kiwix-lib 3.1.1
===============

 * The OPDS feed book's date must be the date of the book, not the date of the
 feed generation.
 * Convert the standard opds date to our format (YYYY-MM-DD)
 * Remove duplicate language attribute in the libxml dumper.
 * Create the datadirectory to not fail to write a file in a non-existent
   directory

kiwix-lib 3.1.0
===============

 * Add a method to get the favicon url of book (if available).
 * Move dump code of library.xml in a specific class.
 * Add a first support to bookmarks

kiwix-lib 3.0.3
===============

 * Add the 'en' language to the mapping alpha2-code ('en') to alpha3-code
 ('eng').
 * Correctly write the 'ArticleCount' and 'MediaCount' in the library.xml.
 * Correctly fill the book size for the zim file size.
 * Fix launch of aria2c.

kiwix-lib 3.0.2
===============

 * Use the correct path separator when computing relativePath on Windows.

kiwix-lib 3.0.1
===============

 * Small fix about parsing the opdsStream.

kiwix-lib 3.0.0
===============

 * Change the downloader to use aria2 using a separated process (with rpc)
   instead of using the libaria2. This simplify a lot the link process to
   libaria2 on Windows.
    - kiwix-lib doesn't depend on libaria2 anymore.
    - kiwix-lib now depends on libcurl.
 * [API break] Library class API has been updated :
    - Books are referenced by id, not index. A lot of methods have been
      updated this way.
    - Books "list" is now private.
    - There is no more "current" book.
    - listBooksIds's filters have been updated.
 * [API break] Book class API has been updated :
    - Move the definition of Book in `book.h`.
    - Use getter/setter methods instead public members.
    - Size (getSize/setSize) is now returned in bytes, not kB.
    - Dependending of how the book has been initialized (opdsfeed), the
      faviconUrl may be stored in the book, the favicon being downloaded when
      using `getFavicon`.
    - The path (and indexPath) are always absolute path.
    - Book has now a downloadId, corresponding to the aria2 download id (if
      exists)
  * [API break] Manager class API has been updated :
    - The manager is mainly use to fill a Libray from a "library.xml" file or
      opds feed. Other operations (has removeBookById, setBookPath, filter, ...)
      have been removed.
    - The manager use a intermediate class (LibraryManipulator) to add book to
      the library. This dependency injection allow caller code to hook the add
      of a book to the library.
    - The manager work on a existing Library. It doesn't how a internal
      Library.
  * [API break] OpdsDumper class API has been updated :
    - dumpOPDSFeed method now take the list of bookIds to dump instead of
      dumping all books in the library.
    - OpdsDumper can now dump openSearch result information (total result
      count, start index, ...).
  * [API break] Common tools API has been updated :
    - `base64_encode` and `base64_decode` take std::string as arguments.
    - New `download` function in networkTools.h using libcurl.
    - New `getDataDirectory` function in pathTools.
    - Better `beautifyInteger` and `beautifyFileSize` functions.
    - New `nodeToString` function serializing a pugi::xml_node to a string.
    - New `converta2toa3` function to convert alpha2 language code to aplha3
      language code.


kiwix-lib 2.0.2
===============

 * [Android] Forward c++ errors message de Java world.
 * Follow redirection of favicon.
 * Make aria2 dependency optional.
 * Inculde unistd.h only on unix platform.

kiwix-lib 2.0.1
===============

 * Fix parsing of url.
 * Remove unused static resources.
 * Correctly decode reserved characters in URLs.
 * Explicitly use icu namespace to allow use of packaged icu lib.

kiwix-lib 2.0.0
===============

 * Introduce a new API to retrive content from a reader.
   * Introduce the `Entry` class.
   * Reader's methods return an `Entry`.
   * Content and other information can be retrieved from the `Entry`.
   * Older Reader's methods are depreciated.
 * Add an `OPDSDumper` class to dump a whole `Library` as an OPDS feed.
 * Add a tool function to get the content of a file.
 * Add a tool function to create a tempory directory.
 * Add a `Downloader` class to download a file.
 * Allow the manager to populate a `Library` from an OPDS feed.
 * Try to locate libctpp2 in default system libdir and then fallback in 'lib'
   directory.
 * Build kiwix-lib setting RPATH.
 * Build kiwix-lib without warning (werror=true)
 * Build kiwix-lib on macos.

kiwix-lib 1.1.1
===============

 * Correct the name of kiwix-lib (from `kiwixlib`) in meson.build to generate
   dist archive with the correct name.
 * Libzim version need to be at least 3.2.0

kiwix-lib 1.1.0
===============

 * Allow for more than 70 search result per page in html results rendering
   (kiwix/kiwix-tools#92)
 * Add a small api to do geo queries.
 * Add multi-search support in the JNI (#67)
 * Add an API to get only one part of an article.
 * Add an API to get direct location of an article content in the zim file.
 * Improve urlencoding
 * Fix pagination in html results rendering.
 * Compile using gcc-5 on Travis.
 * Allow JNI to access search snippets.
 * JNI throw an exception instead of returning an invalid object if something
   goes wrong.
 * Add doctext documentation. (#116)
 * Various bug fixes.

kiwix-lib 1.0.0
===============

 * Correctly regenerate template resource using cttp2c at compilation time.
 * Suggestion use xapian database when available
 * Support multi-zim search in kiwix-lib (a search can now search on several
   embedded database in zims in the same time)
 * Fix some wording
 * Fix license issues
 * Add out argument to jni getContent* method to get the title of article in
   the same time we get the content
 * Rename `compile_resources.py` script to `kiwix-compile-resources`
 * Use static lib when building for android or in "static mode"
 * Make the ResourceNotFound exception public

kiwix-lib 0.2.0
===============

 * Generate the snippet from the article content if the snippet is not
 directly in the database.
   This provide better snippets as they now depending of the query.
 * Use the stopwords and the language stored in the fulltext index database to
 parse the user query.
 * Remove the indexer functionnality.
 * Move to C++11 standard.
 * Use the fulltext search of the zimlib.
   We still have the fulltext search code in kiwix-lib to be able to search in
   fulltext index by side of a zim file. (To be remove in the future)
 * Few API hanges
   * Change a lot of `Reader` methods to const methods.
 * Fix some crashes.
