Add some more string encodings.

Add an encoding for "unpacked" 3GPP TS 23.038 7-bit strings, in which
each code position is in a byte of its own, rather than with the code
positions packed into 7 bits.  Rename the packed encoding to explicitly
indicate that it's packed.

Add an encoding for ETSI TS 102 221 Annex A strings.

Use the new encodings.

8 files changed, 37 insertions, 107 deletions

diff --git a/epan/dissectors/packet-ansi_637.c b/epan/dissectors/packet-ansi_637.c
index c6a84a6827..97a760e5ed 100644
--- a/epan/dissectors/packet-ansi_637.c
+++ b/epan/dissectors/packet-ansi_637.c
@@ -451,7 +451,7 @@ text_decoder(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree, guint32 offset
         offset = 0;
         bit = fill_bits;
 
-        proto_tree_add_ts_23_038_7bits_item(tree, hf_index, tvb_out, (offset << 3) + bit, num_fields);
+        proto_tree_add_ts_23_038_7bits_packed_item(tree, hf_index, tvb_out, (offset << 3) + bit, num_fields);
         break;
 
     case 0x10: /* KSC5601 (Korean) */
diff --git a/epan/dissectors/packet-cell_broadcast.c b/epan/dissectors/packet-cell_broadcast.c
index afd86f8040..6914be47d1 100644
--- a/epan/dissectors/packet-cell_broadcast.c
+++ b/epan/dissectors/packet-cell_broadcast.c
@@ -240,7 +240,7 @@ tvbuff_t * dissect_cbs_data(guint8 sms_encoding, tvbuff_t *tvb, proto_tree *tree
    switch(sms_encoding){
    case SMS_ENCODING_7BIT:
    case SMS_ENCODING_7BIT_LANG:
-      text = tvb_get_ts_23_038_7bits_string(pinfo->pool, tvb, offset<<3, (length*8)/7);
+      text = tvb_get_ts_23_038_7bits_string_packed(pinfo->pool, tvb, offset<<3, (length*8)/7);
       tvb_out = tvb_new_child_real_data(tvb, text, (guint)strlen(text), (guint)strlen(text));
       add_new_data_source(pinfo, tvb_out, "unpacked 7 bit data");
       break;
diff --git a/epan/dissectors/packet-etsi_card_app_toolkit.c b/epan/dissectors/packet-etsi_card_app_toolkit.c
index 5da5e75526..1c0f97e8bb 100644
--- a/epan/dissectors/packet-etsi_card_app_toolkit.c
+++ b/epan/dissectors/packet-etsi_card_app_toolkit.c
@@ -997,89 +997,25 @@ typedef struct {
 	cat_nmr_type nmr_type;
 } cat_transaction_t;
 
