mirrors/freetype
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

removing obsolete files:

Browse Source 
the design of FT2 has evolved considerably since
these documents were written, and it's better not to
include them in order to avoid confusion among
users of the library..
This commit is contained in:
David Turner 2000-10-03 17:07:28 +00:00
parent 59266f23c9
commit 43764232b8
12 changed files with 0 additions and 3173 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

910
docs/design/build-system.html
View File

@ -1,910 +0,0 @@
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="Author" content="David Turner">
<meta name="GENERATOR" content="Mozilla/4.5 [fr] (Win98; I) [Netscape]">
<title>FreeType 2 Internals - I/O Frames</title>
<basefont face="Georgia, Arial, Helvetica, Geneva">
<style content="text/css">
P { text-align=justify }
H1 { text-align=center }
H2 { text-align=center }
LI { text-align=justify }
</style>
</head>
<body text="#000000"
bgcolor="#FFFFFF"
link="#0000EF"
vlink="#51188E"
alink="#FF0000">
<center>
<h1>FreeType 2.0 Build System</h1></center>
<center>
<h3>
&copy; 2000 David Turner (<a href="fichier :///david@freetype.org">david@freetype.org</a>)<br>
&copy; 2000 The FreeType Development Team
(<a href="mailto:devel@freetype.org">devel@freetype.org</a>)
</h3></center>
<center><table width=650><tr><td>
<p><hr WIDTH="100%"></p>
<h2>Table of Content</h2>
<center><table><tr><td>
<p><font size="+1"><a href="#introduction">Introduction</a></font></p>
<p><font size="+1"><a href="#features">I. Features & Background</a></font></p>
<ul>
<li><a href="#features-1">1. Convenience, not Requirement</a>
<li><a href="#features-2">2. Compiler and platform independence</a>
<li><a href="#features-3">3. Uses GNU Make</a>
<li><a href="#features-4">4. Automatic host platform detection</a>
<li><a href="#features-5">5. User-selectable builds</a>
<li><a href="#features-6">6. Robustness</a>
<li><a href="#features-7">7. Simple modules management</a>
</ul>
<p><font size="+1"><a href="#overview">II. Overview of the build process</a></font></p>
<ul>
<p><li><a href="#overview-1">1. Build setup</a>
<ul>
<li><a href="#overview-1-a">a. Default build setup</a>
<li><a href="#overview-1-b">b. Selecting another build configuration</a>
</ul>
</p>
<li><a href="#overview-2">2. Library compilation</a>
</ul>
<p><font size="+1"><a href="#setup">III. Build setup details</a></font></p>
<p><font size="+1"><a href="#compilation">IV. Library compilation details</a></font></p>
<ul>
<li><a href="#compilation-1">a. Compiling the <tt>ftsystem</tt> component</a>
<li><a href="#compilation-2">b. Compiling the base layer and optional components</a>
<li><a href="#compilation-3">c. Compiling the modules</a>
<li><a href="#compilation-4">d. Compiling the <tt>ftinit</tt> component</a>
</ul>
</ul>
</td></tr></table></center>
<hr><a name="introduction">
<h2>Introduction:</h2>
<p>This document describes the new build system that was introduced
with FreeType 2.</p>
<p><hr></p>
<a name="features">
<h2>I. Features and Background:</h2>
<p>The FreeType 2 build system is a set of Makefiles and sub-Makefiles that
are used to build the library on a very large variety of systems easily.
One of its main features are the following:</p>
<a name="features-1">
<h3>1. Convenience, not Requirement</h3>
<ul>
<p>Even though the build system is rather sophisticated, it simply is a
convenience that was written simply to allow the compilation of the
FreeType 2 library on as many platforms as possible, as easily as
possible. However, it is not a requirement and the library can be
compiled manually or in a graphical IDE without using it, with minimal
efforts</p>
<p>(for more information on this topic, see the <tt>BUILD</tt>
document that comes with your release of FreeType, in its <em>Detailed
Compilation Guide</em> section).</p>
</ul>
<a name="features-2">
<h3>2. Compiler and platform independence</h3>
<ul>
<p>The FreeType 2 build system can be used with any compiler, on any platform.
It is independent of object file suffix, executable file suffix, directory
separator convention (i.e. "/" or "\"), and compiler flags for path
inclusion, macro definition, output naming, ansi compliance, etc..</p>
<p>Supporting a new compiler is trivial and only requires writing a minimal
configuration sub-makefile that contains several Makefile variables
definitions that are later used by the rest of the build system. This is
described in details later in the document.</p>
</ul>
<a name="features-3">
<h3>3. Uses GNU Make</h3>
<ul>
<p>The build system works <em>exclusively</em> with <b>GNU Make</b>. Reason
is that it is the only make utility that has all the features required to
implement the build system as described below. Moreover, it is already
ported to hundreds of various distinct platforms and is widely and
freely available.</p>
<p>It also uses the native command line shell. <em>You thus
don't need a Unix-like shell on your platform</em>.
For example, FreeType 2 already compiles on Unix, Dos, Windows
and OS/2 right "out of the box" (assuming you have GNU Make
installed).</p>
<p>Finally, note that the build system is <em>specifically</em> designed
for gnu make and will <em>fail</em> with any other make tool. We have
<em>no plans</em> to support a different tools, as you'll rapidly
understand by reading this document or looking at the sub-makefiles
themselves.</p>
</ul>
<a name="features-4">
<h3>4. Automatic host platform detection</h3>
<ul>
<p>When you launch the build system for the first time, by simply invoking
GNU make in the top-level directory, it automatically tries to detect
your current platform in order to choose the best configuration
sub-makefile available. It then displays what it found. If everything
is ok, you can then launch compilation of the library, by invoking make
a second time.</p>
<p>The following platforms are currently automatically detected:</p>
<ul>
<li>Dos (plain-dos, windows in Dos mode, or Dos session under OS/2)
<li>Windows 95, 98 + Windows NT (a.k.a win32)
<li>OS/2
<li>Unix (uses Autoconf/Automake)
</ul>
<p>Note that adding support for a new platform requires writing a minimal
number of very small files, and simply putting them in a new sub-directory
of <tt>freetype2/config</tt>.</p>
</ul>
<a name="features-5">
<h3>5. User-selectable builds</h3>
<ul>
<p>The platform auto-detection rules try to setup the build for a default
compiler (<em>gcc</em> for most platforms), with default build options
for the library (which normally is
<em>"all features enable, no debugging"</em>), as well as the default
list of modules (which is <em>"all modules in <tt>freetype2/src</tt>"</em>)</p>
<p>There are cases where it is important to specify a different compiler,
different build options or simply a different module list. The FreeType 2
build system is designed in such a way that all of this is easily possible
from the command line, <em>without having to touch a single file</em>.
The latter is crucial when dealing with projects that need specific
builds of the library without modifying a single file from the FreeType
distribution.</p>
<p>The exact mechanism and implementation to do this is described later in
this document. It allows, for example, to compile FreeType with any of
the following compilers on Win32: gcc, Visual C++, Win32-LCC.</p>
</ul>
<a name="features-6">
<h3>6. Robustness</h3>
<ul>
<p>The build system uses a single top-level Makefile that includes
one or more sub-makefiles to build the entire library (base layer
plus all modules).
<font color="red">
To understand why this is important, we <em>strongly</em> recommend
the following article to all of our readers:</font></p>
<p>
<center>
<font size="+2"><a href="http://www.pcug.org.au/~millerp/rmch/recu-make-cons-harm.html">
Recursive Make Considered Dangerous
</a>
</font>
</center>
</p>
<p>As an example, here's a short list of files that make up the
build system. Note that each sub-makefile contains rules corresponding
to a very specific purpose, and that they all use the "<tt>.mk</tt>"
suffix:</p>
<ul>
<li><tt>freetype2/Makefile</tt>
<li><tt>freetype2/config/detect.mk</tt>
<li><tt>freetype2/config/freetype.mk</tt>
<li><tt>freetype2/config/<em>&lt;system&gt;</em>/detect.mk</tt>
<li><tt>freetype2/src/<em>&lt;module&gt;</em>/rules.mk</tt>
<li><tt>freetype2/src/<em>&lt;module&gt;</em>/module.mk</tt>
</ul>
</ul>
<a name="features-7">
<h3>7. Simple Module Management</h3>
<ul>
<p>FreeType 2 has a very modular design, and is made of a core
<em>base layer</em> that provides its high-level API as well as
generic services used by one or more <em>modules</em>.
Most modules are used to support a specific font format (like TrueType
or Type 1), and they are called <em>font drivers</em>. However, some of
them do not support font files directly, but rather provide helper
services to the font drivers.</p>
<p>FreeType 2 is designed so that adding modules at run-time is possible
and easy. Similarly, we expect many more modules to come in the near
future and wanted a build system that makes such additions to the
source package itself dead easy.
Indeed, all source code (base + modules) is located in the
<tt>freetype2/src</tt> directory hierarchy. And the build system is
capable of re-generating automatically the list of known modules
from the contents of this directory. Hence, adding a new font driver
to the FreeType sources simply requires to:</p>
<ul>
<li><p>Add a new sub-directory to <tt>freetype2/src</tt>
<li><p>Re-launch the build system</p>
</ul>
<p>There is thus no need to edit a source file</p>
</ul>
<p><hr><p>
<a name="overview">
<h2>II. Overview of the build process(es):</h2>
<p>Before describing in details how the build system works, it is essential
to give a few examples of how it's used. This section presents
what's the build process is to the typical developer:</p>
<p>Compiling the library is normally done in two steps: the first one
configures the build according to the current platform and possible
additional parameters, while the second simply compiles the library with
the information gathered in the configuration step.</p>
<a name="overview-1">
<h3>1. Build Setup</h3>
<a name="overview-1-a">
<h4>a. Default build setup</h4>
<ul>
<p>To configure the build, simply invoke gnu make from the top-level FreeType
directory. This will launch a series of rules that will detect your current
host platform, and choose a configuration file for you. It will then display
what it found. For example, here's the output of typing the command "make"
on a win32 platform (assuming this calls GNU make):</p>
<pre><font color="blue">
<font color="purple">C:\FreeType> make</font>
FreeType build system -- automatic system detection
The following settings are used:
platform win32
compiler gcc
configuration directory ./config/win32
configuration rules ./config/win32/w32-gcc.mk
If this does not correspond to your system or settings please remove the file
'config.mk' from this directory then read the INSTALL file for help.
Otherwise, simply type 'make' again to build the library.
<font color="purple">C:\FreeType></font>
</font></pre>
<p>Note that this step copies the selected configuration file (here
<tt>./config/win32/w32-gcc.mk</tt>) to <em>the current directory</em>, under
the name <tt><b>config.mk</b></tt>. This file contains data that is used
to drive the library compilation of the second step. It correspond to
the platform and compiler selected by the auto-detection phase.</p>
<p>Note that you can re-generate the <tt><b>config.mk</b></tt> file anytime
by invoking <tt>make setup</tt> whenever you need it, even when the file is
already present in the current directory.</p>
