charsets: use new bytearray_from_utf8() instead of byte_array_append()

7 files changed, 59 insertions, 112 deletions

diff --git a/src/mm-broadband-modem-mbim.c b/src/mm-broadband-modem-mbim.c
index 513017a8..5168b0e1 100644
--- a/src/mm-broadband-modem-mbim.c
+++ b/src/mm-broadband-modem-mbim.c
@@ -4784,13 +4784,10 @@ ussd_encode (const gchar  *command,
         packed = mm_charset_gsm_pack (gsm->data, gsm->len, 0, &packed_len);
         array = g_byte_array_new_take (packed, packed_len);
     } else {
-        g_autoptr(GError) inner_error = NULL;
-
         *scheme = MM_MODEM_GSM_USSD_SCHEME_UCS2;
-        array = g_byte_array_sized_new (strlen (command) * 2);
-        if (!mm_modem_charset_byte_array_append (array, command, MM_MODEM_CHARSET_UCS2, &inner_error)) {
-            g_set_error (error, MM_CORE_ERROR, MM_CORE_ERROR_UNSUPPORTED,
-                         "Failed to encode USSD command in UCS2 charset: %s", inner_error->message);
+        array = mm_modem_charset_bytearray_from_utf8 (command, MM_MODEM_CHARSET_UCS2, FALSE, error);
+        if (!array) {
+            g_prefix_error (error, "Failed to encode USSD command in UCS2 charset: ");
             return NULL;
         }
     }
diff --git a/src/mm-broadband-modem-qmi.c b/src/mm-broadband-modem-qmi.c
index 620f763f..4c2a1070 100644
--- a/src/mm-broadband-modem-qmi.c
+++ b/src/mm-broadband-modem-qmi.c
@@ -7256,8 +7256,8 @@ ussd_encode (const gchar                  *command,
         return (GArray *) g_steal_pointer (&barray);
     }
 
-    barray = g_byte_array_sized_new (command_len * 2);
-    if (!mm_modem_charset_byte_array_append (barray, command, MM_MODEM_CHARSET_UCS2, &inner_error)) {
+    barray = mm_modem_charset_bytearray_from_utf8 (command, MM_MODEM_CHARSET_UCS2, FALSE, &inner_error);
+    if (!barray) {
         g_set_error (error, MM_CORE_ERROR, MM_CORE_ERROR_UNSUPPORTED,
                      "Failed to encode USSD command in UCS2 charset: %s", inner_error->message);
         return NULL;
diff --git a/src/mm-broadband-modem.c b/src/mm-broadband-modem.c
index 7ba46c91..43a9e229 100644
--- a/src/mm-broadband-modem.c
+++ b/src/mm-broadband-modem.c
@@ -5877,31 +5877,31 @@ modem_3gpp_ussd_send (MMIfaceModem3gppUssd *_self,
 /* USSD Encode/Decode (3GPP/USSD interface) */
 
 static gchar *
-modem_3gpp_ussd_encode (MMIfaceModem3gppUssd  *self,
+modem_3gpp_ussd_encode (MMIfaceModem3gppUssd  *_self,
                         const gchar           *command,
                         guint                 *scheme,
                         GError               **error)
 {
-    MMBroadbandModem      *broadband = MM_BROADBAND_MODEM (self);
-    gchar                 *hex = NULL;
+    MMBroadbandModem      *self = MM_BROADBAND_MODEM (_self);
     g_autoptr(GByteArray)  ussd_command = NULL;
 
-    ussd_command = g_byte_array_new ();
-
     /* Encode to the current charset (as per AT+CSCS, which is what most modems
      * (except for Huawei it seems) will ask for. */
-    if (mm_modem_charset_byte_array_append (ussd_command, command, broadband->priv->modem_current_charset, NULL)) {
-        /* The scheme value does NOT represent the encoding used to encode the string
-         * we're giving. This scheme reflects the encoding that the modem should use when
-         * sending the data out to the network. We're hardcoding this to GSM-7 because
-         * USSD commands fit well in GSM-7, unlike USSD responses that may contain code
-         * points that may only be encoded in UCS-2. */
-        *scheme = MM_MODEM_GSM_USSD_SCHEME_7BIT;
-        /* convert to hex representation */
-        hex = mm_utils_bin2hexstr (ussd_command->data, ussd_command->len);
-    }
-
-    return hex;
+    ussd_command = mm_modem_charset_bytearray_from_utf8 (command, self->priv->modem_current_charset, FALSE, error);
+    if (!ussd_command) {
+        g_prefix_error (error, "Failed to encode USSD command: ");
+        return NULL;
+    }
+
+    /* The scheme value does NOT represent the encoding used to encode the string
+     * we're giving. This scheme reflects the encoding that the modem should use when
+     * sending the data out to the network. We're hardcoding this to GSM-7 because
+     * USSD commands fit well in GSM-7, unlike USSD responses that may contain code
+     * points that may only be encoded in UCS-2. */
+    *scheme = MM_MODEM_GSM_USSD_SCHEME_7BIT;
+
+    /* convert to hex representation */
+    return (gchar *) mm_utils_bin2hexstr (ussd_command->data, ussd_command->len);
 }
 