+/*
+ * ETSI TS 102 221 Annex A.
+ */
 static void
 dissect_cat_efadn_coding(tvbuff_t *tvb, proto_tree *tree, guint32 pos, guint32 len, int hf_entry)
 {
 	if (len) {
-		guint32 i;
-
 		guint8 first_byte = tvb_get_guint8(tvb, pos);
+
 		if ((first_byte & 0x80) == 0) {
-			wmem_strbuf_t *strbuf = wmem_strbuf_sized_new(wmem_packet_scope(), len+1, 0);
-			for (i = 0; i < len; i++) {
-				guint8 gsm_chars[2];
-				gsm_chars[0] = tvb_get_guint8(tvb, pos+i);
-				if (gsm_chars[0] == 0x1b) {
-					/* Escape character */
-					guint8 second_byte;
-					i++;
-					second_byte = tvb_get_guint8(tvb, pos+i);
-					gsm_chars[0] |= second_byte << 7;
-					gsm_chars[1] = second_byte >> 1;
-					wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 2));
-				} else {
-					wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 1));
-				}
-			}
-			proto_tree_add_string(tree, hf_entry, tvb, pos, len, wmem_strbuf_finalize(strbuf));
-		} else if (first_byte == 0x80) {
-			proto_tree_add_item(tree, hf_entry, tvb, pos+1, len-1, ENC_UCS_2|ENC_BIG_ENDIAN);
-		} else if (first_byte == 0x81) {
-			guint8 string_len = tvb_get_guint8(tvb, pos+1);
-			guint16 ucs2_base = tvb_get_guint8(tvb, pos+2) << 7;
-			wmem_strbuf_t *strbuf = wmem_strbuf_sized_new(wmem_packet_scope(), 2*string_len+1, 0);
-			for (i = 0; i < string_len; i++) {
-				guint8 byte = tvb_get_guint8(tvb, pos+3+i);
-				if ((byte & 0x80) == 0) {
-					guint8 gsm_chars[2];
-					gsm_chars[0] = byte;
-					if (gsm_chars[0] == 0x1b) {
-						/* Escape character */
-						guint8 second_byte;
-						i++;
-						second_byte = tvb_get_guint8(tvb, pos+3+i);
-						gsm_chars[0] |= second_byte << 7;
-						gsm_chars[1] = second_byte >> 1;
-						wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 2));
-					} else {
-						wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 1));
-					}
-				} else {
-					guint8 ucs2_char[2];
-					ucs2_char[0] = ucs2_base >> 8;
-					ucs2_char[1] = (ucs2_base & 0xff) + (byte & 0x7f);
-					wmem_strbuf_append(strbuf, get_ucs_2_string(wmem_packet_scope(), ucs2_char, 2, ENC_BIG_ENDIAN));
-				}
-			}
-			proto_tree_add_string(tree, hf_entry, tvb, pos, len, wmem_strbuf_finalize(strbuf));
-		} else if (first_byte == 0x82) {
-			guint8 string_len = tvb_get_guint8(tvb, pos+1);
-			guint16 ucs2_base = tvb_get_ntohs(tvb, pos+2);
-			wmem_strbuf_t *strbuf = wmem_strbuf_sized_new(wmem_packet_scope(), 2*string_len+1, 0);
-			for (i = 0; i < string_len; i++) {
-				guint8 byte = tvb_get_guint8(tvb, pos+4+i);
-				if ((byte & 0x80) == 0) {
-					guint8 gsm_chars[2];
-					gsm_chars[0] = byte;
-					if (gsm_chars[0] == 0x1b) {
-						/* Escape character */
-						guint8 second_byte;
-						i++;
-						second_byte = tvb_get_guint8(tvb, pos+4+i);
-						gsm_chars[0] |= second_byte << 7;
-						gsm_chars[1] = second_byte >> 1;
-						wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 2));
-					} else {
-						wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 1));
-					}
-				} else {
-					guint8 ucs2_char[2];
-					ucs2_char[0] = ucs2_base >> 8;
-					ucs2_char[1] = (ucs2_base & 0xff) + (byte & 0x7f);
-					wmem_strbuf_append(strbuf, get_ucs_2_string(wmem_packet_scope(), ucs2_char, 2, ENC_BIG_ENDIAN));
-				}
-			}
-			proto_tree_add_string(tree, hf_entry, tvb, pos, len, wmem_strbuf_finalize(strbuf));
+			/*
+			 * Unpacked GSM alphabet.
+			 */
+			proto_tree_add_item(tree, hf_entry, tvb, pos, len, ENC_3GPP_TS_23_038_7BITS_UNPACKED|ENC_NA);
+		} else {
+			/*
+			 * Annex A.
+			 */
+			proto_tree_add_item(tree, hf_entry, tvb, pos, len, ENC_ETSI_TS_102_221_ANNEX_A);
 		}
 	}
 }
diff --git a/epan/dissectors/packet-gmr1_rr.c b/epan/dissectors/packet-gmr1_rr.c
index 52fb29a5d7..00df109157 100644
--- a/epan/dissectors/packet-gmr1_rr.c
+++ b/epan/dissectors/packet-gmr1_rr.c
@@ -652,7 +652,7 @@ GMR1_IE_FUNC(gmr1_ie_rr_pos_display)
 	txt_packed_tvb = tvb_new_real_data(txt_packed, 11, 11);
 
