# This is the main manual for PDF::Writer. It is interpreted by PDF::TechBook
# which implements an interpreter for the simple markup language presented
# here.
#
# 0.  Comments, headings, and directives must begin in column 0 and consume
#     the whole line.
# 1.  Comments begin with '#' at the beginning of the line.
# 2.  Directives are in the form '.<directive>[ [<args>] <comment>]', where
#     both args (arguments) and comment are optional. The arguments must be
#     separated from the directive by a space and comments must be separated
#     from the arguments by a space. The format of the arguments is directive
#     dependent. Directive names must begin with a letter, may only contain
#     the word characters.
# 2.1.  .newpage starts a new page.
# 2.2.  .pre starts a "preserved newlines" block of text. This is similar to
#       normal text (see #4) but the lines are not flowed. This is useful for
#       creating lists or other text that must be in a particular form
#       without switching to a code font. .pre blocks may not be nested in
#       any other block (that is to say, when an .endpre is encountered, the
#       text state is set to normal text).
# 2.3.  .endpre ends a preserved newlines block of text.
# 2.4.  .code starts a "code" block of text. Similar to .pre, this changes
#       the display font to the "code" font for the text, which is generally
#       a Courier font. .code blocks may be nested in .pre blocks (when an
#       .endcode is encountered, the text state is restored to the previous
#       text state, either normal or preserved newlines).
# 2.5.  .endcode ends a code block of text.
# 2.6.  .eval starts an eval block. Text between this line and .endeval will
#       be collected and passed to Ruby's Kernel#eval at $SAFE level 3. .eval
#       blocks may be nested in .pre or .code blocks.
# 2.7.  .endeval ends an eval block.
# 2.8.  .done stops processing on the document.
# 2.9.  .columns turns on and off column output. The number of columns is
#       specified as an argment, e.g., ".columns 3". Turn columns off with
#       ".columns 1" or ".columns 0". A second parameter may be provided
#       specifying the gutter width (e.g., ".columns 3 15").
# 2.10. .toc indicates that the table of contents should be generated with
#       the argument value as the title of the table of contents page. This
#       is an advisory value only. If this is not present, a table of
#       contents will not be automatically generated.
# 2.11. .author, .title, .subject, and .keywords set values in the PDF info
#       object. The arguments are used as the values.
# 2.12. .blist starts a list block. Lists may be nested within each other,
#       and will successively shift the left margin by 10% of the initial
#       left margin value. Each line in the list block will be treated as a
#       separate list item with a bullet inserted in front. This will be
#       inserted as a callback (e.g., <C:bullet/> or <C:disc/>) before the
#       text of the list item.
# 2.13. .endblist will end the current list block.
# 3.  Headings are added to the text with the form '#<heading-text>' where #
#     is the level of the heading. The only heading levels currently
#     supported are 1 - 5. Heading levels 1 and 2 will automatically be added
#     to the table of contents. Headings *may* include markup (see #5), but
#     should not exceed a single line (the heading box is not currently set
#     up for this). All heading levels may also have an optional name
#     comprised word characters (a-zA-Z0-9_) for automated cross-reference
#     building. The form will be like this: '1<chapter>name'. If the name is
#     not provided, the cross-reference identifier will be the position of
#     the identifier.
# 3.1.  1<chapter> is a first-level heading with text of "chapter". Usually
#       used for chapter headings.
# 3.2.  2<section> is a second-level heading with text of "section". Usually
#       used for section headings.
# 3.3.  3<section>, 4<section>, and 5<section>, are headings that use the
#       text of "section" and identify various subsections of the text.
# 3.4.  Automated cross-reference handling is achived with the replacing text
#       directive 'xref': <r:xref name='name' label='labeltype'>. The name
#       will be the name associated with the heading level and labeltype will
#       be either 'page', 'title', or 'text'. The cross-reference directive
#       is replaced with either the page number or the section label. The
#       page type can only be replaced if the cross-reference target has
#       already been placed on a document page.
# 4. Normal text is treated as flowed text. That is, two lines back to back
#    will be joined into a single line and then wrapped by the text
#    formatting within PDF::Writer. Paragraphs are separated by an empty
#    line, a valid directive, or a heading.
# 5. Within normal, preserved, code text, or headings, HTML-like markup may
#    be used for bold (<b></b>), italic (<i></i>), superscript (<sup></sup>),
#    and subscript (<sub></sub>).
#    Other markup options may be used with the callback tag. Clickable
#    external hyperlinks may be added by using the 'alink' callback:
#         <c:alink uri="URI">LABEL</c:alink>
# 6. Special characters (e.g., <, &, and >) must be escaped in the HTML
#    style.
#
# Please note that this syntax will be evolving in future versions of
# PDF::Writer.
#
# Austin Ziegler
# Tuesday, 3 May 2005
#--
# We want a table of contents for this document.
.toc Table of Contents
# Set the PDF document info.
.author Austin Ziegler
.title  PDF::Writer for Ruby Manual
.subject The use and care of PDF::Writer for Ruby.
# Use .eval to do encoding configuration.
.eval {{{
pdf.techbook_encoding  = {
  :encoding     => "WinAnsiEncoding",
  :differences  => {
    169 => "copyright",
    215 => "multiply",
  }
}
.endeval }}}
# Use .eval to create the heading and footing.
.eval {{{
pdf.select_font     "Helvetica", pdf.techbook_encoding
  # Put a heading on all pages.
pdf.open_object do |heading|
  pdf.save_state
  pdf.stroke_color! Color::RGB::Black
  pdf.stroke_style! PDF::Writer::StrokeStyle::DEFAULT

  s = 6
  t = "PDF::Writer for Ruby ~ Manual"
  w = pdf.text_width(t, s) / 2.0
  x = pdf.margin_x_middle
  y = pdf.absolute_top_margin
  pdf.add_text(x - w, y, t, s)

  x = pdf.absolute_left_margin
  w = pdf.absolute_right_margin
  y -= (pdf.font_height(s) * 1.01)
  pdf.line(x, y, w, y).stroke

  pdf.restore_state
  pdf.close_object
  pdf.add_object(heading, :all_following_pages)
end

  # Put a footing on all pages.
pdf.open_object do |footing|
  pdf.save_state
  pdf.stroke_color! Color::RGB::Black
  pdf.stroke_style! PDF::Writer::StrokeStyle::DEFAULT

  s = 6
  t = "http://rubyforge.org/projects/ruby-pdf"
  x = pdf.absolute_left_margin
  y = pdf.absolute_bottom_margin
  pdf.add_text(x, y, t, s)

  x = pdf.absolute_left_margin
  w = pdf.absolute_right_margin
  y += (pdf.font_height(s) * 1.05)
  pdf.line(x, y, w, y).stroke

  pdf.restore_state
  pdf.close_object
  pdf.add_object(footing, :all_following_pages)
end
.endeval }}}
# Use .eval to create the title page and start page numbering.
.eval {{{
  # Get the info we need for placing page numbering after we draw the title
  # page.
pny = pdf.absolute_bottom_margin
pnx = pdf.absolute_right_margin

pdf.margins_pt(54)    # 54 point (3/4") margins on all sides...
pdf.left_margin = 72  # ...except the left margin (1")

image = File.join(pdf.techbook_source_dir, "images", "bluesmoke.jpg")
pdf.add_image_from_file(image, 300, 50)

  # Draw the title page.
title = "PDF::Writer for Ruby"
size  = 72
fh    = pdf.font_height(size) * 1.01
fd    = pdf.font_descender(size) * 1.01
pdf.save_state
pdf.fill_color    Color::RGB::Red
pdf.stroke_color  Color::RGB::Red 
pdf.rectangle(pdf.absolute_left_margin, 0, fh, pdf.page_height).fill
pdf.fill_color    Color::RGB::White
pdf.stroke_color  Color::RGB::White
pdf.add_text(pdf.absolute_left_margin + fh + fd, 70, title, size, 90)
pdf.restore_state

pdf.select_font   pdf.techbook_textfont, pdf.techbook_encoding
pdf.text("\nNative Ruby PDF Document Creation\n", :font_size => 24, :justification => :right)

info = <<-"INFO"