<p>Finally, if your platform is not correctly detected, the build system will
display and use configuration information for the virtual "ansi" platform.
</p>
</ul>
<a name="overview-1-b">
<h4>b. Selecting another build configuration</h4>
<ul>
<p>You may not be really satisfied by the configuration file selected by the
auto-detection routines. Typically, you might be using a compiler that is
not the default one for your platform. It is however possible to re-launch
the build setup phase with an additional argument, used to specify a
different compiler/config file. For example, you can type the following
commands on Win32 systems:</p>
<p align=center><table width="80%" cellpadding=10><tr valign=top><td>
<p><b><tt>make&nbsp;setup</tt></b></p>
</td><td>
<p>re-run the platform detection phase, and select the default compiler for it.
On Win32, this is <em>gcc</em>.</p>
</td></tr><tr valign=top><td>
<p><b><tt>make&nbsp;setup&nbsp;visualc</tt></b></p>
</td><td>
<p>re-run the platform detection phase, and select a config file that
corresponds to the <em>Visual C++</em> compiler</p>
</td></tr><tr valign=top><td>
<p><b><tt>make&nbsp;setup&nbsp;lcc</tt></b></p>
</td><td>
<p>re-run the platform detection phase, and select a config file that
corresponds to the <em>Win32-LCC</em> compiler</p>
</td></tr></table>
</p>
<p>Note that a specific configuration is selected with a command that
looks like : <tt><b>make&nbsp;setup&nbsp;<em>compiler</em></b></tt>,
where the <em><tt>compiler</tt></em> keywords depends on the platform.
Moreover, each one of them corresponds to a specific configuration
sub-makefile that is copied as <b><tt>config.mk</tt></b> in the current
directory.</p>
</ul>
<a name="overview-2">
<h3>2. Library compilation</h3>
<p>Once you're satisfied with the version of <b><tt>config.mk</tt></b> that
has been copied to your current directory, you can simply re-invoke
gnu make <em>with no arguments</em>. The top-level Makefile will
automatically detect the config sub-makefile in the current directory,
and use it to drive the library compilation. The latter can be seen
as a series of different steps decribed here:</p>
<ul>
<li><p><b>Compiling the <tt>ftsystem</tt> component</b><br><ul>
It encapsulates all low-level operations (memory management +
i/o access) for the library. Its default version, located in
<tt>./src/base/ftsystem.c</tt> uses the ANSI C library but
system-specific implementations are also available to
improve performance (e.g. memory-mapped files on Unix).
</ul></p>
<li><p><b>Compiling the <em>base layer</em> and optional components</b><br><ul>
They provide the library's high-level API as well as various useful
routines for client applications. Many features of the base layer can
be activated or not depending on a configuration file named
<tt>ftoption.h</tt>
</ul></p>
<li><p><b>Compiling the <em>modules</em></b><br><ul>
Each module is used to support a specific font format (it is then
called a <em>font driver</em>), or to provide helper services to
the drivers (e.g. the auto-hinter). They are all located in
sub-directories of <tt>./src</tt>, like <tt>./src/truetype</tt>,
<tt>./src/type1</tt>.
</ul></p>
<li><p><b>Compiling the <tt>ftinit</tt> component</b><br><ul>
This one is in charge of implementing <tt>FT_Init_FreeType</tt>,
the library initialisation routine. It also selects what modules
are activated when a new library instance is created.
</ul></p>
</ul>
<p><hr><p>
<a name="setup">
<h2>II. Details of the build setup.</h2>
<p>When the top-level <tt>Makefile</tt> is invoked, it looks for a
file named <b><tt>config.mk</tt></b> in the <em>current directory</em>.
If this file is found, it is used directly to build the library
(skip to <a href="library">Section III</a> for details then).</p>
<p>Otherwise, the file <b><tt>./config/detect.mk</tt></b> is included
by the top-level <tt>Makefile</tt> and parsed. Its purpose is to drive the
platform-detection phase, by:</p>
<ul>
<li><p>Defining the <tt>PLATFORM</tt> variable, which indicates
what the currently detected platform is. It is initially
set to the default value "<tt><b>ansi</b></tt>".
</p>
<li><p>Searching for a <tt>detect.mk</tt> file in <em>all
subdirectories</em> of <b><tt>./config</tt></b>.
Each such file is included and parsed. Each of these files must
try to detect if the host platform is a system it knows
about. If so, it changes the value of the <tt>PLATFORM</tt> variable
accordingly.</p>
<li><p>Copying the selected configuration submakefile to the current directory
under the name <tt><b>config.mk</b></tt>.</p>
</ul>
<p>This is illustrated by the following graphics :</p>
<p><center>
<img src="platform-detection.png" border=0>
</center></p>
<p>Each system-specific <b><tt>detect.mk</tt></b> works as follows:</p>
<ul>
<li><p>It checks that the value of <tt>PLATFORM</tt> is currently set
to <b>ansi</b>, which indicates that no platform was detected
for now. If this isn't true, it doesn't do anything</p>
<li><p>Otherwise, it runs a series of test to see wether it is on a
system it knows about. Here are a few examples of tests:</p>
<p><center><table width="80%" cellpadding=5><tr valign=top><td>
<em><b>Unix</b></em>
</td><td>
<p>checks for a file named <tt>/sbin/init</tt>, and runs, when it found
it, a 'configure' script to generate the relevant config sub-makefile</p>
</td></tr><tr valign=top><td>
<em><b>Dos</b></em>
</td><td>
<p>checks for the <tt>COMSPEC</tt> environment variable, then tries to
run the "<tt>ver</tt>" command on the current shell to check that there
is a "Dos" substring in its output; if not, it tries to find the
substring "<tt>MDOS\COMMAND</tt>" in <tt>COMSPEC</tt>, which indicates
a Dos session under OS/2.</p>
</td></tr><tr valign=top><td>
<em><b>Win32</b></em>
</td><td>
<p>if the environment variable <tt>OS</tt> is defined and has the value
<tt>Windows_NT</tt>, or if <tt>COMSPEC</tt> is defined and the
"<tt>ver</tt>" returns a string that contains <tt>Windows</tt> in it,
we're on a Win32 system.</p>
</td></tr></table></center>
</p>
<li><p>It sets the value of <tt>PLATFORM</tt> to a new value corresponding
to its platform.</p>
<li><p>It then tries to select a configuration
sub-makefile, depending on the current platform and any optional
make target (like "visualc" or "devel", etc..). Note that it can
even generate the file, as on Unix through Autoconf/Automake.</p>
<li><p>It copies the selected configuration sub-makefile to the current
directory, under the name <tt><b>config.mk</b></tt>
</ul>
<p>If one wants to support a new platform in the build system, it simply needs
to provide:</p>
<ul>
<li>A new subdirectory, in <tt>./config</tt>, with a file named
<tt>detect.mk</tt> in it, containing relevant checks for the system.
<li>One or more configuration sub-makefiles that will get copied to
<tt>config.mk</tt> at build setup time. You can use the one in
<tt>./config/ansi/config.mk</tt> as a template.
</ul>
<p>Similary, supporting a new compiler on an existing system simply means:</p>
<ul>
<li>Writing a new config sub-makefile that contains definitions used to
specify the compiler and flags for the build.
<li>Change your <tt>./config/<em>system</em>/detect.mk</tt> to recognize
a new optional build target that will copy your new config sub-makefile
instead of the default one.
</ul>
<p><hr><p>
<h2>III. Details of the library compilation.</h2>
<p>When the top-level Makefile is invoked, it looks for a file named
<tt>config.mk</tt> in the current directory. If one is found, it
defines the <tt>BUILD_FREETYPE</tt> variable, then includes and parses it.
The structure of this file is the following:
</p>
<ul>
<li><p>First, it defines a series of Make variables that describe
the host environment, like the compiler, compilation flags,
object file suffix, the directory where all object files are
placed, etc..</p>
<li><p>If <tt>BUILD_FREETYPE</tt> is defined, it includes the file
<tt><b>./config/freetype.mk</b></tt>, which is in charge of
defining all the rules used to build the library object files.
(The test is useful to use the <tt>config.mk</tt> file to
compile other projects that rely on FreeType 2, like its
demonstration programs).</p>
<li><p>Finally, it defines the rule(s) used to link FreeType 2 object files
into a library file (e.g. <tt>libfreetype.a</tt>, <tt>freetype.lib</tt>,
<tt>freetype.dll</tt>, ...). Unfortunately, the command line interface of link tools is
a <em>lot less</em> standardized than those of compilers, which
explains why this rule must be defined in the system-specific
<tt>config.mk</tt>.</p>
</ul>
<p>The following is an explanation of what <tt><b>./config/freetype.mk</b></tt>
does to build the library objects:
</p>
<h4>a. Include paths</h4>
<ul>
<p>To avoid namespace pollution, the <tt><em>freetype</em></tt> directory prefix
is used to include all public header files of the library. This means
that a client application will typically use lines like:</p>
<pre><font color="blue">
#include &lt;freetype/freetype.h&gt;
#include &lt;freetype/ftglyph.h&gt;
</font></pre>
<p>to include one the FreeType 2 public header files. <tt>freetype.mk</tt>
uses a variable named <tt><b>INCLUDES</b></tt> to hold the inclusion
paths list, and thus starts by adding <tt>./include</tt> to it. However,
nothing prevents
<p><tt>freetype.mk</tt> uses a variable named <tt><b>INCLUDES</b></tt>
to hold directory inclusion-path to be used when compiling the library.
It always add <tt>./include</tt> to this variable, which means
</ul>
<h4>b. Configuration header files:</h4>
<ul>
<p>Three header files used to configure the compilation of the
FreeType 2 library. Their default versions are all located in the
directory <tt><b>./include/freetype/config/</b></tt>, even though
project specific versions can be provided on a given build, as
described later:</p>
<ul>
<p><b><tt>#include &lt;freetype/config/ftoption.h&gt;</tt></b><br><ul>
This file contains a set of configuration macro definitions that
can be toggled to activate or deactivate certain features of the
library. By changing one of these definitions, it is possible to
compile <em>only the features that are needed</em> for a specific
project. Note that by default, all options are enabled.
<br><br>
You might need to provide an alternative version of <tt>ftoption.h</tt>
for one of your own projects.
</ul></p>
<p><b><tt>#include &lt;freetype/config/ftconfig.h&gt;</tt></b><br><ul>
This file includes <tt>ftoption.h</tt> but also contains some automatic
macro definitions used to indicate some important system-specific
features (e.g: word size in bytes, DLL export prefix macros, etc..).
<br><br>
You shouldn't normally need to change or provide an alternative
version of this file.
</ul></p>
<p><b><tt>#include &lt;freetype/config/ftmodule.h&gt;</tt></b><br><ul>
This file is very special, as it is normally machine-generated, and
used by the <tt>ftinit</tt> component that is described below. To
understand it, one must reminds that FreeType 2 has an extremely
modular design and that it's possible to change, <em>at run-time</em>,
the modules it's using. The <tt>ftmodule.h</tt> file simply contains
the list of modules that are registered with each new instance of
the library.
<br><br>
Note that the file can be re-generated automatically by invoking
<tt>make setup</tt> from the top-level directory. The re-generated
list contains all the modules that were found in subdirectories of
<tt>./src</tt>.
</ul></p>
</ul>
<p>Note that we strongly advise you to avoid modifying the config files
within the FreeType 2 source directory hierarchy. Rather, it's possible
to specify alternative versions through the help of a build-specific
