* csv import filter
  */
 
+/* This is used by both allcsv_export_database() and csv_export_common()
+   to distinguish between standard and defined fields.
+   To avoid confusions this should stay > ITEM_FIELDS  */
+#define CUSTOM_FIELD_START_INDEX       (ITEM_FIELDS + 10)
+
 /* FIXME
  * these files should be parsed according to a certain
  * lay out, or the default if layout is not given, at
                        else if(CSV_IS_SPECIAL(fields[i])) {
                                if(special_func)
                                        (*special_func)(out, e.item, fields[i]);
-                       } else
+                       }
+                       else if(fields[i] >= CUSTOM_FIELD_START_INDEX) {
+                               fprintf(out, "\"%s\"",
+                                       safe_str(db_fget_byid(e.item, fields[i] - CUSTOM_FIELD_START_INDEX)));
+                       }
+                       else
                                /*fprintf(out,(
                        strchr(safe_str(database[e.item][field_idx(fields[i])]), ',') ||
                        strchr(safe_str(database[e.item][field_idx(fields[i])]), '\"')) ?
         * TODO: Should get these atomatically from abook_fileds
         *  - JH
         */
-       int allcsv_export_fields[] = {
+       int allcsv_export_fields[ITEM_FIELDS + 6] = { // only the 5 custom fields are allowed so far
                NAME,
                EMAIL,
                ADDRESS,
                PHONE,
                WORKPHONE,
                FAX,
-               MOBILEPHONE, // spelled "mobile" in standard_fields
+               MOBILEPHONE, // spelt "mobile" in standard_fields
                NICK,
                URL,
                NOTES,
        while(allcsv_export_fields[i+1] != CSV_LAST) {
                fprintf(out, "\"%s\",", standard_fields[i++].key);
        }
-       fprintf(out, "\"%s\"\n", standard_fields[i].key);
+       fprintf(out, "\"%s\"", standard_fields[i].key);
+
+       /*
+         Custom fields handling:
+         This loop appends custom fields' id at the end of allcsv_export_fields and shift
+         the CSV_LAST sentinel value each time one is found.
+         CUSTOM_FIELD_START_INDEX is added to these index values so csv_export_common()
+         can later recognize them and call db_fget_byid() instead of the traditional db_fget()
+
+         It only search for defined the [legacy?] "custom" fields.
+       */
+
+       // pointer to the end of the field list
+       int append_field = ITEM_FIELDS;
+       // custom field's trailing number (between 1 and 5)
+       int j;
+       // full custom field name, eg "custom4"
+       char custom_field_key[8];
+       // index used by custom_field_key
+       int field_no;
+       // name of the defined field <field_no> as chosen by the user
+       char *custom_field_name;
+
+       for (j = 1; j <= 5; j++) {
+               snprintf(custom_field_key, 8, "custom%d", j++);
+               if(find_declared_field(custom_field_key)) {
+                       find_field_number(custom_field_key, &field_no);
+                       get_field_info(field_no, NULL, &custom_field_name, NULL);
+                       // append the field to the list
+                       allcsv_export_fields[append_field] = field_no + CUSTOM_FIELD_START_INDEX;
+                       allcsv_export_fields[++append_field] = CSV_LAST;
+                       // print column name
+                       fprintf(out, ",\"%s\"", custom_field_name);
+               }
+       }
+       free(custom_field_name);
+       fprintf(out, "\n");
 
        csv_export_common(out, e, allcsv_export_fields, NULL);