The Ruby PDF Project
<c:alink uri="http://rubyforge.org/projects/ruby-pdf/">http://rubyforge.org/projects/ruby-pdf</c:alink>
version #{PDF::Writer::VERSION}
INFO

pdf.text(info, :font_size => 20, :justification => :right)

info = <<-INFO

Copyright © 2003–2005
<c:alink uri="mailto:austin@rubyforge.org">Austin Ziegler</c:alink>
INFO

pdf.text(info, :font_size => 18, :justification => :right)
pdf.open_here("Fit")

pdf.techbook_fontsize = 12
.endeval }}}
.newpage
.eval {{{
pnx = pdf.absolute_right_margin
pdf.start_page_numbering(pnx, 36, 6, :right, "<PAGENUM>", 1)
.endeval }}}
# The main document.
1<Introduction to PDF::Writer for Ruby>Introduction

PDF::Writer is designed to provide a pure Ruby way to dynamically create PDF
documents. Obviously, this will not be as fast as one that uses a compiled
extension, but it is surprisingly fast. This manual is, in some ways, a worst
case scenario because of the number of examples that must be displayed.

PDF::Writer does not yet implement the full PDF specification (any version),
but it implements a substantial portion of it, with more to come. It also
implements a more complex document layer on top of the portion that is
implemented.

This manual (manual.pdf) is generated by PDF::TechBook using the application
runner (bin/techbook) and the text version of the manual (manual.pwd). It is
a comprehesive introduction to the use of PDF::Writer and a simple markup
language interpreter. PDF::Writer itself only implements a few markup items,
mostly relating to bold and italic text.

PDF::Writer is based on Adobe’s PDF Reference, Fifth Edition, version 1.6.
This and earlier editions are available from the
<c:alink
uri="http://partners.adobe.com/public/developer/pdf/index_reference.html">Adobe
website</c:alink>. The original implementation was a port of the public
domain <c:alink uri="http://www.ros.co.nz/pdf/">R&OS PDF Class for PHP</c:alink>.

Other demo programs are available in the <b>demo/</b> directory.
Alternatively, they may be downloaded separately from the <c:alink
uri="http://rubyforge.org/projects/ruby-pdf">Ruby PDF Tools</c:alink> project on RubyForge.

2<Installation>
If this manual was generated from a local installation of PDF::Writer, then
congratulations! PDF::Writer is installed. Otherwise, you are reading a
manual generated otherwise. If you want to install PDF::Writer, you can
download it from the Ruby PDF Tools project on RubyForge or install it with
RubyGems.

PDF::Writer has dependencies on <c:alink
uri="http://rubyforge.org/frs/?group_id=295&release_id=2130">Transaction::Simple
1.3.0</c:alink> or later and <c:alink
uri="http://rubyforge.org/frs/?group_id=81&release_id=2127">color-tools
1.0.0</c:alink> or later. These must be installed for PDF::Writer to work.
RubyGems installation will automatically detect and offer to download and
install these dependencies.

PDF::Writer is installed with:

.code {{{
  % ruby setup.rb
.endcode }}}
.newpage
1<Making Documents with PDF::Writer>MakingDocuments

Document creation with PDF::Writer is quite easy. The following code will
create a single-page document that contains the phrase “Hello, Ruby.”
centered at the top of the page in 72-point (1”) type.

.code {{{
  # This code is demo/hello.rb
require "pdf/writer"

pdf = PDF::Writer.new
pdf.select_font "Times-Roman"
pdf.text "Hello, Ruby.", :font_size =&gt; 72, :justification =&gt; :center

File.open("hello.pdf", "wb") { |f| f.write pdf.render }
.endcode }}}

This one, on the other hand, uses a very (in)famous phrase and image. Note
that the images are JPEG and PNG—PDF::Writer cannot use GIF images.

.code {{{
  # This code is demo/chunkybacon.rb
require "pdf/writer"

pdf = PDF::Writer.new
pdf.select_font "Times-Roman"
pdf.text "Chunky Bacon!!", :font_size =&gt; 72, :justification =&gt; :center

  # PDF::Writer#image returns the image object that was added.
i0 = pdf.image "../images/chunkybacon.jpg", :resize =&gt; 0.75
pdf.image "../images/chunkybacon.png",
          :justification =&gt; :center, :resize =&gt; 0.75

  # It can reinsert an image if wanted.
pdf.image i0, :justification =&gt; :right, :resize =&gt; 0.75

pdf.text "Chunky Bacon!!", :font_size =&gt; 72, :justification =&gt; :center

File.open("chunkybacon.pdf", "wb") { |f| f.write pdf.render }
.endcode }}}

If those aren’t enough to whet your appetite for how easy PDF::Writer can be,
well, no chunky bacon for you!
.newpage
1<Fundamental PDF Concepts>FundamentalConcepts

This section covers fundamental concepts to the creation of PDF documents
that will assist in the understanding of the details that follow in later
sections.

2<PDF Coordinate Space and User Units>CoordinateSpace
PDF documents don’t use inches or millimetres for measurement, and the
coordinate space is slightly different than that used by other graphics
canvases.

3<PDF User Units>
PDF user units are by default points. A modern point is exactly 1/72” (1/72
inch); it is generally accepted that 72 point type is one inch in height.
Historically, one point was 0.0138”, or just under 1/72”.

.eval {{{
pdf.select_font("Helvetica")

PDF::SimpleTable.new do |tab|
  tab.title = "PDF User Unit Conversions"
  tab.column_order.push(*%w(from1 to1 from2 to2))

  tab.columns["from1"] = PDF::SimpleTable::Column.new("from1") { |col|
    col.heading = "From"
  }
  tab.columns["to1"] = PDF::SimpleTable::Column.new("to1") { |col|
    col.heading = "To"
  }
  tab.columns["from2"] = PDF::SimpleTable::Column.new("from2") { |col|
    col.heading = "From"
  }
  tab.columns["to2"] = PDF::SimpleTable::Column.new("to2") { |col|
    col.heading = "To"
  }

  tab.show_lines    = :all
  tab.show_headings = true
  tab.orientation   = :center
  tab.position      = :center

  data = [
    { "from1" => "1 point", "to1" => "0.3528 mm",
      "from2" => "1 point", "to2" => "1/72”" },
    { "from1" => "10 mm", "to1" => "28.35 pts",
      "from2" => "", "to2" => "" },
    { "from1" => "A4", "to1" => "210 mm × 297 mm",
      "from2" => "A4", "to2" => "595.28 pts × 841.89 pts" },
    { "from1" => "LETTER", "to1" => "8½” × 11”",
      "from2" => "LETTER", "to2" => "612 pts × 792 pts" },
  ]

  tab.data.replace data
  tab.render_on(pdf)
end
.endeval }}}

It is possible to scale the user coordinates with PDF::Writer#scale_axis.
With scale_axis(0.5, 0.5), each coordinate is ½ point in size (1/144”) in
both the x and y directions. See the following discussion on PDF Coordinate
Space for the limitation of performing transformations on the coordinate
axis.

PDF::Writer provides utility methods to convert between normal measurement
units and and points. These utility methods are available either as class
mehods or as instance methods.
.eval {{{
$rm = pdf.right_margin
pdf.right_margin += 175

y = pdf.absolute_bottom_margin + 40
pdf.stroke_style(PDF::Writer::StrokeStyle.new(1))
pdf.rectangle(400, y, 150, 200).stroke
x0y0 = "(0, 0)"
hh = pdf.font_height(9) + pdf.font_descender(9)
pdf.add_text(405, y + hh - 2, x0y0, 9)
x1y1 = "(150, 200)"
ww = pdf.text_width(x1y1, 9)
pdf.add_text(545 - ww, y + 198 - hh, x1y1, 9)
.endeval }}}

4<PDF::Writer.cm2pts(x)>
    Converts from centimetres to points.