include path that is include before <tt>./include</tt> in the inclusion
path.</p>
<p>For example, imagine that your platform, named <em>foo</em>, needs a
specific version of <tt>ftoption.h</tt>
</ul>
<h4>a. Compiling the <b><tt>ftsystem</tt></b> component:</h4>
<ul>
<p>FreeType 2 encapsulates all low-level operations (i.e. memory management
and i/o access) within a single component called <tt><b>ftsystem</b></tt>.
Its default implementation uses the <em>ANSI C Library</em> and is located
in <tt>./src/base/ftsystem.c</tt>.</p>
<p>However, some alternate, system-specific, implementations of
<tt>ftsystem</tt> are provided with the library in order to support more
efficient and advanced features. As an example, the file
<tt>./config/unix/ftsystem.c</tt> is an implementation that
uses memory-mapped files rather than the slow ANSI <tt>fopen</tt>,
<tt>fread</tt> and <tt>fseek</tt>, boosting performance significantly.</p>
<p>The build system is thus capable of managing alternate implementations
of <tt>ftsystem</tt></p>
</ul>
<h4>b. Compiling the base layer and optional components:</h4>
<ul>
<p>The high-level API of the library is provided by a component called the
<em>base layer</em>, whose source is located in <tt>./src/base</tt>. This
directory also contains one or more components that are optional, i.e.
that are not required by the library but provide valuable routines to
client applications.</p>
<p>The features of the base library and other components are selected through
a single configuration file named
<tt><b>./include/freetype/config/ftoption.h</b></tt>. It contains a list
of commented configuration macro definitions, that can be toggled to
activate or de-activate a certain feature or component at build time.</p>
<p>For example, the code in <tt>./src/base/ftdebug.c</tt> will be compiled
only if one of these two macros are defined in <tt>ftoption.h</tt>:
<tt>FT_DEBUG_LEVEL_ERROR</tt> or <tt>FT_DEBUG_LEVEL_TRACE</tt></p>
</ul>
<h4>c. Compiling the modules:</h4>
<ul>
<p>Once the base layer is completed, the build system starts to compile each
additional module independently. These are simply defined as all source
code located in a sub-directory of <tt>./src</tt> that contains a file
named <tt><b>rules.</b></tt>, for example:
<tt>src/sfnt</tt>, <tt>src/truetype</tt>, <tt>src/type1</tt>, ...</p>
<p>The <tt><b>rules.</b></tt> file simply contains directives used by the
build system to compile the corresponding module into a single object
file.</p>
</ul>
<h4>d. Compiling the <b><tt>ftinit</tt></b> component:</h4>
<ul>
<p>The file <tt><b>./src/base/ftinit.c</b></tt> is special because it is used
to implement the library initialisation function <tt>FT_Init_FreeType</tt>.
</p>
</ul>
<p>Typically, you will end up with all object files, as well as the
corresponding library file, residing in the <tt>freetype2/obj</tt>
directory.</p>
<h3>1. Purpose of the configuration sub-makefile</h3>
<h3>2. Managing module dependencies</h3>
<h3>3. </h3>
<p><hr><p>
<a name="modules">
<h2>IV. Managing the modules list</h2>
<p><hr><p>
The build system features some important points, which are all detailed
in the following sections:<p>
<ul>
<li><b>Automatic host platform detection</b><br>
The first time the top <tt>Makefile</tt> is invoked, it will
run a series of rules to detect your platform. It will then
create a system-specific configuration sub-Makefile in the
current directory, called <b><tt>config.mk</tt></b>. You can now
invoke the top <tt>Makefile</tt> a second time to compile the
library directly.
<p>
The configuration sub-makefile can be regenerated any time
by invoking "<tt>make setup</tt>", which will re-run the
detection rules even if a <tt>config.mk</tt> is already present.
<p>
<li><b>User-selectable builds</b><br>
<p>
<li><b>Automatic detection of font drivers</b><br>
FreeType is made of a "base" layer that invokes several
separately-compiled modules. Each module is a given
font driver, in charge of supporting a given font format.
<p>
The list of font drivers is located in the file
"<tt>freetype2/config/<em>system</em>/ftmodule.h</tt>", however
it can be regenerated on-demand. Adding a new module to the
FreeType source tree is thus as easy as:<p>
<ul>
<li>create a new directory in "<tt>freetype2/src</tt>" and
put the new driver's source code and sub-makefiles there.
<p>
<li>invoke the top <tt>Makefile</tt> with target
"<tt>modules</tt>" (as in "<tt>make modules</tt>"),
as this will automatically regenerate the list of
available drivers by detecting the new directory and
its content.
</ul>
<p>
</ul>
</ul>
<p><hr><p>
<h2>II. Host Platform Detection</h2>
<ul>
When the top-level <tt>Makefile</tt> is invoked, it looks for a
file named <tt>config.mk</tt> in the current directory. If this
file is found, it is used to build the library
(see <a href="library">Section III</a>).
<p>
Otherwise, the file <tt>freetype2/config/detect.mk</tt> is included
and parsed. Its purpose is to:<p>
<ul>
<li>Define the <tt>PLATFORM</tt> variable, which indicates
what is the currently detected platform. It is initially
set to the default value "<tt>ansi</tt>".
<p>
<li>It searches for a <tt>detect.mk</tt> file in all
subdirectories of <tt>freetype2/config</tt>. Each such
file is included and parsed. Each of these files must
try to detect if the host platform is a system it knows
about. If so, it changes the value of the <tt>PLATFORM</tt>
accordingly.
</ul>
<p>
This is illustrated by the following graphics :<p>
<center>
<img src="platform-detection.png" border=0>
</center>
<p>
Note that each system-specific <tt>detect.mk</tt> is in charge
of copying a valid configuration makefile to the current directory
(i.e. the one where <tt>make</tt> was invoked), depending on the
current targets. For example, the Win32 <tt>detect.mk</tt> will
be able to detect a "<tt>visualc</tt>" or "<tt>lcc</tt>" target,
as described in section I. Similarly, the OS/2 <tt>detect.mk</tt>
can detect targets like "<tt>borlandc</tt>", "<tt>watcom</tt>"
or "<tt>visualage</tt>", etc..
</ul>
<p><hr><p>
<h2>III. Building the library</h2>
<ul>
When the top-level <tt>Makefile</tt> is invoked and that it finds
a <tt>config.mk</tt> file in the current directory, it defines
the variable <tt>BUILD_FREETYPE</tt>, then includes and parses the
configuration sub-makefile.
<p>
The latter defines a number of important variables that describe
the compilation process to the build system. Among other things:<p>
<ul>
<li>the extension to be used for object files and library files
(i.e. <tt>.o</tt> and <tt>.a</tt> on Unix, <tt>.obj</tt>
and <tt>.lib</tt> on Dos-Windows-OS/2, etc..).
<p>
<li>the directory where all object files will be stored
(usually <tt>freetype2/obj</tt>), as well as the one
containing the library file (usually the same as for
objects).
<p>
<li>the command line compiler, and its compilation flags for
indicating a new include path (usually "<tt>-I</tt>"),
a new macro declaration (usually "<tt>-D</tt>") or
the target object file (usually "<tt>-o&nbsp;</tt>")
</ul>
<p>
Once these variable are defined, <tt>config.mk</tt> test for the
definition of the <tt>BUILD_FREETYPE</tt> variable. If it exists,
the makefile then includes "<tt>freetype2/config/freetype.mk</tt>"
which contains the rules required to compile the library.
<p>
Note that <tt>freetype.mk</tt> also scans the subdirectories of
"<tt>freetype2/src</tt>" for a file called "<tt>rules.mk</tt>".
Each <tt>rules.mk</tt> contains, as it names suggests, the rules
required to compile a given font driver or module.
<p>
Once all this parsing is done, the library can be compiled. Usually,
each font driver is compiled as a standalone object file (e.g.
<tt>sfnt.o</tt>, <tt>truetype.o</tt> and <tt>type1.o</tt>).
<p>
This process can be illustrated by the following graphics:<p>
<center>
<img src="library-compilation.png" border=0>
</center>
<p>
</ul>
<p><hr><p>
<h2>IIV. Managing the list of modules</h2>
<ul>
The makefile <tt>freetype.mk</tt> only determines how to compile
each one of the modules that are located in the sub-directories of
<tt>freetype2/src</tt>.
<p>
However, when the function <tt>FT_Init_FreeType</tt> is invoked at
the start of an application, it must create a new <tt>FT_Library</tt>
object, and registers all <em>known</em> font drivers to it by
repeatly calling <tt>FT_Add_Driver</tt>.
<p>
The list of <em>known</em> drivers is located in the file
"<tt>freetype2/config/<em>system</em>/ftmodule.h</tt>", and is used
exclusively by the internal function <tt>FT_Default_Drivers</tt>. The
list in <tt>ftmodule.h</tt> must be re-generated each time you add
or remove a module from <tt>freetype2/src</tt>.
<p>
This is normally performed by invoking the top-level <tt>Makefile</tt>
with the <tt>modules</tt> target, as in:<p>
<ul>
<tt>make modules</tt>
</ul>
<p>
This will trigger a special rule that will re-generate
<tt>ftmodule.h</tt>. To do so, the Makefile will parse all module
directories for a file called "<tt>module.mk</tt>". Each
<tt>module.mk</tt> is a tiny sub-Makefile used to add a single
module to the driver list.
<p>
This is illustrated by the following graphics:<p>
<center>
<img src="drivers-list.png" border=0>
</center>
<p>
Note that the new list of modules is displayed in a very human-friendly
way after a "<tt>make modules</tt>". Here's an example with the current
source tree (on 11 Jan 2000):<p>
<ul><pre>
Regenerating the font drivers list in ./config/unix/ftmodule.h
* driver: sfnt ( pseudo-driver for TrueType & OpenType formats )
* driver: truetype ( Windows/Mac font files with extension *.ttf or *.ttc )
* driver: type1 ( Postscript font files with extension *.pfa or *.pfb )
-- done --
</pre></ul>
</ul>
<p><hr><p>
<h2>V. Building the demonstration programs</h2>
<ul>
Several demonstration programs are located in the
"<tt>freetype2/demos</tt>" directory hierarchy. This directory also
includes a tiny graphics sub-system that is able to blit glyphs to
a great variety of surfaces, as well as display these in various
graphics libraries or windowed environments.
<p>
This section describes how the demonstration programs are compiled,
using the configuration <tt>freetype2/config.mk</tt> and their own
<tt>freetype2/demos/Makefile</tt>.
<p>
To compile the demonstration programs, <em>after the library</em>,
simply go to <tt>freetype2/demos</tt> then invoke GNU make with no
arguments.
<p>
The top-level Makefile will detect the <tt>config.mk</tt> in the
<em>upper</em> directory and include it. Because it doesn't define
the <tt>BUILD_FREETYPE</tt> variable, this will not force the
inclusion of <tt>freetype2/config/freetype.mk</tt> as described in
the previous section.
<p>
the <tt>Makefile</tt> will then include the makefile called
"<tt>freetype2/demos/graph/rules.mk</tt>". The graphics <tt>rules.mk</tt>
defines the rules required to compile the graphics sub-system.
<p>
Because the graphics syb-system is also designed modularly, it is able
to use any number of "modules" to display surfaces on the screen.
The graphics modules are located in the subdirectories of
<tt>freetype2/demos/config</tt>. Each such directory contains a file
named <tt>rules.mk</tt> which is in charge of:<p>
<ul>
<li>detecting wether the corresponding graphics library is
available at the time of compilation.
<p>
<li>if it is, alter the compilation rules to include the graphics
module in the build of the <tt>graph</tt> library.
</ul>
<p>
When the <tt>graph</tt> library is built in <tt>demos/obj</tt>, the
demonstration programs executables are generated by the top-level
Makefile.
<p>
This is illustrated by the following graphics:<p>
<center>
<img src="demo-programs.png" border="0">
</center>
</ul>
<p><hr>
</td></tr></table></center>
</body>
</html>

