.. include:: headings.inc


.. _font encodings:

=================================================
|phoenix_title|  **Font Encodings**
=================================================


wxPython has support for multiple font encodings.

By encoding we mean here the mapping between the character codes and
the letters. Probably the most well-known encoding is (7 bit) ASCII
one which is used almost universally now to represent the letters of
the English alphabet and some other common characters. However, it is
not enough to represent the letters of foreign alphabets and here
other encodings come into play. Please note that we will only discuss
8-bit fonts here and not Unicode.

Font encoding support is ensured by several classes: `Font`     
itself, but also :ref:`wx.FontEnumerator`
and :ref:`wx.FontMapper`. :ref:`wx.Font` encoding support is reflected by a
(new) constructor parameter encoding which takes one of the following
values (elements of enumeration type :ref:`wx.FontEncoding`):

======================================== =========================================================
``wx.FONTENCODING_SYSTEM``               The default encoding of the underlying operating system (notice that this might be a "foreign" encoding for foreign versions of Windows 9x/NT).
``wx.FONTENCODING_DEFAULT``              The applications default encoding as returned by :meth:`wx.Font.GetDefaultEncoding`. On program startup, the applications default encoding is the same as ``wx.FONTENCODING_SYSTEM``, but may be changed to make all the fonts created later to use it (by default).
``wx.FONTENCODING_ISO8859_1..15``        ISO8859 family encodings which are usually used by all non-Microsoft operating systems.
``wx.FONTENCODING_KOI8``                 Standard Cyrillic encoding for the Internet (but see also ``wx.FONTENCODING_ISO8859_5`` and ``wx.FONTENCODING_CP1251``).
``wx.FONTENCODING_CP1250``               Microsoft analogue of ISO8859-2
``wx.FONTENCODING_CP1251``               Microsoft analogue of ISO8859-5
``wx.FONTENCODING_CP1252``               Microsoft analogue of ISO8859-1
======================================== =========================================================


As you may see, Microsoft's encoding partly mirror the standard
ISO8859 ones, but there are (minor) differences even between ISO8859-1
(Latin1, ISO encoding for Western Europe) and CP1251 (WinLatin1,
standard code page for English versions of Windows) and there are more
of them for other encodings.

The situation is particularly complicated with Cyrillic encodings for
which (more than) three incompatible encodings exist: KOI8 (the old
standard, widely used on the Internet), ISO8859-5 (ISO standard for
Cyrillic) and CP1251 (WinCyrillic).

This abundance of (incompatible) encodings should make it clear that
using encodings is less easy than it might seem. The problems arise
both from the fact that the standard encodings for the given language
(say Russian, which is written in Cyrillic) are different on different
platforms and because the fonts in the given encoding might just not
be installed (this is especially a problem with Unix, or, in general,
non-Win32 systems).

To clarify, the :ref:`wx.FontEnumerator` class may be used to enumerate
both all available encodings and to find the facename(s) in which the
given encoding exists. If you can find the font in the correct
encoding with :ref:`wx.FontEnumerator` then your troubles are over, but,
unfortunately, sometimes this is not enough. For example, there is no
standard way (that I know of, please tell me if you do!) to find a
font on a Windows system for KOI8 encoding (only for WinCyrillic one
which is quite different), so :ref:`wx.FontEnumerator` will never return
one, even if the user has installed a KOI8 font on his system.

To solve this problem, a :ref:`wx.FontMapper` class is provided.

This class stores the mapping between the encodings and the font face
names which support them in :ref:`wx.ConfigBase` object. Of course, it
would be fairly useless if it tried to determine these mappings by
itself, so, instead, it (optionally) asks the user and remembers his
answers so that the next time the program will automatically choose
the correct font.