4<PDF::Writer.in2pts(x)>
    Converts from inches to points.

4<PDF::Writer.mm2pts(x)>
    Converts from millimetres to points.

3<PDF Coordinate Space>PDFCoordinateSpace
The standard PDF coordinate space is a little different than might be
expected. Most graphics canvases place the base coordinate (0, 0) in the
upper left-hand corner of the canvas; the common PDF coordinate sytem places
the base coordinate in the lower left-hand corner of the canvas.
.eval {{{
pdf.right_margin = $rm
.endeval }}}

PDF::Writer uses the standard coordinate space. Any transformations on the
coordinate space (or, more accurately, on the coordinate axis) may affect the
usability of PDF::Writer. Standard transformations available against the
coordinate space axis are rotations (counter-clockwise), scaling, translation
(repositioning of the base coordinate) and skew. Each of these will be
discussed in more detail in a later section of this manual, after other
concepts have been introduced.

<b>NOTE:</b> As of PDF::Writer 1.1.0, angle rotations are now
counter-clockwise, not clockwise as in earlier versions. This is a necessary
incompatible change to make transformations more compatible with other vector
graphics systems.

2<Fonts, Special Characters, and Character Maps in PDF::Writer>FontBasics
All PDF readers support fourteen standard fonts; twelve of these fonts are
bold, italic (or oblique), and bold-italic (bold-oblique) variants of three
base font families and the other two are symbolic fonts.

.eval {{{
pdf.select_font("Helvetica")

PDF::SimpleTable.new do |tab|
  tab.title = "Default Fonts in PDF"
  tab.column_order.push(*%w(family name filename))

  tab.columns["family"] = PDF::SimpleTable::Column.new("family") { |col|
    col.heading = "Family"
  }
  tab.columns["name"] = PDF::SimpleTable::Column.new("name") { |col|
    col.heading = "Name"
  }
  tab.columns["filename"] = PDF::SimpleTable::Column.new("filename") { |col|
    col.heading = "Filename"
  }

  tab.show_lines      = :outer
  tab.show_headings   = true
  tab.shade_headings  = true
  tab.orientation     = :center
  tab.position        = :center

  data = [
    { "family"    => "Courier",
      "name"      => "Courier",
      "filename"  => "Courier.afm" },
    { "family"    => "",
      "name"      => "Courier-Bold",
      "filename"  => "Courier-Bold.afm" },
    { "family"    => "",
      "name"      => "Courier-Oblique",
      "filename"  => "Courier-Oblique.afm" },
    { "family"    => "",
      "name"      => "Courier-BoldOblique",
      "filename"  => "Courier-BoldOblique.afm" },
    { "family"    => "Helvetica",
      "name"      => "Helvetica",
      "filename"  => "Helvetica.afm" },
    { "family"    => "",
      "name"      => "Helvetica-Bold",
      "filename"  => "Helvetica-Bold.afm" },
    { "family"    => "",
      "name"      => "Helvetica-Oblique",
      "filename"  => "Helvetica-Oblique.afm" },
    { "family"    => "",
      "name"      => "Helvetica-BoldOblique",
      "filename"  => "Helvetica-BoldOblique.afm" },
    { "family"    => "Symbol",
      "name"      => "Symbol",
      "filename"  => "Symbol.afm" },
    { "family"    => "Times-Roman",
      "name"      => "Times-Roman",
      "filename"  => "Times-Roman.afm" },
    { "family"    => "",
      "name"      => "Times-Bold",
      "filename"  => "Times-Bold.afm" },
    { "family"    => "",
      "name"      => "Times-Italic",
      "filename"  => "Times-Italic.afm" },
    { "family"    => "",
      "name"      => "Times-BoldItalic",
      "filename"  => "Times-BoldItalic.afm" },
    { "family"    => "ZapfDingbats",
      "name"      => "ZapfDingbats",
      "filename"  => "ZapfDingbats.afm" },
  ]

  tab.data.replace data
  tab.render_on(pdf)
end
.endeval }}}

In addition to these fourteen standard fonts, PDF supports the embedding of
PostScript Type 1, TrueType, and OpenType fonts. PDF::Writer explicitly
supports Type 1 and TrueType fonts. Inasmuch as OpenType fonts may be
compatible with TrueType fonts, they may be supported. Embedded fonts require
font metrics information. See “<r:xref name="EmbeddingFonts" label="title"
/>”, “<r:xref name="Type1Fonts" label="title" />” and “<r:xref
name="TrueTypeFonts" label="title" />” below for more information.

PDF::Writer will find (or attempt to find) system font location(s) so that
any existing Type 1 or TrueType fonts can be loaded directly from the system
locations. These paths are found in <b>PDF::Writer::FONT_PATH</b>. By
modifying this value, it is possible to tell PDF::Writer to find font files
in alternate locations. Adobe font metrics (AFM) files are found in the paths
described by <b>PDF::Writer::FontMetrics::METRICS_PATH</b>, the current
directory, and the Ruby <b>$LOAD_PATH</b>.

Fonts are selected with PDF::Writer#select_font. This will load the
appropriate font metrics file and—if the font is not one of the fourteen
built-in fonts—search for the associated Type 1 or TrueType font to embed in
the generated PDF document.

3<PDF::Writer#select_font(font_name, encoding = nil)>
This selects, and possibly loads, a font to be used from this point in the
document. The font name is as noted above. The format and the use of the
<b>encoding</b> parameter is discussed extensively in “<r:xref
name="CharMaps" label="title" />”.

The encoding directive will be effective <b>only when the font is first
loaded</b>.

.code {{{
  # use a Times-Roman font with MacExpertEncoding
pdf.select_font("Times-Roman", "MacExpertEncoding")
  # this next line should be equivalent
pdf.select_font("Times-Roman", { :encoding =&gt; "MacExpertEncoding" })

  # Set up the Helvetica font for use with German characters as an offset
  # of the WinAnsiEncoding.
diff= {
  196 =&gt; "Adieresis",
  228 =&gt; "adieresis",
  214 =&gt; "Odieresis",
  246 =&gt; "odieresis",
  220 =&gt; "Udieresis",
  252 =&gt; "udieresis",
  223 =&gt; "germandbls"
}
pdf.select_font("Helvetica", { :encoding =&gt; "WinAnsiEncoding", :differences => diff })
.endcode }}}

3<Font Families>FontFamilies
It is possible to define font families in PDF::Writer so that text state
translations may be performed. The only text state translations that
PDF::Writer currently recognises are ‘b’ (bold), ‘i’ (italic), ‘bi’ (bold
italic), and ‘ib’ (italic bold).

4<PDF::Writer#font_families>
PDF::Writer#font_families is a Hash that maps the default font name (such as
“Helvetica”) to a mapping of text states and font names, as illustrated
below.

.code {{{
pdf = PDF::Writer.new

  # Note: "bi" (&lt;b&gt;&lt;i&gt;) can be implemented as a
  # different font than "ib" (&lt;i&gt;&lt;b&gt;).
pdf.font_families["Helvetica"] =
{
  "b"   =&gt; "Helvetica-Bold",
  "i"   =&gt; "Helvetica-Oblique",
  "bi"  =&gt; "Helvetica-BoldOblique",
  "ib"  =&gt; "Helvetica-BoldOblique"
}
.endcode }}}

The use of font families will allow the in-line switching between text states
and make it unnecessary to use <b>#select_font</b> to change between these
fonts. This will also ensure that the encoding and differences (see “<r:xref
name="SpecialCharacters" label="title" />” below) for the selected font family
will be consistent. The default font families detailed above (for Helvetica,
Times Roman, and Courier) have already been defined for PDF::Writer.

3<Embedding Fonts>EmbeddingFonts
PDF::Writer will properly embed both TrueType and PostScript Type 1 fonts,
but these fonts will require Adobe Font Metrics (AFM) files (extension .afm).
These files should exist for PostScript Type 1 files but may not exist for
TrueType files on any given system.