BIN
docs/design/demo-programs.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 2.8 KiB

BIN
docs/design/drivers-list.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.8 KiB

791
docs/design/index.html
View File

@ -1,791 +0,0 @@
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="Author" content="David Turner">
<meta name="GENERATOR" content="Mozilla/4.5 [fr] (Win98; I) [Netscape]">
<title>FreeType 2 Internals</title>
</head>
<body>
<body text="#000000"
bgcolor="#FFFFFF"
link="#0000EF"
vlink="#51188E"
alink="#FF0000">
<center>
<h1>
FreeType 2.0 Internals</h1></center>
<center>
<h2>
Version 1.2</h2></center>
<center>
<h3>
&copy; 1999-2000 David Turner (<a href="fichier :///david@freetype.org">david@freetype.org</a>)<br>
&copy; 1999-2000 The FreeType Development Team (<a href="fichier :///devel@freetype.org">devel@freetype.org</a>)</h3></center>
<p><br>
<hr WIDTH="100%">
<br>&nbsp;
<h2>Introduction:</h2>
<p>This document describes in great deatils the internals of FreeType 2.
It is a must read for porters and developers alike. Its purpose is to
present the
<blockquote>This document describes in great details the internals of the
FreeType 2.0 library. It is a must read for porters and developers alike.
Its purpose is to present the engine's objects, their roles and interactions.
It is assumed that the <b><i>FreeType Glyph Conventions</i></b> document
has been read.
<p>We advise porters to also read the <b><i>FreeType Porting Guide</i></b>
after this document. Would-be hackers and maintainers are of course encouraged
to read the <b><i>FreeType Coding Conventions</i></b> document too. The
development of a new driver is described in more details in the <b><i>FreeType
Driver HowTo</i></b> document.</blockquote>
<p><br>
<hr WIDTH="100%">
<h2>
I. Overview :</h2>
<blockquote>
<h3>
1. Features (and what's new) :</h3>
<blockquote>FreeType 2.0 has a number of important new features that were
not found in the 1.x releases :
<br>&nbsp;
<blockquote><b>font-format independent API</b>
<br>FreeType 2.0 is able to support any kind of font format, be it fixed
or scalable, through the use of pluggable "font drivers". These drivers
can be added or replaced at run time, while applications use a new font
format-independent API.
<p><b>advanced stream caching</b>
<br>2.0 is able to control the number of concurrently opened streams when
using fonts. It is thus possible to open dozens or hundreds of font faces
without running out of system resources.
<p><b>real reentrancy support</b>
<br>It is now possible to use FreeType as a shared library with no static
data in a multi-threaded environment. The synchronization model has also
been simplified in order to make font driver writing easier. Of course,
you can build FreeType with no thread support to get a smaller library.
<p><b>support for cubic beziers and 17-levels anti-aliasing</b>
<br>The FreeType scan-line converter (a.k.a. raster) now supports cubic
bezier arcs seamlessly. It also provides a new anti-aliasing mode which
uses a palette of 17 levels of grays.
<br>&nbsp;</blockquote>
It also features the following :
<blockquote><b>performance improvements :</b>
<br>The FreeType raster has been optimized, and the generation of anti-aliased
pixmaps is now 60% faster than in the 1.x release. Moreover, the TrueType
bytecode interpreter has been profiled and greatly optimised.
<p><b>easier portability</b>
<br>Porting and configuring FreeType is now much easier. A single file
must be provided for system-specific operations (like memory, i/o, thread
management), and a single configuration header is used to select the build
you need.
<br>&nbsp;</blockquote>
</blockquote>
<h3>
2. Architecture :</h3>
<blockquote>The engine is now split in several parts, which are :
<h4>
a. The base layer :</h4>
<blockquote>This part contains all the font-format independent features
of the engine which are :
<ul>
<li>
computations/scaling</li>
<li>
list processing</li>
<li>
outline processing</li>
<li>
scan-line converter</li>
<li>
stream manager</li>
<li>
base object classes</li>
<li>
debugging &amp; traces</li>
<li>
high-level API functions</li>
<li>
low-level system object (memory, i/o, threads)</li>
</ul>
</blockquote>
<h4>
b. The font drivers :</h4>
<blockquote>Each font format is managed with the use of a single font driver
object. The base layer is able to manage several drivers, and these can
be easily added, removed or upgraded at runtime. Each driver has the following
features and functions :
<ul>
<li>
auto-check font format when opening a font resource (i.e. file)</li>
<li>
access, load and/or extract all tables and data from the font file</li>
<li>
grid-fit/hint the glyph outlines (in the case of scalable formats like
TrueType or Type1)</li>
<li>
provide extensions to access font format-specific data and tables from
the font file</li>
</ul>
Note that FreeType 2.0 is a font service. Its purpose is to provide a unified
API for all kinds of fonts and extract individual glyph images and metrics.
However, it does not render text itself, as this operation is left to the
developer, or to higher-level libraries built on top of FreeType. Here
are a few features that are thus not implemented :
<blockquote>1) Text string rendering
<br>2) Glyph bitmap/outline caching for improved performance
<br>3) Synthetic fonts (i.e. italicising, emboldening, underlining)
<br>4) Contextual glyph substitution and other advanced layout processes</blockquote>
Note that features 1 through 3 should be provided by the SemTex library,
which may soon become part of the standard FreeType distribution.</blockquote>
</blockquote>
</blockquote>
<p><br>
<hr WIDTH="100%">
<h2>
II. Design :</h2>
<blockquote>
<h3>
1. Objects :</h3>
<blockquote>They are several kinds of objects in FreeType, which can be
described as follows :
<blockquote><b>Base objects</b>
<br>These objects do not relate directly to font data, but to the way it
is organised and managed. It is the basic core and provides functions that
are heavily used by each font driver. Examples are the resource objects,
used to describe font files, the system object used to manage low-level
system operations, or the raster object, used to convert vector outlines
into bitmaps or anti-aliased pixmaps. Most of the base objects are not
directly visible for client applications of FreeType.
<p><b>Font objects</b>
<br>The font objects directly model the data as it is found in font files.
The root classes implemented in the base layer like <tt>FT_Face</tt>, <tt>FT_Size</tt>,
<tt>FT_GlyphSlot</tt>,
must be derived in each font driver.</blockquote>
Objects are defined in the files "<tt>base/freetype.h</tt>" and "<tt>base/ftobjs.h</tt>".
The former contains all the public object definitions usable by client
applications. The latter contains private definitions used by the rest
of the base layer and each font driver.</blockquote>
<h3>
2. List management</h3>
<blockquote>The "<tt>base/ftlist.c</tt>" component a very simple doubly-linked
list facility which is used by the rest of the engine to create and process
lists, including iteration and finalisation. The definition of the list
node and functions are placed in the "<tt>base/freetype.h</tt>" to let
client applications access listed objects as they like.
<p>The base list type is <tt>FT_List</tt>, which links nodes of type <tt>FT_ListNode</tt>
together.
<br>&nbsp;</blockquote>
<h3>
3. Limited encapsulation</h3>
<blockquote>Unlike what happened in the 1.x releases, the <tt>FT_Face</tt>,
<tt>FT_Size</tt>,
<tt>FT_GlyphSlot</tt> and <tt>FT_CharMap</tt> types are no longer blind
pointers to opaque types. Rather, the corresponding structures are now
public (and defined in "<tt>base/freetype.h</tt>", see <tt>FT_FaceRec</tt>,
<tt>FT_SizeRec</tt>,
etc..) in order to let client applications read directly the various object
attributes they're interested in.
<p>This breaks encapsulation of implementation, famed by OOP, but was chosen
because:
<br>&nbsp;
<ul>
<li>
it simplifies a lot the work of client applications and libraries which
don't need to perform a function call everytime they want to read one important
object attribute (nor does it force them to cache these attributes in their
own structures).</li>
</ul>
<ul>
<li>
It reduces greatly the API, as many <tt>FT_Get_XXX</tt> functions are avoided.</li>
</ul>
<ul>
<li>
Higher-level libraries are able to&nbsp; access data directly. When it
is used frequently, they don't need to cache it in their own structures.</li>
</ul>
<ul>
<li>
It is possible to tightly link FreeType objects with higher-level ones,
in a clearer and more efficient way. This is very important when one wants
to write a C++ wrapper or a text rendering library on top of FreeType (actually,
both projects were performed in an earlier version of FreeType 2.0 which
featured classic encapsulation through get/set methods. The resulting code
was ugly and slow. Moving to a limited encapsulation approach simplified
so many things that the compiled code size was reduced by a factor of two
!).</li>
</ul>
<ul>
<li>
Finally, the API and font object structures were designed after the creation
of two scalable font drivers and one bitmap font driver. They are now very
stable and the public (visible) attributes are not going to change.</li>
</ul>
</blockquote>
</blockquote>
<p><br>
<hr WIDTH="100%">
<h2>
III. Base objects :</h2>
<blockquote>This section describes the FreeType base object classes :
<br>&nbsp;
<h3>
1. System objects :</h3>
<blockquote>The system class is in charge of managing all low-level and
system-specific operations. This means simply memory management, i/o access
and thread synchronisation. It is implemented by the "<tt>ftsys.c</tt>"
component, whose source must be located in the configuration directory
when building FreeType. (e.g. "<tt>lib/arch/ansi/ftsys.c</tt>" for an ANSI
build, "<tt>lib/arch/unix/ftsys.c</tt>" for a Unix one, etc..).
<p>Porting FreeType 2.0 really means providing a new implementation of
<tt>ftsys</tt>
(along with a few configuration file changes). Note however that its interface
is common to all ports, and located in "<tt>base/ftsys.h</tt>".</blockquote>
<h3>
2. Resources and Streams:</h3>
<blockquote>The concepts of files as storages, and files as streams has
been separated for FreeType 2.0. The "<b><i>resource</i></b>" concept was
introduced while the "<b><i>stream</i></b>" one has been redefined. Here
is how they work together :
<ul>
<li>
a "<b>resource</b>" is an object which models a file, seen as a storage.
There are several classes of resources, which differ usually in two ways
: the way their data is accessed by applications, and the way they're named
within the system.</li>
</ul>
<ul>For example, when parsing files with the ANSI C library, data has to
be read (through fseek/fread) into intermediate buffers before it can be
decoded. This scheme is highly portable, but rather inefficient; when using
it, we'll describe the file as a disk-based resource.
<p>As most modern operating systems now provide memory-mapped files, which
allow direct access while improving performance and reducing memory usage.
Because data can be read directly in memory, we'll speak of a memory-based
resource in this case. For embedded systems (like printers, PDAs, etc..),
ROM-fonts fit into this category as well.
<p>Regarding naming, most systems use a string to name files in their storage
hierarchy. Though a typical pathname is an ASCII string (<tt>'c:\windows\fonts\times.ttf'</tt>
on Windows, <tt>'/home/fonts/times.ttf'</tt> on Unix), some OSes use different
schemes, varying from Unicode character strings to file i-node numbers.
These details are platform-specific and must be hidden to the rest of the
library in resource objects.
<p>A resource encapsulates the lowest details regarding a file, though
it should have NO STATE. Note that the nature or type of a resource (i.e.
disk or memory based) is important to the "stream" component only. The
rest of the library and font drivers work transparently from their implementation.
<p>Note also that it is perfectly possible to mix resources of distinct
natures in a single build</ul>
<ul>
<li>
a "<b>stream</b>" is an object which is used to extract bytes from a resource.
Only resource objects can create streams, through its <i><tt>Open_Stream()</tt></i>
method. A stream has state, which typically consist of a file "cursor",
some intermediate buffers, a "current frame" and, of course, methods used
to extract the data from streams, resolving endianess and alignement issues.</li>
</ul>
Data can be extracted from streams through direct reads, or through the
use of <b>frames</b>. A frame models <i>a run of contiguous bytes</i> starting
from the current stream position, and of liberal size.
<p>Methods exist to extract successive integers of any sizes, while resolving
endianess and alignement issues. Rather than a long rethorical explanation,
here's how frames are typically used :
<blockquote><tt>{</tt>
<br><tt>&nbsp; …</tt>
<br><tt>&nbsp; FT_Error&nbsp; error;</tt>
<p><tt>&nbsp; error = FT_Access_Frame( stream, 14 );</tt>
<br><tt>&nbsp; if (error) goto Fail;</tt>
<p><tt>&nbsp; val1 = FT_Get_Short(stream);</tt>
<br><tt>&nbsp; val2 = FT_Get_Long(stream);</tt>
<br><tt>&nbsp; val3 = FT_Get_Long(stream);</tt>
<br><tt>&nbsp; val4 = FT_Get_Long(stream);</tt>
<p><tt>&nbsp; FT_Forget_Frame(stream);</tt>
<br><tt>&nbsp; …</tt>
<br><tt>}</tt></blockquote>
This code does the following :
<blockquote>
<ol>
<li>
&nbsp;first, it "loads" the next 14 bytes from the current cursor position
into the stream's frame, using the <tt>FT_Access_Frame</tt> API. An error
is returned if, for example, less than 14 bytes are left in the stream
when the call occurs..</li>
</ol>
<ol>
<li>
&nbsp;it extract four integers (one 16-bit short, three 32-bit longs) from
the frame using <tt>FT_Get_Short</tt> and <tt>FT_Get_Long</tt>. These function
increment the frame's cursor finally, it "releases" the stream's frame.</li>
</ol>
<ol>
<li>
&nbsp;Each stream has its own frame which can be accessed independently,
however, nested frame accesses are not allowed. Note also that the bytes
are effectively read from the stream on the call to <tt>FT_Access_Frame</tt>.
Any subsequent read will occur after these 14 bytes, even if less are extracted
through <tt>FT_Get_xxxx</tt> functions.</li>
</ol>
</blockquote>
The implementation of the resource class is located in the system component
(i.e. "<tt>arch/<i>&lt;system></i>/ftsys.c</tt>") and can thus be tailored
for a specific port of the engine.
<p>A resource can be created through the <tt>FT_New_Resource</tt> API;
however this function only accepts an 8-bit pathname to name the target
font file, which may be inappropriate for systems using a different naming
scheme (e.g. UTF-16 pathname, i-node number, etc..). It's up to the porter
then to provide its own resource creation function (like. <tt>FT_New_UTF16_Resource</tt>,
for example) in its version of "<tt>ftsys.c</tt>".
<p>Note that <tt>FT_New_Resource</tt> will fail and return an error code
if the font file cannot be found, or when its font format isn't recognized
by one of the drivers installed in the library. The list or resources created
for a given library instance is thus the list of "installed font files".
<br>&nbsp;</blockquote>
<h3>
3. Stream Manager :</h3>
<blockquote>As said before, resources do not bear states, while streams
do. Stream creation is also a very lengthy process, depending on the target
operating system (e.g. "<tt>fopen</tt>" is usually very slow).
<p>Because a typical font driver will want to use a new stream on each
access to individual glyphs, being able to cache the most recently used
streams is a requirement in order to avoid considerable performance penalties.
<p>Stream caching is thus implemented in the "<tt>ftstream</tt>" component.
It maintains a simple LRU list of the least recently used streams. Each
stream in the cache is still opened and available for immediate processing.
When a resource is destroyed, the stream cache is parsed to remove all
related cached streams.