 static gchar *
diff --git a/src/mm-charsets.c b/src/mm-charsets.c
index ec5bb404..b0f4ea60 100644
--- a/src/mm-charsets.c
+++ b/src/mm-charsets.c
@@ -96,40 +96,6 @@ charset_iconv_from (MMModemCharset charset)
     return settings ? settings->iconv_name : NULL;
 }
 
-static const gchar *
-charset_iconv_to (MMModemCharset charset)
-{
-    return charset_iconv_from (charset);
-}
-
-gboolean
-mm_modem_charset_byte_array_append (GByteArray      *array,
-                                    const gchar     *utf8,
-                                    MMModemCharset   charset,
-                                    GError         **error)
-{
-    g_autofree gchar *converted = NULL;
-    const gchar      *iconv_to;
-    gsize             written = 0;
-
-    g_return_val_if_fail (array != NULL, FALSE);
-    g_return_val_if_fail (utf8 != NULL, FALSE);
-
-    iconv_to = charset_iconv_to (charset);
-    g_assert (iconv_to);
-
-    converted = g_convert (utf8, -1, iconv_to, "UTF-8", NULL, &written, error);
-    if (!converted) {
-        g_prefix_error (error, "Failed to convert '%s' to %s character set",
-                        utf8, iconv_to);
-        return FALSE;
-    }
-
-    g_byte_array_append (array, (const guint8 *) converted, written);
-
-    return TRUE;
-}
-
 gchar *
 mm_modem_charset_byte_array_to_utf8 (GByteArray     *array,
                                      MMModemCharset  charset)
diff --git a/src/mm-charsets.h b/src/mm-charsets.h
index 1f635aa1..8300ad6c 100644
--- a/src/mm-charsets.h
+++ b/src/mm-charsets.h
@@ -37,15 +37,6 @@ MMModemCharset  mm_modem_charset_from_string (const gchar    *string);
 
 /*****************************************************************************************/
 
-/* Append the given string to the given byte array but re-encode it
- * into the given charset first.  The original string is assumed to be
- * UTF-8 encoded.
- */
-gboolean mm_modem_charset_byte_array_append (GByteArray      *array,
-                                             const gchar     *utf8,
-                                             MMModemCharset   charset,
-                                             GError         **error);
-
 /* Take a string encoded in the given charset in binary form, and
  * convert it to UTF-8. */
 gchar *mm_modem_charset_byte_array_to_utf8 (GByteArray     *array,
diff --git a/src/mm-sms-part-3gpp.c b/src/mm-sms-part-3gpp.c
index 4bd22626..f467c97b 100644
--- a/src/mm-sms-part-3gpp.c
+++ b/src/mm-sms-part-3gpp.c
@@ -1023,10 +1023,9 @@ mm_sms_part_3gpp_get_submit_pdu (MMSmsPart *part,
         g_autoptr(GByteArray) array = NULL;
         g_autoptr(GError)     inner_error = NULL;
 
-        /* Try to guess a good value for the array */
-        array = g_byte_array_sized_new (strlen (mm_sms_part_get_text (part)) * 2);
         /* Always assume UTF-16 instead of UCS-2! */
-        if (!mm_modem_charset_byte_array_append (array, mm_sms_part_get_text (part), MM_MODEM_CHARSET_UTF16, &inner_error)) {
+        array = mm_modem_charset_bytearray_from_utf8 (mm_sms_part_get_text (part), MM_MODEM_CHARSET_UTF16, FALSE, &inner_error);
+        if (!array) {
             g_set_error (error,
                          MM_MESSAGE_ERROR,
                          MM_MESSAGE_ERROR_INVALID_PDU_PARAMETER,
diff --git a/src/mm-sms-part-cdma.c b/src/mm-sms-part-cdma.c
index befc557c..d79a0f91 100644
--- a/src/mm-sms-part-cdma.c
+++ b/src/mm-sms-part-cdma.c
@@ -1368,57 +1368,49 @@ write_bearer_data_message_identifier (MMSmsPart  *part,
     return TRUE;
 }
 