4<Type 1 Fonts>Type1Fonts
PostScript Type 1 fonts are fully supported by PDF, but only the binary
(.pfb) forms of Type 1 fonts may be embedded into a PDF document. If only
the ASCII (.pfa) form of the Type 1 font is available, then the program
“t1binary” available as part of the Type 1 utilities package found at the
<c:alink uri="http://www.lcdf.org/~eddietwo/type/#t1utils">LCDF type software
page</c:alink> will create the binary form from the ASCII form. There is a
suggestion on that page that older Macintosh Type 1 fonts will need
additional conversion with one of the provided utilities.

4<TrueType Fonts>TrueTypeFonts
AFM files can be generated for TrueType fonts with the program “ttf2afm”,
which is part of the package <c:alink
uri="http://www.tug.org/applications/pdftex/">pdfTex</c:alink> by Han The
Thanh. In a future release of PDF::Writer or another program from the Ruby
PDF project, this requirement should be eliminated.

4<Embedded Font Licensing Restrictions>
As noted above, PDF::Writer will embed Type 1 or TrueType font programs in
the PDF document. Fonts are recognised as copyrightable intellectual property
in some jurisdictions. TrueType fonts encode some licensing rights in the
font programs and PDF::Writer will check and warn if fonts not licensed for
inclusion in a document are selected. It is up to the users of PDF::Writer to
ensure that there are no licence violations of fonts embedded in the
generated PDF documents.

3<Special Characters, Character Maps, and Unicode>SpecialCharacters
Fonts in PDF documents can include encoding information. The PDF standard
encodings are ‘none’, ‘WinAnsiEncoding’, ‘MacRomanEncoding’, or
‘MacExpertEncoding’. Symbolic fonts should be encoded with the ‘none’
encoding. The default encoding used in PDF::Writer is ‘WinAnsiEncoding’.

‘WinAnsiEncoding’ encoding is not quite the same as Windows code page 1252
(roughly equivalent to latin 1). Appendix D of the Adobe PDF Reference
version 1.6 contains full information on all encoding mappings (including
mappings for the two included symbolic fonts).

4<Encoding Character Maps>CharMaps
The main way of providing particular character support with PDF::Writer is
through a differences map, or a character subtitution table. This is done
only when initially selecting the font; it will not be reapplied after the
font has been loaded from disk once (this limitation applies to the fourteen
built-in fonts as well).

.code {{{
encoding = {
  :encoding     =&gt; "WinAnsiEncoding",
  :differences  =&gt; {
    215 =&gt; "multiply",
    148 =&gt; "copyright",
  }
}

pdf.select_font("Helvetica", encoding)
.endcode }}}

The above code maps the bytes 215 (0xd7) and 148 (0x94) to the named
characters “multiply” and “copyright” in the Helvetica font. These byte
values are the characters “©” and “×” in Windows 1252 but are undefined in
the font and therefore require additional information to present and space
these characters properly.

#TODO
As of PDF::Writer version 1.1, these difference maps will be inserted into
the PDF documents as well. This change is experimental but should be safe.

4<Unicode>
PDF supports UTF-16BE encoding of strings—but each such string must begin
with the UTF-16BE byte order mark (BOM) of U+FEFF (0xFE followed by 0xFF).
PDF::Writer does not, at this point, explicitly support either UTF-8 or
UTF-16BE strings. If all of the conditions are correct, the following code
should display the Japanese characters for “Nihongo” (the name of the
Japanese language). As of 2005.05.02, this will not work (at least in this
manual).

.code {{{
pdf.text("\xfe\xff\x65\xe5\x67\x2c\x8a\x9e")
.endcode }}}
# .eval {{{
# pdf.text("\xfe\xff\x65\xe5\x67\x2c\x8a\x9e", :justification => :center)
# .endeval }}}

There are additional limitations and features for Unicode support, but a
later version of PDF::Writer will support Unicode (at least UTF-8) documents.
.newpage
1<Working with PDF Documents>WorkingWith

The PDF::Writer class is used to create PDF documents in Ruby. It is a
“smart” writing canvas on which both text and graphics may be drawn. <r:xref
name="GraphicsOps" label="title" /> covers the methods that are used to draw
graphics in PDF::Writer. This chapter covers text drawing and most other PDF
document operations.

The canvas provided by PDF::Writer is described as a “smart” canvas because
it is aware of common text writing conventions including page margins and
multi-column output.

2<Creating PDF Documents>CreatingDocuments
There are two ways to create PDF documents with PDF::Writer. Both will
provide instances of PDF::Writer; these are <b>PDF::Writer.new</b> and
<b>PDF::Writer.prepress</b>.

3<PDF::Writer.new>
This method creates a new PDF document as a writing canvas. Without any
parameters, it will create a PDF version 1.3 document that uses pages that
are standard US Letter (8½” × 11”) in portrait orientation. It accepts three
named parameters, <b>:paper</b>, <b>:version</b>, and <b>:orientation</b>.

.code {{{
require "pdf/writer"

  # US Letter, portrait, 1.3
PDF::Writer.new
  # 1. A4, portrait, 1.3
PDF::Writer.new(:paper =&gt; "A4")
  # 2. 150cm × 200cm, portrait, 1.3
PDF::Writer.new(:paper =&gt; [ 150, 200 ])
  # 3. 150pt × 200pt, portrait, 1.3
PDF::Writer.new(:paper =&gt; [ 0, 0, 150, 200 ])
  # US Letter, landscape, 1.3
PDF::Writer.new(:orientation =&gt; :landscape)
  # US Letter, portrait, 1.5
PDF::Writer.new(:version =&gt; PDF_VERSION_15)
.endcode }}}

The <b>:paper</b> parameter specifies the size of a page in PDF::Writer. It
may be specified as: (1) a standard paper size name (see the table
“PDF::Writer Page Sizes” below for the defined paper size names); (2) a
<b>[width, height]</b> array measured in centimetres; or (3) a <b>[x0, y0,
x1, y1]</b> array measured in points where <b>(x0, y0)</b> represents the
lower left-hand corner and <b>(x1, y1)</b> represent the upper right-hand
corner of the page.

The <b>:orientation</b> parameter specifies whether the pages will be
<b>:portrait</b> (the long edge is the height of the page) or
<b>:landscape</b> (the long edge is the width of the page). These are the
only allowable values.

The <b>:version</b> parameter specifies the minimum specification version
that the document will adhere to. As of this version, the PDF document
version is inserted but not used to control what features may be inserted
into a document. The latest version of the PDF specification is PDF 1.6
(associated with Adobe Acrobat 7); it is recommended that the default version
(1.3) be kept in most cases as that will ensure that most users will have
access to the features in the document. A later version of PDF::Writer will
include version controls so that if when creating a PDF 1.3 document,
features from PDF 1.6 will not be available. PDF::Writer currently supports
features from PDF 1.3 and does not yet provide access to the advanced
features of PDF 1.4 or higher.