<p>Stream caching can also be disabled with a configuration macro when
using only ROM based resources (where stream opening is really quick).
It is implemented through a Stream Manager object (see <tt>ftstream.c</tt>).
<br>&nbsp;</blockquote>
<h3>
4. Raster :</h3>
<blockquote>The raster is the component is charge of generating bitmaps
and anti-aliased pixmaps from vectorial outline definitions. It is also
sometimes called the scan-line converter. It has been completely rewritten
for FreeType 2.0 in order to support third-order bezier arcs, 17-levels
anti-aliasing (through 4x4 sub-sampling), improved performance, as well
as stand-alone compilation (in order to include it in other graphics package
without requiring the rest of the FreeType engine).
<p>Because it was designed for easy re-use and embedded systems, the raster
is a rtaher 'unusual' piece of code, because it doesn't perform a single
memory allocation, nor contain any static or global variable. Rather, it
is up to client applications to allocate a raster object in their own heap
or memory space.
<p>Each raster object also needs a rather large block of memory called
its render pool. The pool is used during rendering (and only during it)
in order to perform the scan-line conversion. Because it accesses and manages
data directly within the pool, the raster yelds impressive performance
as well as bounded memory consumption. It can also automatically decompose
large requests into smaller individual sub-tasks.
<p>Finally, it never creates bitmaps or pixmaps, but simply renders into
them (providing clipping too). These must be described to the raster with
the help of a <tt>FT_Raster_Map</tt> structure (a very simple bitmap/pixmap
descriptor).
<p>Note that when rendering anti-aliased pixmaps, the raster doesn't use
an intermediate bitmap buffer, as filtering is part of the scan-line conversion
process.
<br>&nbsp;</blockquote>
<h3>
5. Library objects :</h3>
<blockquote>A library object models a single instance of the FreeType engine.
This is useful when FreeType is compiled as a shared object (DLL), as it
can then be used by several applications, each with its own resources and
objects.
<p>The <tt>FT_Library</tt> type is an opaque handle to a library object.
Such an object is created through a call&nbsp; to <tt>FT_Init_FreeType</tt>.
Once you don't need it anymore, one can destroy a library object through
<tt>FT_Done_FreeType</tt>.
<p>Note that in reentrant builds, several threads can access a single library
object concurrently. Such a build can be chosen by switching one configuration
macro in the file '<tt>arch/<i>&lt;system></i>/ftconfig.h</tt>'</blockquote>
<h3>
6. Driver objects :</h3>
<blockquote>A driver object models an instance of a given font driver,
i.e. an element of FreeType code in charge of handling a given font format,
like TrueType, Type1, FNT, PCF, etc..
<p>Each library object contains a given set of driver objects when it is
created through FT_Init_FreeType, this set being determined at compile
time (see the file 'base/ftapi.c'). However, removing or adding drivers
is possible at run-time, in order to make upgrades easy.</blockquote>
<h3>
7. Diagram</h3>
<blockquote>This diagram show the object relationships for the sole base
layer. The library object is the root of the object graph :
<center>
<p><img SRC="objects_diagram.png" height=300 width=562></center>
<p>It can be read as follows :
<br>&nbsp;
<ul>
<li>
Each library object has one system, one raster and one stream manager objects.
These objects can only belong to one given library.</li>
</ul>
<ul>
<li>
Each library contains one list of 0 or more resources, as well as one list
of 0 or more driver objects.</li>
</ul>
<ul>
<li>
Each stream manager holds a bounded list ("0..n" where 'n' is the stream
cache's size) of stream objects. Each stream is related to one given resource
object. Each resource may be related to zero or one stream.</li>
</ul>
<ul>
<li>
Each resource is related to one driver object. A driver is related to 0
or more resources.</li>
</ul>
</blockquote>
</blockquote>
<p><br>
<hr WIDTH="100%">
<h2>
IV. Font objects :</h2>
<blockquote>Font objects are used to directly map the information found
in font files into several categories :
<br>&nbsp;
<h3>
1. Face objects :</h3>
<blockquote>Face objects are used to model individual font faces. They
encapsulate data which isn't related to a specific character size, or a
specific glyph or glyph set. Usually, this means :
<ul>
<li>
the font face's family and style names (e.g. "Palatino" + "Regular")</li>
<li>
some flags indicating which kind of font this is (scalable or fixed ? fixed-width
or proportional ? horizontal or vertical ? etc…)</li>
<li>
the number of glyphs, charmaps and eventually fixed character sizes (for
bitmap formats) found in the font face.</li>
<li>
for scalable formats, some important metrics like the ascender, descender,
global font bounding box, maximum advance width, etc.. expressed in notional
font/grid units (as well as the number of units on the EM grid).</li>
</ul>
A face is created from a resource object, with the <tt>FT_New_Face</tt>
API. Each driver contains a list of opened face objects for the resources
it manages. When a driver is removed or destroyed, all its child faces
are discarded automatically with it.</blockquote>
<h3>
2. Size objects :</h3>
<blockquote>Size objects are used to model a given character dimension
for a given device resolution (which really means a given character pixel
dimensions).
<p>Each size object is created from a parent face object. The object can
be reset to new dimensions at any time. Each face object holds a list of
all its child sizes, these are destroyed automatically when the face object
is discarded.
<p>The metrics contains metrics, expressed in pixels, for the ascender,
descender, maximum advance width, etc..
<br>&nbsp;</blockquote>
<h3>
3. Glyph Slot objects :</h3>
<blockquote>A glyph slot is a container where one can load individual glyphs,
be they in vector of bitmap format. Each slot also contains metrics for
the glyph it contains.
<p>Each face object contains one or more glyph slot object : the first
glyph slot is created automatically with its parent face, and it is possible
to add new glyph slots (this is rarely used outside of debugging purposes).
<br>&nbsp;</blockquote>
<h3>
4. CharMap objects :</h3>
<blockquote>A charmap object is a sort of dictionary whose task is to translate
character codes in a given character encoding (like ShiftJIS, Unicode,
ANSI, etc..) into glyph indexes in a given font face.
<p>A face object contains one or more charmap objects. All charmap objects
are created when the parent face is created, though they're not directly
visible to client applications (rather, they can be enumerated through
FT_Get_First_CharMap and FT_Get_Next_CharMap, or more simply picked adequately
with FT_Find_CharMap for a set of given encodings).
<br>&nbsp;</blockquote>
<h3>
5. Diagram</h3>
<blockquote>The following diagram illustrates the relationships between
font objects :
<center>
<p><img SRC="objects_diagram2.png" height=327 width=561></center>
<p>Which can be read as :
<br>&nbsp;
<ul>
<li>
each resource may have zero or more child face objects "opened" for it.
The number of faces is bounded by the number of font faces within the font
resource.</li>
</ul>
<ul>
<li>
each driver holds a list of all the faces opened for the resources it manages.
When the driver is removed, its child faces are discarded automatically.</li>
</ul>
<ul>
<li>
each face object has one single parent resource, and one single driver.</li>
</ul>
<ul>
<li>
each face has one or more charmaps, and one or more glyph slots</li>
</ul>
<ul>
<li>
each face holds a list of zero or more child size objects</li>
</ul>
<ul>
<li>
each charmap, glyph slot and size is related to one given parent face.
These objects are destroyed automatically when the parent face is discarded.</li>
</ul>
</blockquote>
</blockquote>
<p><br>
<hr WIDTH="100%">
<h2>
V. Driver Interface :</h2>
<blockquote>A font driver is added to a given library object through the
<tt>FT_Add_Driver</tt>
API. This function receives a structure known as a <tt>FT_DriverInterface</tt>,
which describes the driver's basic properties.
<p>The <tt>FT_DriverInterface</tt> contains a set of function pointers
used for the base FreeType functionalities. However, each driver can also
provide a font-format-specific extended interface to allow client applications
to use more advanced features.
<br>&nbsp;
<h3>
1. Common Interface</h3>
<blockquote>The structure of <tt>FT_DriverInterface</tt> is rather simple,
and defined in "<tt>base/ftdriver.h</tt>". It must be well known by any
developer who wants to write a new driver for the engine. We advise reading
the <b><i>FreeType Driver HowTo</i></b> as well as the source code of existing
drivers. Source comments.</blockquote>
<h3>
2. Driver-specific extensions</h3>
<blockquote>The field of the <tt>FT_DriverInterface</tt> structure is a
typeless pointer to a format-specific interface. This extended interface
is usually a structure containing function pointers as well as other kind
of information related to the driver.
<p>It is assumed that client applications that wish to use the driver-specific
extensions are able to <tt>#include</tt> the relevant header files to understand
the format-specific interface structure.</blockquote>
</blockquote>
<hr WIDTH="100%">
<h2>
VI. Configuration:</h2>
<blockquote>This section relates to the configuration of the FreeType library.
By configuration, we mean selection of build options as well as the choice
of font drivers to be used for each new library object.
<br>&nbsp;
<h3>
1. Configuration files :</h3>
<blockquote>A single file is used to configure the FreeType base engine.
As it is considered system-specific, it is located in the architecture
directories of the library, under the name "arch/&lt;system>/ftconfig.h".
Note that the same directory should also contain a platform-specific implementation
of "ftsys.c".
<p>The configuration files is a simple C header which is included by the
engine's sources during compilation. It is not included in "freetype.h",
and hence doesn't need to be copied when installing the FreeType headers
on your system.
<p>It is made of a series of #define or #undef statements, which are used
to select or turn off a specific option. Each option is documented with
heavy comments, and some of them are explained below.</blockquote>
<h3>
2. Building and Makefiles :</h3>
<blockquote>FreeType 2.0 is more complex than its 1.x release. In order
to facilitate maintenance, as well as ease considerably the writing of
new font drivers, <b><i>only GNU Make is supported with FreeType 2.0</i></b>.
However, it is possible to use any compiler, as well as any object or library
prefix (<tt>.o, .obj, .a, .lib</tt> etc..) with them.
<p>To build FreeType 2.0, one has to be in the library directory, then
invoke its platform-specific makefile. For a Unix system, this would be
:
<blockquote>
<blockquote><tt>% cd freetype2/lib</tt>
<br><tt>% make -f arch/unix/Makefile</tt>
<p>where '<tt>make</tt>' is really GNU Make !</blockquote>
</blockquote>
The system-specific <tt>Makefile</tt> located in '<tt>arch/<i>&lt;system></i></tt>'
is a tiny file used to define several variables. It then includes the file
<tt>freetype2/lib/Makefile.lib</tt>,
which contains all the gory details about library compilation. The system-specific
<tt>Makefile</tt> can be very easily modified to accomodate a new compiler/platform
(see the comments within one of these files).
<p>Each font driver is located in a directory like "<tt>freetype2/lib/drivers/<i>&lt;formatdir></i></tt>".
For example, the TrueType driver is located in "<tt>drivers/truetype</tt>".
Each driver directory must contain a <tt>Makefile</tt> which will be included
by <tt>Makefile.lib</tt>. The former is used to define and build driver
object files.
<br>&nbsp;
<p><br>
<center>
<p><img SRC="build_diagram.png" height=284 width=559></center>
</blockquote>
<h3>
3. Make options :</h3>
<blockquote>The base layer, as well as each font driver, are made up of
several C sources. Traditionally, one compiles each source (i.e. '<tt>.c</tt>'
file) into an object ('<tt>.o</tt>' or '<tt>.obj</tt>') file, and all of
them are grouped into a library file (i.e. '<tt>.a</tt>' or '<tt>.lib</tt>').
<p>By default, FreeType takes a slightly different approach when it comes
to compiling each part of the engine. Usually, a single tiny source is
compiled, which includes all other component sources. This results in a
single object files, with the benefits or reduced code size, usually better
compilation as well as a drastic reduction of the number of symbols exported
by the library. Of course, it is made possible through the use of specific
declaration macros in the FreeType source (see the definition of <tt>LOCAL_DEF</tt>
and <tt>LOCAL_FUNC</tt> in <tt>ftconfig.h</tt> for details).
<p>For a concrete example, see the source code in "<tt>base/ftbase.c</tt>"
which generates the whole base layer in a single object file. The same
build process is applied to font drivers, in order to generate one single
object file per given font format (e.g. <tt>truetype.o</tt>, <tt>type1.o</tt>,
etc..).
<p>Compiling the library and drivers in "normal" mode is possible, through
the use of the '<tt>multi</tt>' target (which really means &laquo; multiple
objects &raquo;). For example, calling :
<blockquote><tt>% make -f arch/ansi/Makefile multi</tt></blockquote>
Will build the FreeType library by compiling each source file to an individual
object, then linking them together. You'll notice that the library is significantly
bigger in this case. Creating a shared dll from a 'multi' build is certainly
a very poor idea, as this will export a huge quantity of symbols that aren't
useful to any client application.</blockquote>
<h3>
4. Adding a driver at compile time</h3>
<blockquote>A driver can be included very easily in the build process by
including its <tt>Makefile</tt> in <tt>Makefile.lib</tt>. For example,
the TrueType driver is simply included with the following lines (see <tt>Makefile.lib</tt>):
<blockquote><tt># TrueType driver rules</tt>
<br><tt>#</tt>
<br><tt>include $(DRIVERS_DIR)/truetype/Makefile</tt></blockquote>
<p><br>Where <tt>DRIVERS_DIR</tt> really is "<tt>freetype2/lib/drivers</tt>",
though this can be redefined. You can, of course specify a different path
if you want to place your driver sources in another location.
<p>Note that this only adds the driver's object files to the generated
library file. A few more steps are needed to make your <tt>FT_Library</tt>
objects use the driver. They consist in modifying the file "<tt>base/ftinit.c</tt>",
whose sole purpose is to define the set of driver objects that are to be
created with each new library object.
<br>&nbsp;</blockquote>
<h3>
5. Adding a driver at run time</h3>
<blockquote>New driver objects can be added at run-time through the <tt>FT_Add_Driver</tt>
API. This function takes a handle to an existing library object, as well
as a pointer to a given driver interface. This interface is used to create
a new driver object and register it within the library.
<p>Similarly, a single driver can be removed from a library anytime through
<tt>FT_Remove_Driver</tt>.
This will automatically discard the resources and face objects managed
by the driver.</blockquote>
<h3>
6. Custom library objects :</h3>
<blockquote>Finally, it is possible to build custom library objects. You
need to pass a handle to a valid <tt>FT_System</tt> object to the <tt>FT_Build_Library</tt>
API. The function will return a handle to the new fresh library object.
Note that the library has no registered drivers after the call, developers
have to add them by hand with <tt>FT_Add_Driver</tt>.
<p>It is thus possible to create two distinct library objects with distinct
<tt>FT_System</tt>
implementations in the same session, which can be useful for debugging
purpose.</blockquote>
<br>&nbsp;</blockquote>
</body>
</html>

