Fonts
Font setup
Terminology:
Index, font index: The index of a character in a
font, i.e. the place of the glyph for a character in a font.
Code point: The same as font index.
Character value: The two-byte
(char) value by which a character is represented in
memory and in Unicode. Note that this only works straightforwardly for
the basal plane (BMP) of Unicode, i.e. for characters <=
0xFFFF.
Unicode code point: The same as the character
value.
During compilation for each of the 14 base fonts a class is
generated from the XML font metric files. Each font class contains
the metric information and an encoding table (a
CodePointMapping object). The metric information is
static, the encoding table is an object member.
During compilation also a class
CodePointMapping is generated, which contains the
known encodings as static values. For each known encoding it contains
a table as a static final array of int. The array holds an alternation
of font index and character value; in fact it is a mapping from
table[2i] to table[2i+1], where
table[2i] is the font index and
table[2i+1] is the character.
When an encoding is needed in the process, a
CodePointMapping object is created from the
encoding table. It contains a table (char array)
called latin1Map of the font indices for the
characters of the Latin1 range (0-0xFF) in
character value order. It also contains two tables (char
arrays), called characters and codepoints, for the higher
character values. The table characters contains the character values
in order, and the table codepoints contains the corresponding font
indexes for this encoding in the same order. The characters can be
retrieved from these tables as follows:
char <= 0xFF: index = latin1Map[character]
char > 0xFF:
find i such that characters[i] == char;
then index = codepoints[i]
In the code the characters are retrieved from the
CodePointMapping object with its method
mapChar(char c).
In FOP's preparation stage the fonts are set up in the
method Driver.getContentHandler. It calls the
renderer's method
setupFontInfo(currentDocument). The
Document object currentDocument
(which is the foTreeControl object) is able to
store the font setup info and has methods to access the fonts
registered with it.
The PrintRenderer (PostScript and PDF) then
calls FontSetup.setup(fontInfo, fontList), where
fontInfo is the Document object
and fontList is the list of user configured fonts
registered with the renderer in its member
fontList.
[1] org.apache.fop.fonts.FontSetup.setup (FontSetup.java:98)
[2] org.apache.fop.render.PrintRenderer.setupFontInfo (PrintRenderer.java:77)
[3] org.apache.fop.apps.Driver.getContentHandler (Driver.java:551)
[4] org.apache.fop.apps.Driver.render (Driver.java:602)
[5] org.apache.fop.apps.Driver.render (Driver.java:589)
[6] org.apache.fop.apps.Fop.main (Fop.java:102)
FontSetup.setup takes three actions:
An object is created for each of the base 14 fonts
and registered with the fontInfo object in its
member fonts.
A series of triplets (family, style, weight) is set
up. To each triplet a font is assigned; this font will be used when a
font with the characteristics of that triplet is requested. The
triplets registered with fontInfo in its member
triplets. The member triplets is
a map which uses a string of the form
family,style,weight as a key. There is also a class
FontTriplet, which is not used.
The user configured fonts are added.
In the following listing treeBuilder is
the tree builder object set up in the preparation stage, and
foTreeControl is the document object. The list of
user configured fonts of the renderer is empty, and
the list of used fonts is still empty.