.eval {{{
pdf.select_font("Helvetica")

PDF::SimpleTable.new do |tab|
  tab.title = "PDF::Writer Page Sizes"
  tab.column_order.replace %w(name1 size1 name2 size2)

  tab.columns["name1"] = PDF::SimpleTable::Column.new("name1") { |col|
    col.heading = PDF::SimpleTable::Column::Heading.new("Type")
  }
  tab.columns["size1"] = PDF::SimpleTable::Column.new("size1") { |col|
    col.heading = PDF::SimpleTable::Column::Heading.new("Size")
  }
  tab.columns["name2"] = PDF::SimpleTable::Column.new("name2") { |col|
    col.heading = PDF::SimpleTable::Column::Heading.new("Type")
  }
  tab.columns["size2"] = PDF::SimpleTable::Column.new("size2") { |col|
    col.heading = PDF::SimpleTable::Column::Heading.new("Size")
  }

  tab.show_lines      = :all
  tab.show_headings   = true
  tab.orientation     = :center
  tab.position        = :center

  data = [
    { "name1" => "2A0", "size1" => '46.811" x 66.220" (118.9cm x 168.2cm)', "name2" => "4A0", "size2" => '66.220" x 93.622" (168.2cm x 237.8cm)' },
    { "name1" => "A0",  "size1" => '33.110" x 46.811" (84.1cm x 118.9cm)', "name2" => "A1",  "size2" => '23.386" x 33.110" (59.4cm x 84.1cm)' },
    { "name1" => "A2",  "size1" => '16.535" x 23.386" (42cm x 59.4cm)', "name2" => "A3",  "size2" => '11.693" x 16.535" (29.7cm x 42cm)' },
    { "name1" => "A4",  "size1" => '8.268" x 11.693" (21cm x 29.7cm)', "name2" => "A5",  "size2" => '5.827" x 8.268" (14.8cm x 21cm)' },
    { "name1" => "A6",  "size1" => '4.134" x 5.827" (10.5cm x 14.8cm)', "name2" => "A7",  "size2" => '2.913" x 4.134" (7.4cm x 10.5cm)' },
    { "name1" => "A8",  "size1" => '2.047" x 2.913" (5.2cm x 7.4cm)', "name2" => "A9",  "size2" => '1.457" x 2.047" (3.7cm x 5.2cm)' },
    { "name1" => "A10", "size1" => '1.024" x 1.457" (2.6cm x 3.7cm)' },
    { "name1" => "B0",  "size1" => '39.370" x 55.669" (100cm x 141.4cm)', "name2" => "B1",  "size2" => '27.835" x 39.370" (70.7cm x 100cm)' },
    { "name1" => "B2",  "size1" => '19.685" x 27.835" (50cm x 70.7cm)', "name2" => "B3",  "size2" => '13.898" x 19.685" (35.3cm x 50cm)' },
    { "name1" => "B4",  "size1" => '9.842" x 13.898" (25cm x 35.3cm)', "name2" => "B5",  "size2" => '6.929" x 9.842" (17.6cm x 25cm)' },
    { "name1" => "B6",  "size1" => '4.921" x 6.929" (12.5cm x 17.6cm)', "name2" => "B7",  "size2" => '3.465" x 4.921" (8.8cm x 12.5cm)' },
    { "name1" => "B8",  "size1" => '2.441" x 3.465" (6.2cm x 8.8cm)', "name2" => "B9",  "size2" => '1.732" x 2.441" (4.4cm x 6.2cm)' },
    { "name1" => "B10", "size1" => '1.220" x 1.732" (3.1cm x 4.4cm)' },
    { "name1" => "C0",  "size1" => '36.102" x 51.063" (91.7cm x 129.7cm)', "name2" => "C1",  "size2" => '25.512" x 36.102" (64.8cm x 91.7cm)' },
    { "name1" => "C2",  "size1" => '18.032" x 25.512" (45.8cm x 64.8cm)', "name2" => "C3",  "size2" => '12.756" x 18.032" (32.4cm x 45.8cm)' },
    { "name1" => "C4",  "size1" => '9.016" x 12.756" (22.9cm x 32.4cm)', "name2" => "C5",  "size2" => '6.378" x 9.016" (16.2cm x 22.9cm)' },
    { "name1" => "C6",  "size1" => '4.488" x 6.378" (11.4cm x 16.2cm)', "name2" => "C7",  "size2" => '3.189" x 4.488" (8.1cm x 11.4cm)' },
    { "name1" => "C8",  "size1" => '2.244" x 3.189" (5.7cm x 8.1cm)', "name2" => "C9",  "size2" => '1.575" x 2.244" (4.0cm x 5.7cm)' },
    { "name1" => "C10", "size1" => '1.102" x 1.575" (2.8cm x 4.0cm)' },
    { "name1" => "EXECUTIVE", "size1" => '7.248" x 10.5" (18.410cm x 26.670cm)', "name2" => "FOLIO",     "size2" => '8.5" x 13" (21.590cm x 33.020cm)' },
    { "name1" => "LEGAL",     "size1" => '8.5" x 14" (21.590cm x 35.560cm)', "name2" => "LETTER",    "size2" => '8.5" x 11" (21.590cm x 27.940cm)' },
    { "name1" => "RA0", "size1" => '33.858" x 48.032" (86cm x 122cm)', "name2" => "RA1", "size2" => '24.016" x 33.858" (61cm x 86cm)' },
    { "name1" => "RA2", "size1" => '16.929" x 24.016" (43cm x 61cm)', "name2" => "RA3", "size2" => '12.008" x 16.929" (30.5cm x 43cm)' },
    { "name1" => "RA4", "size1" => '8.465" x 12.008" (21.5cm x 30.5cm)' },
    { "name1" => "SRA0", "size1" => '35.433" x 50.394" (90cm x 128cm)', "name2" => "SRA1", "size2" => '25.197" x 35.433" (64cm x 90cm)' },
    { "name1" => "SRA2", "size1" => '17.717" x 25.197" (45cm x 64cm)', "name2" => "SRA3", "size2" => '12.598" x 17.717" (32cm x 45cm)' },
    { "name1" => "SRA4", "size1" => '8.858" x 12.598" (22.5cm x 32cm)' }
  ]

  tab.data.replace data
  tab.render_on(pdf)
end
.endeval }}}

3<PDF::Writer.prepress>
This is an alternative way to create a new PDF document. In addition to the
named parameters in PDF::Writer.new (<b>:paper</b>, <b>:orientation</b>, and
<b>:version</b>), this method accepts the following options as well.

<b>:left_margin</b>, <b>:right_margin</b>, <b>:top_margin</b>, and
<b>:bottom_margin</b> specify the margins. Prepress marks are placed relative
to the margins, so when creating a document in prepress mode the margins must
be specified as the document is created. Future versions of PDF::Writer will
be more flexible on this. The default margins are 36pts (about ½”).

<b>:bleed_size</b> and <b>:mark_length</b> specify the size of the bleed area
(default is 12 points) and the length of the prepress marks (default is 18
points).

Prepress marks will appear on all pages.

.code {{{
require "pdf/writer"

PDF::Writer.prepress # US Letter, portrait, 1.3, prepress
.endcode }}}

2<Adding Text to PDF Document>AddingText
There are two different and complementary ways of adding text to a PDF
document with PDF::Writer. The easier is to use the generated PDF as a
“smart” canvas, where information internal to the document is used to
maintain a current writing pointer (see “<r:xref name="WritingPointer"
label="title" />”) and text is written within the writing space defined by
document margins (see “<r:xref name="DocumentMargins" label="title" />”). The
harder is to place the text on the document canvas explicitly, ensuring that
the text does not overflow the page horizontally or vertically.

In order to support more flexible and humane layout, PDF::Writer supports
XML-like tags that can be used to change the current text state (discussed
earlier in “<r:xref name="FontFamilies" label="title" />”), substitute
alternate text, or apply custom formatting. This is discussed in detail in
“<r:xref name="TextTags" label="title" />”.

3<Text Wrapping>TextWrapping
PDF::Writer uses a very simple text wrapping formula. If a line looks like
it’s going to exceed its permitted width, then PDF::Writer backtracks to the
previous hyphen (‘-’) or space to see if the text will then fit into the
permitted width. It will do this until the text will fit. If no acceptable
hyphen or space can be found, the text will be forcibly split at the largest
position that will fit in the permitted width.

3<“Smart” Text Writing>
As noted, PDF::Writer has two different ways to write text to a PDF document.
This is expressed with three different methods. The method that uses the PDF
document as a “smart” canvas is <b>PDF::Writer#text</b>.

4<PDF::Writer#text(text, options = nil)>
This method will add text to the document starting at the current drawing
position (described in “<r:xref name="WritingPointer" label="title" />”). It
uses the document margins (see “<r:xref name="DocumentMargins" label="title"
/>”) to flow text onto the page. If an explicit newline (‘\n’) is
encountered, it will be used.

The text is drawn from the left margin until it reaches the right margin. If
the text exceeds the right margin, then the text will be wrapped and
continued on the next line. If the text to be written will exceed the bottom
margin, then a new page will be created (detailed in “<r:xref
name="ColumnOutput" label="title" />”).