343
docs/design/io-frames.html
View File

@ -1,343 +0,0 @@
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type"
content="text/html; charset=iso-8859-1">
<meta name="Author"
content="David Turner">
<title>FreeType 2 Internals - I/O Frames</title>
</head>
<body text="#000000"
bgcolor="#FFFFFF"
link="#0000EF"
vlink="#51188E"
alink="#FF0000">
<h1 align=center>
FreeType 2.0 I/O Frames
</h1>
<h3 align=center>
&copy; 2000 David Turner
(<a href="mailto:david@freetype.org">david@freetype.org</a>)<br>
&copy; 2000 The FreeType Development Team
(<a href="http://www.freetype.org">www.freetype.org</a>)
</h3>
<center>
<table width="70%">
<tr><td>
<hr>
<h2>
Introduction
</h2>
<p>This document explains the concept of I/O <b>frames</b> as used in the
FreeType&nbsp;2 source code. It also enumerates the various functions and
macros that can be used to read them.</p>
<p>It is targeted to FreeType hackers, or more simply to developers who
would like a better understanding of the library's source code.</p>
<hr>
<h2>
I. What frames are
</h2>
<p>Simply speaking, a frame is an array of bytes in a font file that is
`preloaded' into memory in order to be rapidly parsed. Frames are useful
to ensure that every `load' is checked against end-of-file overruns, and
provides nice functions to extract data in a variety of distinct
formats.</p>
<p>But an example is certainly more meaningful than anything else. The
following code</p>
<font color="blue">
<pre>
error = read_short( stream, &str.value1 );
if ( error ) goto ...
error = read_ulong( stream, &str.value2 );
if ( error ) goto ...
error = read_ulong( stream, &str.value3 );
if ( error ) goto ...</pre>
</font>
<p>can easily be replaced with</p>
<font color="blue">
<pre>
error = FT_Access_Frame( stream, 2 + 4 + 4 );
if ( error ) goto ...
str.value1 = FT_Get_Short( stream );
str.value2 = FT_Get_ULong( stream );
str.value3 = FT_Get_ULong( stream );
FT_Forget_Frame( stream );</pre>
</font>
<p>Here, the call to <code>FT_Access_Frame()</code> will</p>
<ul>
<li>
<p>Ensure that there are at least 2+4+4=10 bytes left in the
stream.</p>
</li>
<li>
<p>`Preload' (for disk-based streams) 10&nbsp;bytes from the current
stream position.</p>
</li>
<li>
<p>Set the frame `cursor' to the first byte in the frame.</p>
</li>
</ul>
<p>Each <code>FT_Get_Short()</code> or <code>FT_Get_ULong()</code> call
will read a big-endian integer from the stream (2&nbsp;bytes for
<code>FT_Get_Short()</code>, 4&nbsp;bytes for <code>FT_Get_ULong</code>)
and advance the frame cursor accordingly.</p>
<p><code>FT_Forget_Frame()</code> `releases' the frame from memory.</p>
<p>There are several advantages to using frames:</p>
<ul>
<li>
<p>Single-check when loading tables.</p>
</li>
<li>
<p><em>Making code clearer</em> by providing simple parsing functions
<em>while keeping code safe</em> from file over-runs and invalid
offsets.</p>
</li>
</ul>
<hr>
<h2>
II. Accessing and reading a frame with macros
</h2>
<p>By convention in the FreeType source code, macros are able to use two
implicit variables named <var>error</var> and <var>stream</var>. This is
useful because these two variables are extremely often used in the
library, and doing this only reduces our typing requirements and make the
source code much clearer.</p>
<p>Note that <var>error</var> must be a local variable of type
<code>FT_Error</code>, while <var>stream</var> must be a local variable or
argument of type <code>FT_Stream</code>.</p>
<p>The macro used to access a frame is <font
color="purple"><code><b>ACCESS_Frame(_size_)</b></code></font>, it will
translate to</p>
<font color="blue">
<pre>
( error = FT_Access_Frame( stream, _size_ ) )
!= FT_Err_Ok</pre>
</font>
<p>Similarly, the macro <font
color="purple"><b><code>FORGET_Frame()</code></b></font> translates to</p>
<font color="blue">
<pre>
FT_Forget_Frame( stream )</pre>
</font>
<p>Extracting integers can be performed with the <code>GET_xxx()</code>
macros, like</p>
<table align=center
cellpadding=5>
<tr valign="top">
<th>
Macro name
</th>
<th>
Translation
</th>
<th>
Description
</th>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_Byte()</b></code></font>
</td>
<td>
<font color="blue"><code>FT_Get_Byte(stream)</code></font>
</td>
<td>
<p>Reads an 8-bit unsigned byte.</p>
</td>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_Char()</b></code></font>
</td>
<td>
<font color="blue"><code>(FT_Char)<br>
FT_Get_Byte(stream)</code></font>
</td>
<td>
<p>Reads an 8-bit <em>signed</em> byte.</p>
</td>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_Short()</b></code></font>
</td>
<td>
<font color="blue"><code>FT_Get_Short(stream)</code></font>
</td>
<td>
Reads a 16-bit signed big-endian integer.
</td>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_UShort()</b></code></font>
</td>
<td>
<font color="blue"><code>(FT_UShort)<br>
FT_Get_Short(stream)</code></font>
</td>
<td>
Reads a 16-bit unsigned big-endian integer.
</td>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_Offset()</b></code></font>
</td>
<td>
<font color="blue"><code>FT_Get_Offset(stream)</code></font>
</td>
<td>
Reads a 24-bit signed big-endian integer.
</td>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_UOffset()</b></code></font>
</td>
<td>
<font color="blue"><code>(FT_UOffset)<br>
FT_Get_Offset(stream)</code></font>
</td>
<td>
Reads a 24-bit unsigned big-endian integer.
</td>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_Long()</b></code></font>
</td>
<td>
<font color="blue"><code>FT_Get_Long(stream)</code></font>
</td>
<td>
Reads a 32-bit signed big-endian integer.
</td>
</tr>
<tr valign="top">
<td>
<font color="purple"><code><b>GET_ULong()</b></code></font>
</td>
<td>
<font color="blue"><code>(FT_ULong)<br>
FT_Get_Long(stream)</code></font>
</td>
<td>
Reads a 32-bit unsigned big-endian integer.
</td>
</tr>
</table>
<p>(Note that an <b>Offset</b> is an integer stored with 3&nbsp;bytes on
the file.)</p>
<p>All this means that the following code</p>
<font color="blue">
<pre>
error = FT_Access_Frame( stream, 2 + 4 + 4 );
if ( error ) goto ...
str.value1 = FT_Get_Short( stream );
str.value2 = FT_Get_ULong( stream );
str.value3 = FT_Get_ULong( stream );
FT_Forget_Frame( stream );</pre>
</font>
<p>can be simplified with macros:</p>
<font color="blue">
<pre>
if ( ACCESS_Frame( 2 +4 + 4 ) ) goto ...
str.value1 = GET_Short();
str.value2 = GET_ULong();
str.value3 = GET_ULong();
FORGET_Frame();</pre>
</font>
<p>Which is clearer. Notice that <var>error</var> and <var>stream</var>
must be defined locally though for this code to work!</p>
<hr>
<h2>
III. Alternatives
</h2>
<p>It is sometimes useful to read small integers from a font file without
using a frame. Some functions have been introduced in FreeType&nbsp;2 to
do just that, and they are of the form <font
color="blue"><code>FT_Read_xxxx</code></font>.</p>
<p>For example, <font color="blue"><code>FT_Read_Short(stream,
&error)</code></font> reads and returns a 2-byte big-endian integer from a
<var>stream</var>, and places an error code in the <var>error</var>
variable.</p>
<p>Thus, reading a single big-endian integer is shorter than using a frame
for it.</p>
<p>Note that there are also macros <font
color="purple"><code>READ_xxx()</code></font> which translate to</p>
<font color="blue">
<pre>
FT_Read_xxx( stream, &error ), error != FT_Err_Ok</pre>
</font>
<p>and can be used as in</p>
<font color="blue">
<pre>
if ( READ_UShort( variable1 ) ||
READ_ULong ( variable2 ) )
goto Fail;</pre>
</font>
<p>if <var>error</var> and <var>stream</var> are already defined
locally.</p>
</td></tr>
</table>
</center>
</body>
</html>

BIN
docs/design/library-compilation.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.7 KiB

BIN
docs/design/logo1.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 870 B

BIN
docs/design/objects_diagram.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 3.8 KiB

BIN
docs/design/objects_diagram2.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 3.1 KiB

BIN
docs/design/platform-detection.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.7 KiB

258
docs/design/system-interface.html
View File

@ -1,258 +0,0 @@
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="Author" content="David Turner">
<meta name="GENERATOR" content="Mozilla/4.5 [fr] (Win98; I) [Netscape]">
<title>FreeType 2 Internals</title>
</head>
<body>
<body text="#000000"
bgcolor="#FFFFFF"
link="#0000EF"
vlink="#51188E"
alink="#FF0000">
<center>
<h1>
FreeType 2.0 System Interface</h1></center>
<center>
<h3>
&copy; 2000 David Turner (<a href="fichier :///david@freetype.org">david@freetype.org</a>)<br>
&copy; 2000 The FreeType Development Team (<a href="fichier :///devel@freetype.org">devel@freetype.org</a>)</h3></center>
<p><br>
<hr WIDTH="100%">
<br>&nbsp;
<h2>Introduction:</h2>
<ul>
This document explains how the FreeType 2 library performs the low-level and
system-specific operations of memory management and i/o access. It is targetted
to FreeType hackers, porters and "advanced" developers who want special
features like providing their own memory manager or streams.
<p>
Note that the only system-specific part of the library is a file
named "<tt>ftsystem.c</tt>", normally located in the directory
"<tt>freetype2/config/&lt;system&gt;</tt>" where <tt>&lt;system&gt;</tt> designates
your platform (e.g. "<tt>config/ansi/ftsystem.c</tt>" or
"<tt>config/unix/ftsystem.c</tt>").
<p>
</ul>
<p>
<hr>
<p>
<h2>I. Memory Management</h2>
<ul>
Memory allocation and releases are performed through a <tt>FT_Memory</tt> object in
FreeType. A <tt>FT_Memory</tt> is nothing more than a table of functions plus
an arbitrary user data field. It is defined in the file
"<tt>freetype2/include/ftsystem.h</tt>" and has the following structure:
<p>
<ul>
<tt>typedef struct</tt><br>
<tt>{</tt>
<ul>
<table>
<tr><td><tt><b>void* user</b></tt> <td> // a user-defined pointer. This is zero by default
<tr><td><tt><b>void* (*alloc)( FT_System, int)</b></tt> <td> // a function used to allocate a new block
<tr><td><tt><b>void* (*realloc)( FT_System, int, int, void* )</b></tt><td> // a function used to reallocate a given block
<tr><td><tt><b>void (*free)( FT_System, void*)</b></tt> <td> // a function used to release a given block
</table>
</ul>
<tt>} FT_MemoryRec, *FT_Memory;</tt><br>
</ul>
<p>
You'll notice that:<p>
<ul>
<li>The <tt>FT_Memory</tt> type is really a pointer to a <tt>FT_MemoryRec</tt>.
This is a normal convention for the FreeType code.
<li>The <tt>realloc</tt> takes two integer arguments. The first one is the
current block size, the second one its new size.
</ul>
<p>
All current implementations of "<tt>ftsystem.c</tt>" provide a very simple
implementation of the <tt>FT_Memory</tt> interface by calling directly the
standard C <tt>alloc</tt>, <tt>realloc</tt> and <tt>free</tt>.
<p>
The FreeType source code never invokes directly the function pointers. Rather,
it calls <tt>FT_Alloc</tt>, <tt>FT_Realloc</tt> and <tt>FT_Free</tt> functions
which are defined in "<tt>freetype2/src/base/ftobjs.c</tt>". These will not be
discussed here.
<p>
<b>If you want to use your own memory allocator</b> rather than the one provided
by your build of FreeType, follow these simple steps:<p>
<ol>
<li>Create your own <tt>FT_Memory</tt> object, with pointers that map to
your own memory management routines (beware function signatures though).
<p>
<li>Call <tt>FT_Build_Library(memory,&library)</tt>. This will create a new
<tt>FT_Library</tt> object that uses your own <tt>FT_Memory</tt> exclusively.
Note however that this library has no font drivers loaded in !!
<p>
<li>Load the default font drivers into the new library, either by
calling <tt>FT_Default_Drivers(library)</tt>, or by adding them manually
through repeated calls to <tt>FT_Add_Driver(library,&driver_interface)</tt>
<p>
</ol>
This will replace the <tt>FT_Init_FreeType(&library)</tt> call that an application
must do to initialise one library instance.
<p>
Notice that you <em>don't need to recompile FreeType 2 to use your own memory
manager !!</em>.
<p>
</ul>
<p>
<hr>
<p>
<h2>II. Streams</h2>
<ul>
<h3>1. Basic Stream Structure</h3>
<p>
A stream models the array of bytes found in a font file. FreeType 2 separates
streams into two families :<p>
<ul>
<li><b>memory-based streams:</b><br>
when the stream's content is entirely found in memory. This is the
case for ROM font files, or memory-mapped files.
<p>
<li><b>disk-based streams:</b><br>
when the stream isn't directly accessible in memory. This is the
case for local or remote files.
<p>
</ul>
<p>
Note that a stream's nature only determines how FreeType accesses its content, not
the way it is effectively stored. For example, in the case of a compressed font file,
one implementation may choose to uncompress the font in memory, then provide a memory
based stream to access it. Another one might chose a disk based stream to perform
on-the-fly decompression of the font data. Similarly, the font file can be stored
on a local disk, or obtained from a network. This will be completely transparent to
FreeType.
<p>
The stream structure is:
<p>
<ul>
<tt>typedef struct</tt><br>
<tt>{</tt><br>
<ul><table>
<tr><td><tt><b>char* base</b></tt> <td> for memory-based streams, the address
of its first byte.
<tr><td><tt><b>ulong size</b></tt> <td> the stream's size in bytes.
<tr><td><tt><b>ulong pos</b></tt> <td> the current stream position in the file
<tr><td><tt><b>descriptor</b></tt><td> a union field used to hold either an
integer file descriptor or pointer.
This field is not used by FreeType
itself, but is left to implementations
of "<tt>ftsystem</tt>"
<tr><td><tt><b>pathname</b></tt> <td> a union field that can hold either an
integer or pointer. It is not used by
FreeType itself, but is left to
implementations. These can put the
file pathname's during debugging for
example.
<tr><td><tt><b>read</b></tt> <td> a pointer to a function used to seek the
stream and/or read a run of bytes from it.
<tr><td><tt><b>close</b></tt><td> a pointer to a function called when the
stream is closed.
<tr><td><tt><b>memory</b></tt> <td> a <tt>FT_Memory</tt> object, which is used
to allocate frames for disk-based streams.
This field is set and used by FreeType.
<tr><td><tt><b>cursor</b></tt> <td> a pointer in memory used when accessing
frames. This is set and used by FreeType.
<tr><td><tt><b>limit</b></tt> <td> a pointer in memory used when accessing
frames. This is set and used by FreeType.
</table></ul>
<tt>} FT_StreamRec, *FT_Stream</tt>
</ul>
<p>
The following important things must be noticed here:<p>
<ul>
<li>The <tt>FT_Stream</tt> type is really a pointer to a <tt>FT_StreamRec</tt>.
This is a normal convention for the FreeType source.
<p>
<li>When the <tt>read</tt> field is non NULL, the stream is considered to be
disk-based. Otherwise, the stream is memory-based, and the <tt>base</tt>
field <em>must</em> be set by "<tt>ftsystem.c</tt>" when the stream is
created.
<p>
<li>The <tt>base</tt> field must be set to 0 when a disk-based stream is created.
However, this field will later be set and used by the FreeType library when
accessing frames of bytes within the font file (of course, this doesn't
happen with memory-based streams).
</ul>
<h3>2. Stream lifecyles</h3>
<p>
Each <tt>FT_Face</tt> needs its own stream to access font data. The most common
way to create a new <tt>FT_Stream</tt> object is to call the function
<tt>FT_New_Face</tt>. This function takes a <em>file pathname</em> argument that
is used to create a new stream object.
<p>
This is possible because each implementation of "<tt>ftsystem.c</tt>" provides
a function called <tt>FT_New_Stream</tt> which takes a file pathname and a
<tt>FT_Stream</tt> pointer as an argument. The function simply opens the file
and initialises the stream structure accordingly. It is called by <tt>FT_New_Face</tt>
to create the face's stream object.
<p>
A stream is only closed when the face is destroyed through <tt>FT_Done_Face</tt>.
Its <tt>close</tt> field function will then be called. Note that the function should
<em>never</em> destroy the <tt>FT_Stream</tt>.
<p>
<h3>3. Using your own streams</h3>
<p>
There are cases where it is interesting to provide your own stream to create
a new face object, rather than rely on the default implementation. For example,
a filepathname, which is a C string, might not be useful on a system where files
are named with a UTF-16 string or via an i-node number of memory address (for ROM files).
<p>
For this purpose, the <tt>FT_Open_Face</tt> is defined. It simply takes a
<tt>FT_Stream</tt> pointer as its second argument, instead of a file pathname (the
stream must be allocated and initialised by you, so be careful).
<p>
Actually, the only thing that <tt>FT_New_Face</tt> does is create a new stream
through <tt>FT_New_Stream</tt>, then call <tt>FT_Open_Face</tt> to create the
face with it.
<p>
Note also that you can use the function <tt>FT_New_Memory_Face</tt> to create
a new font face for a memory-based font file, whose address and size can be passed
as arguments. The function automatically creates the corresponding memory-based
stream and use it to create the face.
<p>
</ul>
<p>
<hr>
<p>
<h2>III. Thread synchronisation</h2>
<ul>
The FreeType library uses no static data. It can be used concurrently by two
thread as long as each one uses its own <tt>FT_Library</tt> instance. Otherwise,
one can very simply synchronize access to a single library instance by using a
mutex to protect each call to one of FreeType's API functions.
<p>
</ul>