-static void
+static GByteArray *
 decide_best_encoding (const gchar *text,
                       gpointer     log_object,
-                      GByteArray **out,
                       guint       *num_fields,
                       guint       *num_bits_per_field,
-                      Encoding    *encoding)
+                      Encoding    *encoding,
+                      GError     **error)
 {
-    guint ascii_unsupported = 0;
-    guint i;
-    guint len;
-    g_autoptr(GError) error = NULL;
+    g_autoptr(GByteArray) barray = NULL;
+    MMModemCharset        target_charset = MM_MODEM_CHARSET_UNKNOWN;
+    guint                 len;
 
     len = strlen (text);
 
-    /* Check if we can do ASCII-7 */
-    for (i = 0; i < len; i++) {
-        if (text[i] & 0x80) {
-            ascii_unsupported++;
-            break;
-        }
+    if (mm_charset_can_convert_to (text, MM_MODEM_CHARSET_IRA))
+        target_charset = MM_MODEM_CHARSET_IRA;
+    else if (mm_charset_can_convert_to (text, MM_MODEM_CHARSET_8859_1))
+        target_charset = MM_MODEM_CHARSET_8859_1;
+    else
+        target_charset = MM_MODEM_CHARSET_UCS2;
+
+    barray = mm_modem_charset_bytearray_from_utf8 (text, target_charset, FALSE, error);
+    if (!barray) {
+        g_prefix_error (error, "Couldn't decide best encoding: ");
+        return NULL;
     }
 
-    /* If ASCII-7 already supported, done we are */
-    if (!ascii_unsupported) {
-        *out = g_byte_array_sized_new (len);
-        g_byte_array_append (*out, (const guint8 *)text, len);
+    if (target_charset == MM_MODEM_CHARSET_IRA) {
         *num_fields = len;
         *num_bits_per_field = 7;
         *encoding = ENCODING_ASCII_7BIT;
-        return;
-    }
-
-    /* Check if we can do Latin encoding */
-    if (mm_charset_can_convert_to (text, MM_MODEM_CHARSET_8859_1)) {
-        *out = g_byte_array_sized_new (len);
-        if (!mm_modem_charset_byte_array_append (*out, text, MM_MODEM_CHARSET_8859_1, &error))
-            mm_obj_warn (log_object, "failed to convert to latin encoding: %s", error->message);
-        *num_fields = (*out)->len;
+    } else if (target_charset == MM_MODEM_CHARSET_8859_1) {
+        *num_fields = barray->len;
         *num_bits_per_field = 8;
         *encoding = ENCODING_LATIN;
-        return;
-    }
+    } else if (target_charset == MM_MODEM_CHARSET_UCS2) {
+        *num_fields = barray->len / 2;
+        *num_bits_per_field = 16;
+        *encoding = ENCODING_UNICODE;
+    } else
+        g_assert_not_reached ();
 
-    /* If no Latin and no ASCII, default to UTF-16 */
-    *out = g_byte_array_sized_new (len * 2);
-    if (!mm_modem_charset_byte_array_append (*out, text, MM_MODEM_CHARSET_UCS2, &error))
-        mm_obj_warn (log_object, "failed to convert to UTF-16 encoding: %s", error->message);
-    *num_fields = (*out)->len / 2;
-    *num_bits_per_field = 16;
-    *encoding = ENCODING_UNICODE;
+    return g_steal_pointer (&barray);
 }
 
 static gboolean
@@ -1462,12 +1454,14 @@ write_bearer_data_user_data (MMSmsPart  *part,
 
     /* Text or Data */
     if (text) {
-        decide_best_encoding (text,
-                              log_object,
-                              &converted,
-                              &num_fields,
-                              &num_bits_per_field,
-                              &encoding);
+        converted = decide_best_encoding (text,
+                                          log_object,
+                                          &num_fields,
+                                          &num_bits_per_field,
+                                          &encoding,
+                                          error);
+        if (!converted)
+            return FALSE;
         aux = (const GByteArray *)converted;
     } else {
         aux = data;