treeBuilder.foTreeControl.fonts = "{
F1=org.apache.fop.fonts.base14.Helvetica@e3c624,
F2=org.apache.fop.fonts.base14.HelveticaOblique@e020c9,
F3=org.apache.fop.fonts.base14.HelveticaBold@13e58d4,
F4=org.apache.fop.fonts.base14.HelveticaBoldOblique@15a6029,
F5=org.apache.fop.fonts.base14.TimesRoman@17494c8,
F6=org.apache.fop.fonts.base14.TimesItalic@1e57e8f,
F7=org.apache.fop.fonts.base14.TimesBold@888e6c,
F8=org.apache.fop.fonts.base14.TimesBoldItalic@d3db51,
F9=org.apache.fop.fonts.base14.Courier@5f6303,
F10=org.apache.fop.fonts.base14.CourierOblique@117f31e,
F11=org.apache.fop.fonts.base14.CourierBold@1d7fbfb,
F12=org.apache.fop.fonts.base14.CourierBoldOblique@5d9084
F13=org.apache.fop.fonts.base14.Symbol@39e5b5,
F14=org.apache.fop.fonts.base14.ZapfDingbats@1b5998f,
}"
treeBuilder.foTreeControl.triplets = "{
Computer-Modern-Typewriter,normal,400=F9,
Courier,italic,400=F10,
Courier,italic,700=F12,
Courier,normal,400=F9,
Courier,normal,700=F11,
Courier,oblique,400=F10,
Courier,oblique,700=F12,
Helvetica,italic,400=F2,
Helvetica,italic,700=F4,
Helvetica,normal,400=F1,
Helvetica,normal,700=F3,
Helvetica,oblique,400=F2,
Helvetica,oblique,700=F4,
Symbol,normal,400=F13,
Times Roman,italic,400=F6,
Times Roman,italic,700=F8,
Times Roman,normal,400=F5,
Times Roman,normal,700=F7,
Times Roman,oblique,400=F6,
Times Roman,oblique,700=F8,
Times,italic,400=F6,
Times,italic,700=F8,
Times,normal,400=F5,
Times,normal,700=F7,
Times,oblique,400=F6,
Times,oblique,700=F8,
Times-Roman,italic,400=F6,
Times-Roman,italic,700=F8,
Times-Roman,normal,400=F5,
Times-Roman,normal,700=F7,
Times-Roman,oblique,400=F6,
Times-Roman,oblique,700=F8,
ZapfDingbats,normal,400=F14,
any,italic,400=F6,
any,italic,700=F8,
any,normal,400=F5,
any,normal,700=F7,
any,oblique,400=F6,
any,oblique,700=F8,
monospace,italic,400=F10,
monospace,italic,700=F12,
monospace,normal,400=F9,
monospace,normal,700=F11,
monospace,oblique,400=F10,
monospace,oblique,700=F12,
sans-serif,italic,400=F2,
sans-serif,italic,700=F4,
sans-serif,normal,400=F1,
sans-serif,normal,700=F3,
sans-serif,oblique,400=F2,
sans-serif,oblique,700=F4,
serif,italic,400=F6,
serif,italic,700=F8,
serif,normal,400=F5,
serif,normal,700=F7,
serif,oblique,400=F6
serif,oblique,700=F8,
}"
treeBuilder.foTreeControl.atModel.renderer.fontList = null
treeBuilder.foTreeControl.usedFonts = "{}"
User configured fonts should be listed in the member
fontList of the renderer. The
objects in the list are EmbedFontInfo objects. They
are created from the path to the metrics file, boolean kerning, the
list of triplets for which this font may be used, the path to the font
file. The triplets are FontTriplet objects. The
list may be created from an Avalon configuration object with
FontSetup.buildFontListFromConfiguration(Configuration
cfg).
>FontSetup.addConfiguredFonts creates a
LazyFont font object from each
EmbedFontInfo object. LazyFont
fonts are not loaded until they are actually used. This makes it
possible to register a large number of fonts at low cost.
Font weights are integers between 100 and 900.
Font.NORMAL and Font.BOLD are
set to 400 and 700, respectively. See
FontUtil.parseCSS2FontWeight.
Classes and interfaces used in the font package
IF FontMetrics
SubIF FontDescriptor
IF MutableFont
Abstract Class TypeFace: FontMetrics
Classes Courier etc.
Abstract Class TypeFace: FontMetrics
Abstract Class CustomFont: FontDescriptor, MutableFont
Abstract Class CIDFont, Class SingleByteFont
Class MultiByteFont (sub CIDFont)
Abstract Class TypeFace: FontMetrics
Abstract Class CustomFont: FontDescriptor, MutableFont
Class SingleByteFont
Abstract Class TypeFace: FontMetrics
Class LazyFont: FontDescriptor
Abstract Class TypeFace: FontMetrics
Class FontMetricsMapper, for AWT fonts
SingleByteFont,
MultiByteFont: A font is not really single or
multibyte. Rather the name SingleByteFont indicates
that the font does not contain more than 256 glyphs; the
implementation is optimized for this. In
MultiByteFont (actually CIDFont Type2) the
implementation is optimized for fonts with an unknown number of
glyphs.