 	/* Unpack text */
-	txt_unpacked = tvb_get_ts_23_038_7bits_string(wmem_packet_scope(), txt_packed_tvb, 0, 12);
+	txt_unpacked = tvb_get_ts_23_038_7bits_string_packed(wmem_packet_scope(), txt_packed_tvb, 0, 12);
 	tvb_free(txt_packed_tvb);
 
 	/* Display it */
diff --git a/epan/dissectors/packet-gsm_a_dtap.c b/epan/dissectors/packet-gsm_a_dtap.c
index cb7dd3a0b0..706b069e0f 100644
--- a/epan/dissectors/packet-gsm_a_dtap.c
+++ b/epan/dissectors/packet-gsm_a_dtap.c
@@ -938,7 +938,7 @@ de_network_name(tvbuff_t *tvb, proto_tree *tree, packet_info *pinfo, guint32 off
         {
             expert_add_info(pinfo, item, &ei_gsm_a_dtap_text_string_not_multiple_of_7);
         }
-        proto_tree_add_ts_23_038_7bits_item(tree, hf_gsm_a_dtap_text_string, tvb, curr_offset<<3, num_text_bits/7);
+        proto_tree_add_ts_23_038_7bits_packed_item(tree, hf_gsm_a_dtap_text_string, tvb, curr_offset<<3, num_text_bits/7);
         break;
     case 1:
         proto_tree_add_item(tree, hf_gsm_a_dtap_text_string, tvb, curr_offset, len - 1, ENC_UCS_2|ENC_BIG_ENDIAN);
diff --git a/epan/dissectors/packet-gsm_sms.c b/epan/dissectors/packet-gsm_sms.c
index 922f1b2d30..e82976e25d 100644
--- a/epan/dissectors/packet-gsm_sms.c
+++ b/epan/dissectors/packet-gsm_sms.c
@@ -645,7 +645,7 @@ dis_field_addr(tvbuff_t *tvb, packet_info* pinfo, proto_tree *tree, guint32 *off
     {
     case 0x05: /* "Alphanumeric (coded according to 3GPP TS 23.038 GSM 7-bit default alphabet)" */
         addrlength = (addrlength << 2) / 7;
-        addrstr = tvb_get_ts_23_038_7bits_string(wmem_packet_scope(), tvb, offset << 3,
+        addrstr = tvb_get_ts_23_038_7bits_string_packed(wmem_packet_scope(), tvb, offset << 3,
                                                  (addrlength > MAX_ADDR_SIZE) ? MAX_ADDR_SIZE : addrlength);
         break;
     default:
@@ -2131,29 +2131,23 @@ dis_field_ud(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree, guint32 offset
     {
         if (data && data->stk_packing_required)
         {
-            /* STK requires SMS packing by the terminal */
-            /* Per 3GPP 31.111 chapter 6.4.10: */
-            /* It shall use the SMS default 7-bit coded alphabet */
-            /* as defined in TS 23.038 with bit 8 set to 0 */
+            /*
+             * STK requires SMS packing by the terminal; this means
+             * that the string here is *not* packet 7 bits per
+             * character, but is unpacked, with each character in
+             * an octet, with the expectation that the recipient
+             * will pack it before sending it on the network.
+             *
+             * Per 3GPP 31.111 chapter 6.4.10:
+             * It shall use the SMS default 7-bit coded alphabet
+             * as defined in TS 23.038 with bit 8 set to 0
+             *
+             * I.e., bit 8 of each octet should be 0.
+             */
             if(!(reassembled && pinfo->num == reassembled_in))
             {
-                wmem_strbuf_t *strbuf = wmem_strbuf_sized_new(wmem_packet_scope(), length+1, 0);
-                for (i = 0; i < length; i++) {
-                    guint8 gsm_chars[2];
-                    gsm_chars[0] = tvb_get_guint8(tvb, offset+i);
-                    if (gsm_chars[0] == 0x1b) {
-                            /* Escape character */
-                            guint8 second_byte;
-                            i++;
-                            second_byte = tvb_get_guint8(tvb, offset+i);
-                            gsm_chars[0] |= second_byte << 7;
-                            gsm_chars[1] = second_byte >> 1;
-                            wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 2));
-                    } else {
-                            wmem_strbuf_append(strbuf, get_ts_23_038_7bits_string(wmem_packet_scope(), gsm_chars, 0, 1));
-                    }
-                }
-                proto_tree_add_string(subtree, hf_gsm_sms_text, tvb, offset, length, wmem_strbuf_finalize(strbuf));
+                proto_tree_add_item(subtree, hf_gsm_sms_text, tvb, offset,
+                                    length, ENC_3GPP_TS_23_038_7BITS_UNPACKED);
             }
             else
             {
@@ -2181,7 +2175,7 @@ dis_field_ud(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree, guint32 offset
             if(!(reassembled && pinfo->num == reassembled_in))
             {
                 /* Show unassembled SMS */
-                proto_tree_add_ts_23_038_7bits_item(subtree, hf_gsm_sms_text, tvb, (offset<<3)+fill_bits,
+                proto_tree_add_ts_23_038_7bits_packed_item(subtree, hf_gsm_sms_text, tvb, (offset<<3)+fill_bits,
                                                     (udl > SMS_MAX_MESSAGE_SIZE ? SMS_MAX_MESSAGE_SIZE : udl));
             }
             else
@@ -2202,7 +2196,7 @@ dis_field_ud(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree, guint32 offset
                                                                          &frag_params_key);
 
