3 files changed, 51 insertions, 24 deletions

diff --git a/doc/idest.texi b/doc/idest.texi
index 2b65ab7..f9934c7 100644
--- a/doc/idest.texi
+++ b/doc/idest.texi
@@ -486,13 +486,13 @@ $ idest --fixup *.mp3
 @end example
 
 If the input tags also contain malformed 8-bit encodings
 (@pxref{broken 8bit encoding}), you can fix them as shown below:
 
 @example
-$ idest --broken-8bit-encoding=iso-8859-1 --fixup *.mp3
+$ idest --broken-8bit-charset=iso-8859-1 --fixup *.mp3
 @end example
 
 @node Copy
 @chapter Copying Tags Between Files
 @xopindex{copy, described}
   To copy tags from one file to another, use the @option{--copy}
diff --git a/src/idop.c b/src/idop.c
index bb9f1c3..01c6574 100644
--- a/src/idop.c
+++ b/src/idop.c
@@ -185,23 +185,42 @@ safe_id3_file_update_and_close(struct id3_file *file)
         sigprocmask(SIG_BLOCK, &set, &oldset);
         id3_file_update(file);
         id3_file_close(file);
         sigprocmask(SIG_SETMASK, &oldset, NULL);
 }
 
-static int
-frame_encoding_is_8bit(struct id3_frame const *frame)
+enum frame_recode_option {
+        FRAME_RECODE_NONE,   /* Nothing to recode */
+        FRAME_RECODE_UTF8,   /* Convert all frames to UTF-8 encoding */
+        FRAME_RECODE_BROKEN  /* Convert broken 8-bit encodings */
+};
+
+enum frame_recode_option
+frame_recode_detect(struct id3_frame const *frame)
 {
         union id3_field const *field;
         
         if (!frame)
-                return 0;
+                return FRAME_RECODE_NONE;
         field = id3_frame_field(frame, 0);
         if (!field)
-                return 0;
-        return id3_field_gettextencoding(field) == ID3_FIELD_TEXTENCODING_ISO_8859_1;
+                return FRAME_RECODE_NONE;
+        switch (id3_field_gettextencoding(field)) {
+        case ID3_FIELD_TEXTENCODING_ISO_8859_1:
+                if (broken_8bit_charset)
+                        return FRAME_RECODE_BROKEN;
+                /* fall through */
+        case ID3_FIELD_TEXTENCODING_UTF_16:
+        case ID3_FIELD_TEXTENCODING_UTF_16BE:
+                return FRAME_RECODE_UTF8;
+                
+        case ID3_FIELD_TEXTENCODING_UTF_8:
+        default:
+                break;
+        }
+        return FRAME_RECODE_NONE;
 }
 
 static int
 frame_is_genre(struct id3_frame const *frame)
 {
         if (!frame)
@@ -209,13 +228,13 @@ frame_is_genre(struct id3_frame const *frame)
         return strcmp(frame->id, ID3_FRAME_GENRE) == 0;
 }
 
 char *
 idest_ucs4_cvt(id3_ucs4_t const *ucs4, struct id3_frame const *frame)
 {
-        if (broken_8bit_charset && frame_encoding_is_8bit(frame)) {             
+        if (frame_recode_detect(frame) == FRAME_RECODE_BROKEN) {                
                 char *tempval = (char*)id3_ucs4_latin1duplicate(ucs4);
                 char *output;
                 int rc = utf8_convert(idest_conv_recode, tempval, &output);
                 free(tempval);
                 if (rc == 0) 
                         return output;
@@ -552,15 +571,12 @@ update_frames(struct id3_tag *tag)
 static void
 fixup_charset(struct id3_tag *tag)
 {
         struct id3_frame *frame;
         unsigned i;
 
-        if (!broken_8bit_charset)
-                return;
-
         for (i = 0; (frame = id3_tag_findframe(tag, NULL, i)); i++) {
                 struct ed_item itm;
                 const struct idest_frametab *ft =
                         idest_frame_lookup(frame->id);
 
                 if (!ft) {
@@ -568,27 +584,33 @@ fixup_charset(struct id3_tag *tag)
                                 error(0, 0,
                                       "%s: unsupported text frame",    
                                       frame->id);
                         continue;
                 }
 
-                if (!frame_encoding_is_8bit(frame))
-                        continue;
-                
-                ed_item_zero(&itm);
-                itm.name = xstrdup(frame->id);
-                memcpy(itm.id, frame->id, 4);
-                if (ft->decode(&itm, frame)) {
-                        error(0, 0, "%s: decoding to %s failed",
-                              frame->id, charset);
-                        continue;
+                switch (frame_recode_detect(frame)) {
+                case FRAME_RECODE_NONE:
+                        break;
+
+                case FRAME_RECODE_UTF8:
+                        if (strcasecmp(charset, "UTF-8") != 0)
+                                break;
+                case FRAME_RECODE_BROKEN:
+                        ed_item_zero(&itm);
+                        itm.name = xstrdup(frame->id);
+                        memcpy(itm.id, frame->id, 4);
+                        if (ft->decode(&itm, frame)) {
+                                error(0, 0, "%s: decoding to %s failed",
+                                      frame->id, charset);
+                                continue;
+                        }
+                        if (!input_list)
+                                input_list = ed_list_create();
+                        gl_list_add_last(input_list, ed_item_dup(&itm));
+                        ed_item_free_content(&itm);
                 }
-                if (!input_list)
-                        input_list = ed_list_create();
-                gl_list_add_last(input_list, ed_item_dup(&itm));
-                ed_item_free_content(&itm);
         }
 }
 
 void
 set_tags(const char *name)
 {
diff --git a/src/main.c b/src/main.c
index d084c5a..c83e182 100644
--- a/src/main.c
+++ b/src/main.c
@@ -194,12 +194,17 @@ main(int argc, char **argv)
 
         set_program_name(argv[0]);
 
         get_options(argc, argv);
         if (!default_version_option)
                 default_version_option = IDEST_ID3V_1|IDEST_ID3V_2;
+        if (fixup_option) {
+                if (charset)
+                        error(1, 0, "--charset cannot be used with --fixup");
+                charset = "UTF-8";
+        }
 
         argc -= optind;
         argv += optind;
 
         if (format_name) {
                 if (mode_set)