.code {{{
  pdf.text("Ruby is fun!")
.endcode }}}

5<Text Size>
If the <b>:font_size</b> is not specified, then the last font size
or the default font size of 12 points will be used. If this value is
provided, this <b>changes</b> the current font size just as if the #font_size
attribute were set.

.code {{{
  pdf.text("Ruby is fun!", :font_size =&gt; 16)
.endcode }}}

5<Text Justification>TextJustification
Text may be justified within the margins (normal or overridden) in one of
four ways with the option <b>:justification</b>.

.blist disc {{{
<b>:left</b> justification is also called “ragged-right”, because the text is flush against the left margin and the right edge of the text is uneven. This is the default justification.
<b>:right</b> justification places the text flush agains the right margin; the left edge of the text is uneven.
<b>:center</b> justification centers the text between the margins.
<b>:full</b> justification ensures that the text is flush against both the left and right margins. Additional space is inserted between words to make this happen. The last line of a paragraph is only made flush left.
.endblist }}}

.code {{{
  pdf.text("Ruby is fun!", :justification =&gt; :left)
  pdf.text("Ruby is fun!", :justification =&gt; :right)
  pdf.text("Ruby is fun!", :justification =&gt; :center)
  pdf.text("Ruby is fun!", :justification =&gt; :full)
.endcode }}}

5<Margin Override Options>
The drawing positions and margins may be modified with the <b>:left</b>,
<b>:right</b>, <b>:absolute_left</b>, and <b>:absolute_right</b> options. The
<b>:left</b> and <b>:right</b> values are PDF userspace unit offsets from the
left and right margins, respectively; the <b>:absolute_left</b> and
<b>:absolute_right</b> values are userspace units effectively providing
<i>new</i> margins for the text to be written. Note that
<b>:absolute_right</b> is offset from the left side of the page, not the
right side of the page.

.code {{{
    # The left margin is shifted inward 50 points.
  pdf.text("Ruby is fun!", :left =&gt; 50)
    # The right margin is shifted inward 50 points.
  pdf.text("Ruby is fun!", :right =&gt; 50)
    # The text is drawn starting at 50 points from the left of the page.
  pdf.text("Ruby is fun!", :absolute_left =&gt; 50)
    # The text is drawn ending at 300 points from the left of the page.
  pdf.text("Ruby is fun!", :absolute_right =&gt; 300)
.endcode }}}

5<Line Spacing Options>
Normally, each line of text is separated only by the descender space for the
font. This means that each line will be separated by only enough space to
keep it readable. This may be changed by changing the line spacing
(<b>:spacing</b>) as a multiple of the normal line height (authors’
manuscripts are often printed double spaced or <b>:spacing =&gt; 2</b>). This
can also be changed by redefining the total height of the line (the
<i>leading</i>) independent of the font size. With <b>:font_size =&gt;
12</b>, both <b>:spacing =&gt; 2</b> and <b>:leading =&gt; 24</b> do the same
thing. The <b>:leading</b> value overrides the <b>:spacing</b> value.

.code {{{
    # These are the same
  pdf.text("Ruby is fun!", :spacing =&gt; 2)
  pdf.text("Ruby is fun!", :leading =&gt; 24)
.endcode }}}

5<Test Writing>
Not generally used, the <b>:test</b> option can be set to <b>true</b>. This
will prevent the text from being written to the page and the method will
return <b>true</b> if the text will be overflowed to a new page or column.

3<Explicit Text Placement>
Text can also be placed starting with specific X and Y coordinates using
either <b>PDF::Writer#add_text_wrap</b> or <b>PDF::Writer#add_text</b>.

4<PDF::Writer#add_text_wrap(x, y, width, text, size = nil, justification = :left, angle = 0, test = false)>
This will add text to the page at position <b>(x, y)</b>, but ensures that it
fits within the provided <b>width</b>. If it does not fit, then as much as
possible will be written using the wrapping rules on page <r:xref
name="TextWrapping" label="page" />. The remaining text to be written will be
returned to the caller.

.code {{{
rest = pdf.add_text_wrap(150, pdf.y, 150, "Ruby is fun!", 72)
.endcode }}}

If <b>size</b> is not specified, the current #font_size will be used. This
parameter has changed position with <b>text</b> in PDF::Writer 1.1. If your
code appears to be using the old parameter call, it will be warned.

The optional justification parameter works just as described on page <r:xref
name="TextJustification" label="page" />, except that instead of margin
width, it is the boundaries defined by the <b>x</b> and <b>width</b>
parameters.

Text may be drawn at an angle. This will be discussed fully in the section on
<b>PDF::Writer#add_text</b>.

If the <b>test</b> parameter is <b>true</b>, the text will not be written,
but the remainder of the text will be returned.

4<PDF::Writer#add_text(x, y, text, size = nil, angle = 0, word_space_adjust = 0)>
This places the full text string at the <b>(x, y)</b> position and the
specified <b>angle</b> (measured in degrees of a circle, in a
counter-clockwise direction) with respect to the angle of the coordinate
axis.

<b>NOTE:</b> As of PDF::Writer 1.1.0, angle rotations are now
counter-clockwise, not clockwise as in earlier versions. This is a necessary
incompatible change to make transformations more compatible with other vector
graphics systems.

If <b>size</b> is not specified, the current #font_size will be used. This
parameter has changed position with <b>text</b> in PDF::Writer 1.1. If your
code appears to be using the old parameter call, it will be warned.

.code {{{
0.step(315, 45) do |angle|
  pdf.add_text(pdf.margin_x_middle, pdf.y, "#{angle}º".rjust(8), 12,
               angle)
end
.endcode }}}
.eval {{{
x = pdf.margin_x_middle
y = pdf.y - 50

0.step(315, 45) do |angle|
  pdf.add_text(x, y, "#{angle}º".rjust(8), 12, angle)
end

pdf.y = y - 30
.endeval }}}

The <b>adjust</b> parameter adjusts the space between words by the specified
number of PDF userspace units. This is primarily used by PDF::Writer#text and
PDF::Writer#add_text_wrap to support text justification.

3<Text Tags>TextTags
Text is written in PDF documents as simple strings positioned on the page
canvas. Unlike HTML, there is no easy way to indicate that the font or the
font size should be changed. This has to be managed entirely by the
generating application. PDF::Writer has a mechanism called “text tags” to
manage these changes and provide hooks to introduce new behaviours.

PDF::Writer provides five classes of text tags. Two are used to change font
families and have fixed meaning; the other three are callback tags.

Because text tags use XML semantics, XML escape characters are necessary to
display tag characters.
.blist {{{
&amp;lt; =&gt; &lt;
&amp;gt; =&gt; &gt;
&amp;amp; =&gt; &amp;
.endblist }}}

4<Font Family Tags (&lt;b&gt; and &lt;i&gt;)>
The font family tags have fixed meaning and will add the style identifiers
‘b’ and ‘i’ to the current text state, respectively. Both require closing
tags. The text between the tags will be rendered in <b>bold</b>,
<i>italic</i>, or <b><i>bold italic</i></b> versions of the master font as
appropriate. If the font does not exist, then the master font is used:
PDF::Writer will not simulate italics or boldface.

See <r:xref name="FontFamilies" label="title" /> for an explanation of font
families and text state.

4<Callback Tags and Parameters>
The remaining three tag classes are callback tags. All callback tags have a
regular form of <b>&lt;<i>x</i>:<i>name</i>[ <i>parameters</i>]&gt;</b>. This
is essentially the form of a namespaced XML tag. Parameters must be specified
as with XML, in the form of <b><i>name</i>="<i>value</i>"</b> or
<b><i>name</i>='<i>value</i>'</b>. Parameters will be provided to callbacks
as a name-value hash:
.code {{{
&lt;c:alink uri="http://rubyforge.org/"&gt;RubyForge&lt;/c:alink&gt;
.endcode }}}
When the opening <i>c:alink</i> tag is encountered, the callback will receive
a tag parameter hash of:
.code {{{
{ "uri" =&gt; "http://rubyforge.org/" }
.endcode }}}
When the closing <i>c:alink</i> tag is encountered, the callback will receive
an empty parameter hash (<b>{}</b>).

