/rusxmms/librcc

RusXMMS Project

===============

Originally, the project was aimed to provide means to work with multiple encodings 

of the same language through adapting encoding of ID3 tags, M3U and PLS playlists 

(including file names) to local settings on-the-fly. Both the tag reading and 

writing back using any selected encoding was supported.

Nowadays there are library available providing the same functionality for almost

any program with just a few lines of code. The library is not limited to ID3 tags,

it can be useful for any program working with small titles or file names in 

different languages and encodings. The patches for several music players, ID3 tag

libraries and some other programs are available on the project page.

The Abilities of LibRCC Library

===============================

    * Language Autodetection

    * On the fly translation between languages, using online-services!

    * Encoding Autodetection for most of European Languages1

    * Support for encoding detection plugins (besides Enca and LibRCD)

    * Recoding/translation of multi-language playlists!

    * Cache to speed-up re-recoding.

    * Possibility to configure new languages and encodings.

    * Shared configuration file. For example mentioned TagLib and LibID3 patches

    do not have their own user interface, but will utilize the same recoding 

    configuration as XMMS.

    * As well the separate program for configuration adjustment is available.

    * GTK/GTK2 UI Library: you can add properties page to your GTK application 

    with 3 lines of code.

    * Menu localization opportunity

The Available Patches

=====================

    * RusXMMS: Visualization and editing of the whole range of ID3 tags using any

    of eight byte or unicode encodings. Support for playlists with non-english 

    filenames. The translation of foreign languages to english or locale one is 

    supported as well. The embeded properties page. The patch makes XMMS the best 

    player to  work with ID3 titles.

    * TagLib: Visualization and editing of ID3 v.1 and v.2 titles. Any TagLib 

    based application will correctly work with ID3 tags out of the box. The 

    properties page can be added to application with several lines of code. 

    Additionaly, after applying the patch, the 'tagwriter' program from the 

    TagLib examples can be used to convert titles of all your MP3 files to 

    unicode ID3 v.2 tags just using command: "tagwriter *.mp3".

    * LibID3: Visualization of ID3 v.1 and v.2 titles. Any LibID3 based 

    application will correctly display (but not edit) ID3 tags out of the box.

    * Mpg123: Visualization of ID3 v.1 titles.

    * GFtp: Recoding file names between FTP servers using different encodings.

    * Unzip: Recoding file names from Windows created archives.

Gratitudes

==========

    * Me ;)

    * Michael Shigorin - Ideas and great help in wiping bugs

    * Dmitry A. Koptev - Slackware packages

    * IPE, ForschungsZentrum, Karlsruhe

    * CRD, Yerevan Physics Institute

    * Georgian and Spanish winemakers :)

Important Notes

===============

1. It have much more sense to report problems here, then just claiming nothing is 

   working on miscellaneous forums and mailing lists.

2. If you want patches presented here to be included in the correspondent project 

   trees, please, ask authors. The same thing should be concerned about inclusion 

   of LibRCC and LibRCD in the official Gentoo portage tree.

3. Most of the patches will modify configure.in and Makefile.am files, so the full

   autoconfiguration should be performed. 

   You should run "aclocal; automake; autoconf" prior to using configure script.

4. Output encoding normally must correspond to current "LC_CTYPE" locale. If 

   you would set it to another value without really knowing what are you doing, 

   it can raise problems.

Preferences

===========

    * Current Language. The English, Russian, Ukrainian, Belarussian, Bulgarian,

    Czech, Estonian, Croatian, Hungarian, Lithuanian, Latvian, Polish, Slovak, 

    Slovenian and Chinese are embeded in the library. To get other languages you 

    should configure them in the "/etc/rcc.xml" or in user-defined configuration 

    "~/.rcc/rcc.xml". By default the language will be determined using LC_CTYPE 

    environmental variable.

    * Current encoding for supported encoding classes. For RusXMMS the following

    classes are defined:

          o ID3 Encoding

          o ID3 v.2 Encoding (uses ID3 by default)

          o PlayList Encoding (uses ID3 by default)

          o Encoding for Filenames in the Playlists (defaults to FS encoding)

          o FileSystem Encoding (uses locale encoding by default)

          o Output Encoding (uses locale encoding by default)

    The default encoding will be resolved using:

          o The unicode encoding selected for english language.

          o The encoding of the parrent class if any.

          o The unicode encoding defined by locale variable or any locale encoding 

	  in the case of locale language is used.

          o First available usable encoding.

    * Encoding autodetection engine. First available is used by default.

    * Mode for recoding cache:

          o Off: Do not use recoding cache

          o On: Use recoding cache to find out encoding and language

          o ReLearn: Fill recoding cache with detected values

          o Learn: Try to use recoding cache to find out encoding. If there are 

	  no cached encoding for current title try to detect it and store in the 

	  cache.

    * Autodetect File Name: If option is switched on the encoding of the file 

    will be resolved using search over file system.

    * Autoengine Set Current Encoding: Encoding autodetection engine will 

    automatically set detected encoding to be used by default.

    * Autodetect Language: Try to autodetect used language. Quite slow.

    * Translate Text: Translate text from detected language to the locale 

    language. Very slow and requires internet connection. The different modes is 

    available. In the full mode the string will be translated to the current 

    locale language. If translation to locale language is failed, the string will

    be translated to english. With "Skip Related" and "Skip parrent" options the

    translation between related languages will be omited (The language is 

    considered to be parrent language in the case then it is expected to have 

    words from that language in the strings of child language. The english 

    language is considered to be parrent language for any other.). 

    With "Translate to English" option it is possible to translate all strings 

    to english.

    * Since the translation is slow, it is possible to limit maximum time used to

    recode/translate string. In this case if translation in the specified amount 

    of time is not finished, untranslated string will be returned. However, the 

    string will be queued for translation and in the next access the translated 

    and cached value will be returned.

    * Additionally, for RusXMMS only it is possible to specify font used by shade

    form of xmms playlist.

Using Multi-Language Playlist

=============================

There are two ways of using multi-language playlists. The first is to use 

Language autodetection:

    * The UTF-8 locale should be set.

    * The LibRCC should be compiled with aspell support.

    * The aspell dictionaries for all languages used should be installed.

However this is quite slow, and errnous in the cases then mostly non-dictionary 

words are used. The second option, is to use recodings cache:

    * The UTF-8 locale should be set.

    * Select "Learn" mode for recoding caching policy in the preferences prior

    to loading new files.

    * Afterwards the titles for the loaded files will be recoded correctly 

    whenever the recoding caching is enabled.

Using Language Translation

==========================

It is possible to translate titles to your languages using libtranslate. 

The LibRCC should be compiled with LibTranslate support (It will be quite good 

idea to use patched libtranslated with posibility to limit maximum amount of 

time spent for translation) and you should have internet connection. Since 

libtranslate utilizes online translation services it takes a lot of time to get

translation. To solve this problem the translation queueing and caching is used.

If translation is failed in considered amount of time error will be returned to 

caller and string will be queued for translation. The translated string will be

stored in the DB4 cache and will be returned to caller on the following requests

for translation.