871
docs/tutorial/index.html
View File

@ -1,871 +0,0 @@
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type"
content="text/html; charset=iso-8859-1">
<meta name="Author"
content="David Turner">
<title>FreeType 2 Tutorial</title>
</head>
<body text="#000000"
bgcolor="#FFFFFF"
link="#0000EF"
vlink="#51188E"
alink="#FF0000">
<h1 align=center>
FreeType 2.0 Tutorial
</h1>
<h3 align=center>
&copy; 2000 David Turner
(<a href="mailto:david@freetype.org">david@freetype.org</a>)<br>
&copy; 2000 The FreeType Development Team
(<a href="http://www.freetype.org">www.freetype.org</a>)
</h3>
<center>
<table width="70%">
<tr><td>
<hr>
<h2>
Introduction
</h2>
<p>This short tutorial will teach you how to use the FreeType&nbsp;2
library in your own applications.</p>
<hr>
<h3>
1. Header files
</h3>
<p>To include the main FreeType header file, simply say</p>
<font color="blue">
<pre>
#include &lt;freetype/freetype.h&gt;</pre>
</font>
<p>in your application code. Note that other files are available in the
FreeType include directory, most of them being included by
<tt>"freetype.h"</tt>. They will be described later in this
tutorial.</p>
<hr>
<h3>
2. Initialize the library
</h3>
<p>Simply create a variable of type <tt>FT_Library</tt> named, for
example, <tt>library</tt>, and call the function
<tt>FT_Init_FreeType()</tt> as in</p>
<font color="blue">
<pre>
#include &lt;freetype/freetype.h&gt;
FT_Library library;
...
{
...
error = FT_Init_FreeType( &library );
if ( error )
{
... an error occurred during library initialization ...
}
}</pre>
</font>
<p>This function is in charge of the following:</p>
<ul>
<li>
<p>Creating a new instance of the FreeType&nbsp;2 library, and set
the handle <tt>library</tt> to it.</p>
</li>
<li>
<p>Load each modules that FreeType knows about in the library.
This means that by default, your new <tt>library</tt> object is able
to handle TrueType, Type&nbsp;1, CID-keyed & OpenType/CFF fonts
gracefully.</p>
</li>
</ul>
<p>As you can see, the function returns an error code, like most others
in the FreeType API. An error code of&nbsp;0 <em>always</em> means that
the operation was successful; otherwise, the value describes the error,
and <tt>library</tt> is set to NULL.</p>
<hr>
<h3>
3. Load a font face
</h3>
<h4>
a. From a font file
</h4>
<p>Create a new <em>face</em> object by calling <tt>FT_New_Face</tt>.
A <em>face</em> describes a given typeface and style. For example,
"Times New Roman Regular" and "Times New Roman Italic" correspond to
two different faces.</p>
<font color="blue">
<pre>
FT_Library library; /* handle to library */
FT_Face face; /* handle to face object */
error = FT_Init_FreeType( &library );
if ( error ) { ... }
error = FT_New_Face( library,
"/usr/share/fonts/truetype/arial.ttf",
0,
&face );
if ( error == FT_Err_Unknown_File_Format )
{
... the font file could be opened and read, but it appears
... that its font format is unsupported
}
else if ( error )
{
... another error code means that the font file could not
... be opened or read, or simply that it is broken...
}</pre>
</font>
<p>As you can certainly imagine, <tt>FT_New_Face</tt> opens a font
file, then tries to extract one face from it. Its parameters are</p>
<table cellpadding=5>
<tr valign="top">
<td>
<tt><b>library</b></tt>
</td>
<td>
<p>handle to the FreeType library instance where the face object
is created</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>filepathname</b></tt>
</td>
<td>
<p>the font file pathname (standard C string).</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>face_index</b></tt>
</td>
<td>
<p>Certain font formats allow several font faces to be embedded
in a single file.</p>
<p>This index tells which face you want to load. An error will
be returned if its value is too large.</p>
<p>Index 0 always work though.</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>face</b></tt>
</td>
<td>
<p>A <em>pointer</em> to the handle that will be set to describe
the new face object.</p>
<p>It is set to NULL in case of error.</p>
</td>
</tr>
</table>
<p>To know how many faces a given font file contains, simply load its
first face (use <tt>face_index</tt>=0), then see the value of
<tt>face->num_faces</tt> which indicates how many faces are embedded
in the font file.</p>
<h4>
b. From memory
</h4>
<p>In the case where you have already loaded the font file in memory,
you can similarly create a new face object for it by calling
<tt>FT_New_Memory_Face</tt> as in</p>
<font color="blue">
<pre>
FT_Library library; /* handle to library */
FT_Face face; /* handle to face object */
error = FT_Init_FreeType( &library );
if ( error ) { ... }
error = FT_New_Memory_Face( library,
buffer, /* first byte in memory */
size, /* size in bytes */
0, /* face_index */
&face );
if ( error ) { ... }</pre>
</font>
<p>As you can see, <tt>FT_New_Memory_Face()</tt> simply takes a
pointer to the font file buffer and its size in bytes instead of a
file pathname. Other than that, it has exactly the same semantics as
<tt>FT_New_Face()</tt>.</p>
<h4>
c. From other sources (compressed files, network, etc.)
</h4>
<p>There are cases where using a file pathname or preloading the file
in memory is simply not enough. With FreeType&nbsp;2, it is possible
to provide your own implementation of i/o routines.</p>
<p>This is done through the <tt>FT_Open_Face()</tt> function, which
can be used to open a new font face with a custom input stream, select
a specific driver for opening, or even pass extra parameters to the
font driver when creating the object. We advise you to refer to the
FreeType&nbsp;2 reference manual in order to learn how to use it.</p>
<p>Note that providing a custom stream might also be used to access a
TrueType font embedded in a Postscript Type&nbsp;42 wrapper.</p>
<hr>
<h3>
4. Accessing face content
</h3>
<p>A <em>face object</em> models all information that globally describes
the face. Usually, this data can be accessed directly by dereferencing
a handle, like</p>
<table cellpadding=5>
<tr valign="top">
<td>
<tt><b>face->num_glyphs</b></tt>
</td>
<td>
<p>Gives the number of <em>glyphs</em> available in the font face.
A glyph is simply a character image. It doesn't necessarily
correspond to a <em>character code</em> though.</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>face->flags</b></tt>
</td>
<td>
<p>A 32-bit integer containing bit flags used to describe some
face properties. For example, the flag
<tt>FT_FACE_FLAG_SCALABLE</tt> is used to indicate that the face's
font format is scalable and that glyph images can be rendered for
all character pixel sizes. For more information on face flags,
please read the <a href="#">FreeType&nbsp;2 API Reference</a>.</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>face->units_per_EM</b></tt>
</td>
<td>
<p>This field is only valid for scalable formats (it is set to 0
otherwise). It indicates the number of font units covered by the
EM.</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>face->num_fixed_sizes</b></tt>
</td>
<td>
<p>This field gives the number of embedded bitmap <em>strikes</em>
in the current face. A <em>strike</em> is simply a series of
glyph images for a given character pixel size. For example, a
font face could include strikes for pixel sizes 10, 12
and&nbsp;14. Note that even scalable font formats can have
embedded bitmap strikes!</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>face->fixed_sizes</b></tt>
</td>
<td>
<p>this is a pointer to an array of <tt>FT_Bitmap_Size</tt>
elements. Each <tt>FT_Bitmap_Size</tt> indicates the horizontal
and vertical <em>pixel sizes</em> for each of the strikes that are
present in the face.</p>
</td>
</tr>
</table>
<p>For a complete listing of all face properties and fields, please read
the <a href="#">FreeType&nbsp;2 API Reference</a>.<p>
<hr>
<h3>
5. Setting the current pixel size
</h3>
<p>FreeType 2 uses "<em>size objects</em>" to model all
information related to a given character size for a given face.
For example, a size object will hold the value of certain metrics
like the ascender or text height, expressed in 1/64th of a pixel,
for a character size of 12 points.</p>
<p>When the <tt>FT_New_Face</tt> function is called (or one of its
cousins), it <b>automatically</b> creates a new size object for
the returned face. This size object is directly accessible as
<b><tt>face->size</tt></b>.</p>
<p><em>NOTA BENE: a single face object can deal with one or more size
objects at a time, however, this is something that few programmers
really need to do. We have thus have decided to simplify the API for
the most common use (i.e. one size per face), while keeping this
feature available through additional fuctions.</em></p>
<p>When a new face object is created, its size object defaults to the
character size of 10&nbsp;pixels (both horizontally and vertically) for
scalable formats. For fixed-sizes formats, the size is more or less
undefined, which is why you must set it before trying to load a
glyph.</p>
<p>To do that, simply call <tt>FT_Set_Char_Size()</tt>. Here is an
example where the character size is set to 16pt for a 300x300&nbsp;dpi
device:</p>
<font color="blue">
<pre>
error = FT_Set_Char_Size(
face, /* handle to face object */
0, /* char_width in 1/64th of points */
16*64, /* char_height in 1/64th of points */
300, /* horizontal device resolution */
300 ); /* vertical device resolution */</pre>
</font>
<p>You will notice that:</p>
<ul>
<li>
<p>The character width and heights are specified in 1/64th of
points. A point is a <em>physical</em> distance, equaling 1/72th
of an inch, it's not a pixel..<p>
</li>
<li>
<p>The horizontal and vertical device resolutions are expressed in
<em>dots-per-inch</em>, or <em>dpi</em>. You can use 72 or
96&nbsp;dpi for display devices like the screen. The resolution
is used to compute the character pixel size from the character
point size.</p>
</li>
<li>
<p>A value of&nbsp;0 for the character width means "<em>same as
character height</em>", a value of&nbsp;0 for the character height
means "<em>same as character width</em>". Otherwise, it is possible
to specify different char widths and heights.</p>
</li>
<li>
<p>Using a value of 0 for the horizontal or vertical resolution means
72&nbsp;dpi, which is the default.</p>
</li>
<li>
<p>The first argument is a handle to a face object, not a size
object. That's normal, and must be seen as a convenience.</p>
</li>
</ul>
<p>This function computes the character pixel size that corresponds to
the character width and height and device resolutions. However, if you
want to specify the pixel sizes yourself, you can simply call
<tt>FT_Set_Pixel_Sizes()</tt>, as in</p>
<font color="blue">
<pre>
error = FT_Set_Pixel_Sizes(
face, /* handle to face object */
0, /* pixel_width */
16 ); /* pixel_height */</pre>
</font>
<p>This example will set the character pixel sizes to 16x16&nbsp;pixels.
As previously, a value of&nbsp;0 for one of the dimensions means
"<em>same as the other</em>".</p>
<p>Note that both functions return an error code. Usually, an error
occurs with a fixed-size font format (like FNT or PCF) when trying to
set the pixel size to a value that is not listed in the
<tt><b>face->fixed_sizes</b></tt> array.</p>
<hr>
<h3>
6. Loading a glyph image
</h3>
<h4>
a. Converting a character code into a glyph index
</h4>
<p>Usually, an application wants to load a glyph image based on its
<em>character code</em>, which is a unique value that defines the
character for a given <em>encoding</em>. For example, the character
code&nbsp;65 represents the `A' in ASCII encoding.</p>
<p>A face object contains one or more tables, called
<em>charmaps</em>, that are used to convert character codes to glyph
indices. For example, most TrueType fonts contain two charmaps. One
is used to convert Unicode character codes to glyph indices, the other
is used to convert Apple Roman encoding into glyph indices. Such
fonts can then be used either on Windows (which uses Unicode) and
Macintosh (which uses Apple Roman, bwerk). Note also that a given
charmap might not map to all the glyphs present in the font.</p>
<p>By default, when a new face object is created, it lists all the
charmaps contained in the font face and selects the one that supports
Unicode character codes if it finds one. Otherwise, it tries to find
support for Latin-1, then ASCII.</p>
<p>We will describe later how to look for specific charmaps in a face.
For now, we will assume that the face contains at least a Unicode
charmap that was selected during <tt>FT_New_Face()</tt>. To convert a
Unicode character code to a font glyph index, we use
<tt>FT_Get_Char_Index()</tt> as in</p>
<font color="blue">
<pre>
glyph_index = FT_Get_Char_Index( face, charcode );</pre>
</font>
<p>This will look the glyph index corresponding to the given
<tt>charcode</tt> in the charmap that is currently selected for the
face. If charmap is selected, the function simply returns the
charcode.</p>
<p>Note that this is one of the rare FreeType functions that do not
return an error code. However, when a given character code has no