4<Tag and Callback Class Association>
The constant hash PDF::Writer::TAGS has three hashes that keep track of tag
name and callback class associations. A sample set of associations might be:
.code {{{
TAGS = {
  :pair     =&gt; { "alink"  =&gt; PDF::Writer::TagAlink, },
  :single   =&gt; { "disc"   =&gt; PDF::Writer::TagDisc, },
  :replace  =&gt; { "xref"   =&gt; PDF::TechBook::TagXref, },
}
.endcode }}}

Callback tags must define <b>#[](pdf, params)</b>. The standard callback
classes define <b>CallbackClass.[](pdf, params)</b> so that the callbacks are
used without instantiation.

4<Replacement Tags>
Replacement tags will replace the tag with a value computed by the callback.
It may perform additional processing, but no location information is provided
to the callback. The return from this callback <b>must</b> be the string that
will replace the tag. A replacement tag looks like this in text:
.code {{{
&lt;r:xref name="FontFamilies" label="title" /&gt;
.endcode }}}

Replacement tags always begin with the “namespace” of ‘r:’ and do not
surround text (they are stand-alone tags).

The <b>params</b> parameter sent to <b>#[]</b> is the default parameters
hash, containing only information from the tag itself.

5<Sample Replacement Tag>
The example below is the “xref” tag from PDF::TechBook.
.code {{{
class TagXref
  def self.[](pdf, params)
    name  = params["name"]
    item  = params["label"]

    xref = pdf.xref_table[name]
    if xref
      case item
      when 'page'
        label = xref[:page] || xref[:label]
      when 'label'
        label = xref[:label]
      end

      "&lt;c:ilink dest='#{xref[:xref]}'&gt;#{label}&lt;/c:ilink&gt;"
    else
      warn PDF::Writer::Lang[:techbook_unknown_xref] % [ name ]
      PDF::Writer::Lang[:techbook_unknown_xref] % [ name ]
    end
  end
end
PDF::Writer::TAGS[:replace]["xref"] = PDF::TechBook::TagXref
.endcode }}}

4<Single Drawing Tags>
Single drawing tags are stand-alone tags that will perform drawing behaviours
when the tag is encountered. Location information is provided to the callback
and the callback may return a hash with <b>(x, y)</b> information to adjust
the position of following text. A single drawing tag looks like this in text:
.code {{{
&lt;C:bullet /&gt;
.endcode }}}

Single drawing tags always begin with the “namespace” of ‘C:’ and do not
surround text. The <b>params</b> parameter sent to <b>#[]</b> is a complex
compound object. See <r:xref name="DrawingTagParameters" label="title" /> for
more information on the parameters provided to drawing tags.

A single drawing tag callback will be called once and only once when it is
encountered.

5<Sample Single Drawing Tag>
The example below is the “bullet” tag from PDF::Writer.
.code {{{
class TagBullet
  DEFAULT_COLOR = Color::RGB::Black

  class << self
    attr_accessor :color
    def [](pdf, info)
      @color ||= DEFAULT_COLOR

      desc  = info[:descender].abs
      xpos  = info[:x] - (desc * 2.00)
      ypos  = info[:y] + (desc * 1.05)

      pdf.save_state
      ss = StrokeStyle.new(desc)
      ss.cap  = :butt
      ss.join = :miter
      pdf.stroke_style ss
      pdf.stroke_color @style
      pdf.circle_at(xpos, ypos, 1).stroke
      pdf.restore_state
    end
  end
end
TAGS[:single]["bullet"] = TagBullet
.endcode }}}

4<Paired Drawing Tags>
Paired drawing tags are tags that surround text and perform drawing
behaviours realted to that text when the tag is encountered. Location
information is provided to the callback and the callback may return a hash
with <b>(x, y)</b> information to adjust the position of following text. A
paire ddrawing tag looks like this in text:
.code {{{
&lt;c:uline&gt;text&lt;/c:uline&gt;
.endcode }}}

Paired drawing tags always begin with the “namespace” of ‘c:’ and surround
text. The <b>params</b> parameter sent to <b>#[]</b> is a complex compound
object. See <r:xref name="DrawingTagParameters" label="title" /> for more
information on the parameters provided to drawing tags.

A paired drawing tag callback will be called at least twice, and may be
called many times while it is open. It will be called once when the tag is
opened, once for every line “end” that is reached, once for every line
“start” that is reached, and once when the tag is closed. For
PDF::Writer#text, line start and end positions are whenever a natural or
wrapped newline is encountered. For PDF::Writer#add_text_wrap and
PDF::Writer#add_text, the line start and end are used for <b>each call</b>
while a tag is open.

5<Sample Paired Drawing Tag>
The example below is the “uline” tag from PDF::Writer.
.code {{{
class TagUline
  DEFAULT_STYLE = {
    :factor  =&gt; 0.05
  }

  class << self
    attr_accessor :style

    def [](pdf, info)
      @style ||= DEFAULT_STYLE.dup

      case info[:status]
      when :start, :start_line
        @links ||= {}

        @links[info[:cbid]] = {
          :x         =&gt; info[:x],
          :y         =&gt; info[:y],
          :angle     =&gt; info[:angle],
          :descender =&gt; info[:descender],
          :height    =&gt; info[:height],
          :uri       =&gt; nil
        }

        pdf.save_state
        pdf.stroke_style! StrokeStyle.new(info[:height] * @style[:factor])
      when :end, :end_line
        start = @links[info[:cbid]]
        theta = PDF::Math.deg2rad(start[:angle] - 90.0)
        drop  = start[:height] * @style[:factor] * 1.5
        drop_x = Math.cos(theta) * drop
        drop_y = -Math.sin(theta) * drop
        pdf.move_to(start[:x] - drop_x, start[:y] - drop_y)
        pdf.line_to(info[:x] - drop_x, info[:y] - drop_y).stroke
        pdf.restore_state
      end
    end
  end
end
TAGS[:pair]["uline"] = TagUline
.endcode }}}

4<Drawing Tag Parameters>DrawingTagParameters
Drawing tags expect more information than is provided to replacement tags, as
they are expected to draw something or perform other tasks on the document at
the time that they are encountered in the appropriate position(s). The
parameters provided are described below.

5<:x>
The current <b>x</b> position of the text.

5<:y>
The current <b>y</b> position of the text.

5<:angle>
The current <b>angle</b> of the text in <i>degrees</i>. This will allow the
correct calculation of text or drawing orientation.

5<:params>
The hash of named parameters for the tag. This is the same as the value that
is provided to replacement tags.

5<:status>
One of the values <b>:start</b>, <b>:end</b>, <b>:start_line</b>, or
<b>:end_line</b>
.blist {{{
<b>:start</b> is provided to the callback when encountering the opening tag for a paired drawing tag, or when encountering a single drawing tag.
<b>:end</b> is provided to the callback when the closing tag for a paired drawing tag.
<b>:start_line</b> is provided to the callback when a new line is to be drawn and a paired tag is open.
<b>:end_line</b> is provided to the callback when a line is finished writing and a paired tag is open.
.endblist }}}

5<:cbid>
The callback identifier; this may be used as a key into a variable which
keeps state for the various callbacks. In the “uline” example, the
<b>@links</b> variable keeps this information.

5<:callback>
The name of the tag, used to find the callback object. This is only set for
opening paired tags or single tags.

5<:height>
The font height at the time the tag was encountered.

5<:descender>
The font descender at the time the tag was encountered.

4<Known Text Tags>
A few useful text tags and their callbacks have been defined in PDF::Writer.
Two others are used by the PDF::TechBook class used to generate this manual.

