-3gpp.c

@@ -987,8 +987,11 @@ mm_sms_part_3gpp_get_submit_pdu (MMSmsPart *part,
         guint8 *unpacked, *packed;
         guint32 unlen = 0, packlen = 0;
 
-        unpacked = mm_charset_utf8_to_unpacked_gsm (mm_sms_part_get_text (part), &unlen);
+        unpacked = mm_charset_utf8_to_unpacked_gsm (mm_sms_part_get_text (part), FALSE, &unlen, error);
-        if (!unpacked || unlen == 0) {
+        if (!unpacked)
+            goto error;
+
+        if (unlen == 0) {
             g_free (unpacked);
             g_set_error_literal (error,
                                  MM_MESSAGE_ERROR,

src/tests/test-charsets.c

@@ -33,8 +33,9 @@ common_test_gsm7 (const gchar *in_utf8)
     g_autoptr(GError) error = NULL;
 
     /* Convert to GSM */
-    unpacked_gsm = mm_charset_utf8_to_unpacked_gsm (in_utf8, &unpacked_gsm_len);
+    unpacked_gsm = mm_charset_utf8_to_unpacked_gsm (in_utf8, FALSE, &unpacked_gsm_len, &error);
     g_assert_nonnull (unpacked_gsm);
+    g_assert_no_error (error);
     g_assert_cmpuint (unpacked_gsm_len, >, 0);
 
     /* Pack */