                     if (p_frag_params) {
-                        proto_tree_add_ts_23_038_7bits_item(subtree, hf_gsm_sms_text, sm_tvb,
+                        proto_tree_add_ts_23_038_7bits_packed_item(subtree, hf_gsm_sms_text, sm_tvb,
                             (total_sms_len<<3)+p_frag_params->fill_bits,
                             (p_frag_params->udl > SMS_MAX_MESSAGE_SIZE ? SMS_MAX_MESSAGE_SIZE : p_frag_params->udl));
 
diff --git a/epan/dissectors/packet-mbim.c b/epan/dissectors/packet-mbim.c
index 22841f05c1..7ac2249f82 100644
--- a/epan/dissectors/packet-mbim.c
+++ b/epan/dissectors/packet-mbim.c
@@ -3833,7 +3833,7 @@ static void mbim_decode_sms_cdma_text(tvbuff_t *tvb, proto_tree *tree, const int
             proto_tree_add_item(tree, hfindex, tvb, offset, size_in_bytes, ENC_ISO_8859_1|ENC_NA);
             break;
         case MBIM_ENCODING_GSM_7BIT:
-            proto_tree_add_ts_23_038_7bits_item(tree, hfindex, tvb, (offset << 3), size_in_chars);
+            proto_tree_add_ts_23_038_7bits_packed_item(tree, hfindex, tvb, (offset << 3), size_in_chars);
             break;
         default:
             break;
diff --git a/epan/dissectors/packet-nas_eps.c b/epan/dissectors/packet-nas_eps.c
index 712ceb1f66..496b1d604c 100644
--- a/epan/dissectors/packet-nas_eps.c
+++ b/epan/dissectors/packet-nas_eps.c
@@ -2356,7 +2356,7 @@ de_emm_ext_emerg_num_list(tvbuff_t *tvb, proto_tree *tree, packet_info *pinfo _U
                                      tvb, curr_offset, 1, ENC_NA, &length);
         curr_offset++;
         if (length > 0) {
-            proto_tree_add_ts_23_038_7bits_item(sub_tree, hf_eps_emm_ext_emerg_num_list_sub_serv_field,
+            proto_tree_add_ts_23_038_7bits_packed_item(sub_tree, hf_eps_emm_ext_emerg_num_list_sub_serv_field,
                                                 tvb, curr_offset<<3, (length<<3)/7);
             curr_offset += length;
         }