glyph image in the face, the value&nbsp;0 is returned. By convention,
it always correspond to a special glyph image called the <b>missing
glyph</b>, which usually is represented as a box or a space.</p>
<h4>
b. Loading a glyph from the face
</h4>
<p>Once you have a glyph index, you can load the corresponding glyph
image. Note that the glyph image can be in several formats. For
example, it will be a bitmap for fixed-size formats like FNT, FON, or
PCF. It will also be a scalable vector outline for formats like
TrueType or Type&nbsp;1. The glyph image can also be stored in an
alternate way that is not known at the time of writing this
documentation.</p>
<p>The glyph image is always stored in a special object called a
<em>glyph slot</em>. As its name suggests, a glyph slot is simply a
container that is able to hold one glyph image at a time, be it a
bitmap, an outline, or something else. Each face object has a single
glyph slot object that can be accessed as
<b><tt>face->glyph</tt></b>.</p>
<p>Loading a glyph image into the slot is performed by calling
<tt>FT_Load_Glyph()</tt> as in</p>
<font color="blue">
<pre>
error = FT_Load_Glyph(
face, /* handle to face object */
glyph_index, /* glyph index */
load_flags ); /* load flags, see below */</pre>
</font>
<p>The <tt>load_flags</tt> value is a set of bit flags used to
indicate some special operations. The default value
<tt>FT_LOAD_DEFAULT</tt> is&nbsp;0.</p>
<p>This function will try to load the corresponding glyph image
from the face. Basically, this means that:</p>
<ul>
<li>
<p>If a bitmap is found for the corresponding glyph and pixel
size, it will in the slot (embedded bitmaps are always
favored over native image formats, because we assume that
they are higher-quality versions of the same image. This
can be ignored by using the FT_LOAD_NO_BITMAP flag)</p>
</li>
<li>
<p>Otherwise, a native image for the glyph will be loaded.
It will also be scaled to the current pixel size, as
well as hinted for certain formats like TrueType and
Type1.</p>
</li>
</ul>
<p>The field <tt><b>glyph->format</b></tt> describe the format
used to store the glyph image in the slot. If it is not
<tt>ft_glyph_format_bitmap</tt>, one can immediately
convert it to a bitmap through <tt>FT_Render_Glyph</tt>,
as in:</p>
<font color="blue">
<pre>
error = FT_Render_Glyph(
face->glyph, /* glyph slot */
render_mode ); /* render mode */
</pre>
</font>
<p>The parameter <tt>render_mode</tt> is a set of bit flags used
to specify how to render the glyph image. Set it to 0 to render
a monochrome bitmap, or to <tt>ft_render_mode_antialias</tt> to
generate a high-quality (256 gray levels) anti-aliased bitmap
from the glyph image.</p>
<p>Once you have a bitmap glyph image, you can access it directly
through <tt><b>glyph->bitmap</b></tt> (a simple bitmap descriptor),
and position it through <tt><b>glyph->bitmap_left</b></tt> and
<tt><b>glyph->bitmap_top</b></tt>.</p>
<p>Note that <tt>bitmap_left</tt> is the horizontal distance from the
current pen position to the left-most border of the glyph bitmap,
while <tt>bitmap_top</tt> is the vertical distance from the
pen position (on the baseline) to the top-most border of the
glyph bitmap. <em>It is positive to indicate an upwards
distance</em>.</p>
<p>The next section will detail the content of a glyph slot and
how to access specific glyph information (including metrics).</p>
<h4>
c. Using other charmaps
</h4>
<p>As said before, when a new face object is created, it will look for
a Unicode, Latin-1, or ASCII charmap and select it. The currently
selected charmap is accessed via <b><tt>face->charmap</tt></b>. This
field is NULL when no charmap is selected, which typically happens
when you create a new <tt>FT_Face</tt> object from a font file that
doesn't contain an ASCII, Latin-1, or Unicode charmap (rare
stuff).</p>
<p>There are two ways to select a different charmap with FreeType 2.
The easiest is when the encoding you need already has a corresponding
enumeration defined in <tt>&lt;freetype/freetype.h&gt;</tt>, as
<tt>ft_encoding_big5</tt>. In this case, you can simply call
<tt>FT_Select_CharMap</tt> as in:</p>
<font color="blue"><pre>
error = FT_Select_CharMap(
face, /* target face object */
ft_encoding_big5 ); /* encoding.. */
</pre></font>
<p>Another way is to manually parse the list of charmaps for the
face, this is accessible through the fields
<tt><b>num_charmaps</b></tt> and <tt><b>charmaps</b></tt>
(notice the 's') of the face object. As you could expect,
the first is the number of charmaps in the face, while the
second is <em>a table of pointers to the charmaps</em>
embedded in the face.</p>
<p>Each charmap has a few visible fields used to describe it more
precisely. Mainly, one will look at
<tt><b>charmap->platform_id</b></tt> and
<tt><b>charmap->encoding_id</b></tt> that define a pair of
values that can be used to describe the charmap in a rather
generic way.</p>
<p>Each value pair corresponds to a given encoding. For example,
the pair (3,1) corresponds to Unicode. Their list is
defined in the TrueType specification but you can also use the
file <tt>&lt;freetype/ftnameid.h&gt;</tt> which defines several
helpful constants to deal with them..</p>
<p>To look for a specific encoding, you need to find a corresponding
value pair in the specification, then look for it in the charmaps
list. Don't forget that some encoding correspond to several
values pair (yes it's a real mess, but blame Apple and Microsoft
on such stupidity..). Here's some code to do it:</p>
<font color="blue">
<pre>
FT_CharMap found = 0;
FT_CharMap charmap;
int n;
for ( n = 0; n &lt; face->num_charmaps; n++ )
{
charmap = face->charmaps[n];
if ( charmap->platform_id == my_platform_id &&
charmap->encoding_id == my_encoding_id )
{
found = charmap;
break;
}
}
if ( !found ) { ... }
/* now, select the charmap for the face object */
error = FT_Set_CharMap( face, found );
if ( error ) { ... }</pre>
</font>
<p>Once a charmap has been selected, either through
<tt>FT_Select_CharMap</tt> or <tt>FT_Set_CharMap</tt>,
it is used by all subsequent calls to
<tt>FT_Get_Char_Index()</tt>.</p>
<h4>
d. Glyph Transforms:
</h4>
<p>It is possible to specify an affine transformation to be applied
to glyph images when they're loaded. Of course, this will only
work for scalable (vectorial) font formats.</p>
<p>To do that, simply call <tt>FT_Set_Transform</tt>, as in:</p>
<font color="blue"><pre>
error = FT_Set_Transform(
face, /* target face object */
&amp;matrix, /* pointer to 2x2 matrix */
&amp;delta ); /* pointer to 2d vector */
</pre></font>
<p>This function will set the current transform for a given face
object. Its second parameter is a pointer to a simple
<tt>FT_Matrix</tt> structure that describes a 2x2 affine matrix.
The third parameter is a pointer to a <tt>FT_Vector</tt> structure
that describe a simple 2d vector.</p>
<p>Note that the matrix pointer can be set to NULL, (in which case
the identity transform will be used). Coefficients of the matrix
are in 16.16 fixed float units.</p>
<p>The vector pointer can also be set to NULL (in which case a delta
of (0,0) will be used). The vector coordinates are expressed in
1/64th of a pixel (also known as 26.6 fixed floats).</p>
<p><em>NOTA BENE: The transform is applied every glyph that is loaded
through <tt>FT_Load_Glyph</tt>. Note that loading a glyph bitmap
with a non-trivial transform will produce an error..</em></p>
<hr>
<h3>
7. Accessing glyph image data
</h3>
<p>Glyph image data is accessible through <tt><b>face->glyph</b></tt>.
See the definition of the <tt>FT_GlyphSlot</tt> type for more details.
As stated previously, each face has a single glyph slot, where
<em>one</em> glyph image <em>at a time</em> can be loaded. Each time
you call <tt>FT_Load_Glyph()</tt>, you erase the content of the glyph
slot with a new glyph image.</p>
<p>Note however that the glyph slot object itself doesn't change, only
its content, which means that you can perfectly create a "shortcut" to
access it as in</p>
<font color="blue">
<pre>
{
/* shortcut to glyph slot */
FT_GlyphSlot glyph = face->glyph;
for ( n = 0; n &lt; face->num_glyphs; n++ )
{
... load glyph n ...
... access glyph data as glyph->xxxx
}
}</pre>
</font>
<p>The <tt>glyph</tt> variable will be valid until its parent
<tt>face</tt> is destroyed. Here are a few important fields of the
glyph slot:<p>
<table cellpadding=5>
<tr valign="top">
<td>
<tt><b>glyph->format</b></tt>
</td>
<td>
<p>Indicates the type of the loaded glyph image. Can be either
<tt>ft_glyph_format_bitmap</tt>, <tt>ft_glyph_format_outline</tt>,
or other values.</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>glyph->metrics</b></tt>
</td>
<td>
<p>A simple structure used to hold the glyph image's metrics.
Note that <em>most distances are expressed in 1/64th of
pixels!</em> See the API reference or the user guide for a
description of the <tt>FT_Glyph_Metrics</tt> structure.</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>glyph->bitmap</b></tt>
</td>
<td>
<p>If the glyph slot contains a bitmap, a simple
<tt>FT_Bitmap</tt> that describes it. See the API reference or
user guide for a description of the <tt>FT_Bitmap</tt>
structure.</p>
</td>
</tr>
<tr valign="top">
<td>
<tt><b>glyph->outline</b></tt>
</td>
<td>
<p>When the glyph slot contains a scalable outline, this structure
describes it. See the definition of the <tt>FT_Outline</tt>
structure.</p>
</td>
</tr>
</table>
<h3>
8. Rendering glyph outlines into bitmaps
</h3>
<p>You can easily test the format of the glyph image by inspecting the
<tt>face->glyph->format</tt> variable. If its value is
<tt>ft_glyph_format_bitmap</tt>, the glyph image that was loaded is a
bitmap that can be directly blit to your own surfaces through your
favorite graphics library (FreeType&nbsp;2 doesn't provide bitmap
blitting routines, as you may imagine&nbsp;:-)</p>
<p>If the format is <tt>ft_glyph_format_outline</tt> or something else,
the library provides a means to convert such glyph images to bitmaps
through what are called <b>rasters</b>.</p>
<p>On the other hand, if the image is a scalable outline or something
else, FreeType provides a function to convert the glyph image into a
pre-existing bitmap that you will handle to it, named
<tt>FT_Get_Glyph_Bitmap</tt>. Here's a <em>simple</em> example code
that renders an outline into a <b>monochrome</b> bitmap:</p>
<font color="blue">
<pre>
{
FT_GlyphSlot glyph;
... load glyph ...
glyph = face-&gt;glyph; /* shortcut to glyph data */
if ( glyph->format == ft_glyph_format_outline )
{
FT_Bitmap bit;
/* set-up a bitmap descriptor for our target bitmap */
bit.rows = bitmap_height;
bit.width = bitmap_width;
bit.pitch = bitmap_row_bytes;
/* render into a mono bitmap */
bit.pixel_mode = ft_pixel_mode_mono;
bit.buffer = bitmap_buffer;
/* render the outline directly into the bitmap */
error = FT_Get_Glyph_Bitmap( face, &bit );
if ( error ) { ... }
}
}</pre>
</font>
<p>You should note that <b><em><tt>FT_Get_Glyph_Bitmap()</tt> doesn't
create the bitmap</em></b>. It only needs a descriptor, of type
<tt>FT_Bitmap</tt>, and writes directly into it.</p>
<p>Note that the FreeType scan-converter for outlines can also generate
anti-aliased glyph bitmaps with 128 level of grays. For now, it is
restricted to rendering to 8-bit gray-level bitmaps, though this may
change in the future. Here is some code to do just that:</p>
<font color="blue">
<pre>
{
FT_GlyphSlot glyph;
... load glyph ...
glyph = face->glyph; /* shortcut to glyph data */
if ( glyph->format == ft_glyph_format_outline )
{
FT_Bitmap bit;
/* set-up a bitmap descriptor for our target bitmap */
bit.rows = bitmap_height;
bit.width = bitmap_width;
bit.pitch = bitmap_row_bytes;
/* 8-bit gray-level bitmap */
bit.pixel_mode = ft_pixel_mode_gray;
/* MUST be 128 for now */
bit.grays = 128;
bit.buffer = bitmap_buffer;
/* clean the bitmap - IMPORTANT */
memset( bit.buffer, 0, bit.rows*bit.pitch );
/* render the outline directly into the bitmap */
error = FT_Get_Glyph_Bitmap( face, &bit );
if ( error ) { ... }
}
}</pre>
</font>
<p>You will notice that</p>
<ul>
<li>
<p>As previously, <tt>FT_Get_Glyph_Bitmap()</tt> doesn't generate
the bitmap, it simply renders to it.</p>
</li>
<li>
<p>The target bitmap must be cleaned before calling the function.
This is a limitation of our current anti-aliasing algorithm and is
EXTREMELY important.</p>
</li>
<li>
<p>The anti-aliaser uses 128&nbsp;levels of grays exclusively for
now (this will probably change in a near future). This means that
you <b>must</b> set <tt>bit.grays</tt> to&nbsp;128. The generated
image uses values from 0 (back color) to 127 (foreground color).</p>
</li>
<li>
<p>It is <b>not</b> possible to render directly an anti-aliased
outline into a pre-existing gray-level bitmap, or even any
colored-format one (like RGB16 or paletted 8-bits). We will not
discuss this issue in great details here, but the reason is that we
do not want to deal with graphics composition (or alpha-blending)
within FreeType.<p/>
</li>
</ul>
</td></tr>
</table>
</center>
</body>
</html>