5<&lt;c:alink uri="URI"&gt;>
The &lt;c:alink uri="URI"&gt; tag is used to make a link to a resource
external to the document. This can be any URL handler registered by the
operating system for processing. The text “&lt;c:alink
uri="http://rubyforge.org/"&gt;RubyForge&lt;/c:alink&gt;” will generate a
hyperlink like “<c:alink uri="http://rubyforge.org/">RubyForge</c:alink>”.
This is known to work with HTTP, HTTPS, FTP, and MAILTO style URIs.

The callback object is <b>PDF::Writer::TagAlink</b>. The display style of
linking and link underlining may be modified through
<b>TagAlink.style</b>, which is a hash of five keys:
.blist {{{
<b>:color</b> is colour of the link underline. The default is Color::RGB::Blue. If nil, the current stroke colour will be maintained.
<b>:text_color</b> is the colour of the text. The default is Color::RGB::Blue. If nil, the current fill colour will be maintained.
<b>:factor</b> is the size of the line, as a multiplier of the text height. The default is 5% of the line height (0.05).
<b>:line_style</b> is a style modification hash provided to PDF::Writer::StrokeStyle.new. The default is a solid line with normal cap, join, and miter limit values. See “<r:xref name="StrokeStyleClass" label="title" />” for more information.
<b>:draw_line</b> indicates whether the underline should be drawn or not.
.endblist }}}

5<&lt;c:ilink dest="DEST"&gt;>
The &lt;c:ilink dest="DEST"&gt; tag is used to make a link to a destination
within the document. Destinations must be created manually.

.code {{{
  # Code writing ...
  # ... text, pages, etc.
pdf.add_destination("x3y3z3", "Fit")
  # More code writing ...
  # ... text, pages, etc.
pdf.text("&lt;c:ilink dest='x3y3z3'&gt;Internal Link&lt;/c:ilink&gt;")
.endcode }}}

This manual contains an extensive cross-reference table. <c:ilink
dest="xref5">This link</c:ilink> will goes to the fifth item in that
cross-reference table. (The code was &lt;c:ilink dest="xref5"&gt;This
link&lt;/c:ilink&gt;.)

This is implemented with <b>PDF::Writer::TagIlink</b>. There are no
configuration options.

5<&lt;c:uline&gt;>
The &lt;c:uline&gt; tag is used to underline text between the opening tag and
the closing tag. Therefore, “The quick brown fox &lt;c:uline&gt;is tired
of&lt;/c:uline&gt; jumping over the lazy dog” becomes “The quick brown fox
<c:uline>is tired of</c:uline> jumping over the lazy dog”.

This is implemented with <b>PDF::Writer::TagUline</b>. The display style of
linking and link underlining may be modified through
<b>TagUline.style</b>, which is a hash of three keys:
.blist {{{
<b>:color</b> is colour of the link underline. The default is nil, which means the current stroke colour will be used.
<b>:factor</b> is the size of the line, as a multiplier of the text height. The default is 5% of the line height (0.05).
<b>:line_style</b> is a style modification hash provided to PDF::Writer::StrokeStyle.new. The default is a solid line with normal cap, join, and miter limit values. See “<r:xref name="StrokeStyleClass" label="title" />” for more information.
.endblist }}}

5<&lt;C:bullet&gt;>
The &lt;C:bullet&gt; tag inserts a solid circular bullet, like this:
“ <C:bullet />”. Bullets are implemented in <b>PDF::Writer::TagBullet</b>. The
bullet display colour may be modified through <b>TagBullet.color</b>, which
defaults to <b>Color::RGB::Black</b>.

5<&lt;C:disc&gt;>
The &lt;C:disc&gt; tag inserts a hollow circular bullet, like this: “ <C:disc
/>”. Bullets are implemented in <b>PDF::Writer::TagDisc</b>. The disc display
colours may be modified throguh <b>TagDisc.foreground</b> and
<b>TagDisc.background</b>, which default to <b>Color::RGB::Black</b> and
<b>Color::RGB::White</b>, respectively.

5<&lt;C:tocdots ...&gt;>
This is a stand-alone callback that draws a dotted line over to the right and
appends a page number; it is implemented and used in
<b>PDF::TechBook::TagTocDots</b>. It is of limited configurability in this
version.

5<&lt;r:xref name="XREFNAME" label="page|label" text="TEXT"/&gt;>
This replacement callback returns an internal link (&lt;c:ilink...&gt;) with
either the name (label="title") or the page number (label="page") of the
cross-reference, or an indicator for arbitrary text (label="text") drawn from
the text attribute (e.g., text="random text here"). The page number will only
be used if it is known at the time that the &lt;r:xref&gt; tag is processed;
forward cross-references will always use the text (if present) or the label.
This is implemented through <b>PDF::TechBook::TagXref</b>.

3<Text Height and Width>
These methods return the height of the font or the width of the
text.

4<PDF::Writer#font_height(font_size = nil)>
Returns the height of the current font for the given size, measured in PDF
userspace units. This is the distance from the bottom of the descender to the
top of the capitals or ascenders. Uses the current #font_size if size is not
provided.

4<PDF::Writer#font_descender(font_size = nil)>
Returns the size of the descender—the distance below the baseline of the
font—which will normally be a negative number. Uses the current #font_size if
size is not provided.

4<PDF::Writer#text_width(text, font_size = nil)>
4<PDF::Writer#text_line_width(text, font_size = nil)>
Returns the width of the given text string at the given font size using the
current font, or the current default font size if none is given. The
difference between #text_width and #text_line_width is that the former will
give the width of the largest line in a multiline string.

2<Document Margins and Page Dimensions>DocumentMargins
A document canvas should not be written from edge to edge in most cases;
printers that try to print these items will often lose portions. From a
design perspective, margins increase the amount of whitespace on the page and
make the page easier to read. Drawing methods that do not use the internal
writing pointer (see below) will not use the margin. Typically, these are
methods that explicitly specify <i>(x, y)</i> coordinates. The default
margins are 36pts (about ½”).

3<Setting Margins>
PDF::Writer provides four methods to quickly define all margins at once. With
all four of these methods, if only one value is provided, all four margins
are that value. Two values define the top/bottom margins and the left/right
margins. Three values set separate top/bottom margins, but the left/right
margins will be the same. Four values defines each margin independently, as
shown below. The only difference between the methods is the measurements
used.

.code {{{
                                  # T  L  B  R
  pdf.margins_pt(36)              # 36 36 36 36
  pdf.margins_pt(36, 54)          # 36 54 36 54
  pdf.margins_pt(36, 54, 72)      # 36 54 72 54
  pdf.margins_pt(36, 54, 72, 90)  # 36 54 72 90
.endcode }}}

4<PDF::Writer#margins_pt(top, left = top, bottom = top, right = left)>
Set margins in points.

4<PDF::Writer#margins_cm(top, left = top, bottom = top, right = left)>
Set margins in centimetres.

4<PDF::Writer#margins_mm(top, left = top, bottom = top, right = left)>
Set margins in millimetres.

4<PDF::Writer#margins_in(top, left = top, bottom = top, right = left)>
Set margins in inches.

4<PDF::Writer#top_margin, PDF::Writer#top_margin=>
4<PDF::Writer#left_margin, PDF::Writer#left_margin=>
4<PDF::Writer#bottom_margin, PDF::Writer#bottom_margin=>
4<PDF::Writer#right_margin, PDF::Writer#right_margin=>
In addition, each margin may be examined and modified independently, but all
measurements for direct margin access are in PDF userspace units only.

.code {{{
  pdf.top_margin      # -> 36
  pdf.top_margin = 72
    # Also #left_margin, #bottom_margin, #right_margin
.endcode }}}

3<Using Margins and Page Dimensions>
Margin values in PDF::Writer are offset values. If the right margin is ½”,
the absolute position of the right margin will be the width of the page less
the offset of the right margin (conventionally, it will be the same as
“pdf.page_width - pdf.right_margin”. PDF::Writer provides attributes to read
these and other page dimension values. All measurements are in PDF userspace
units.

4<PDF::Writer#page_width>
The width of the page.

4<PDF::Writer#page_height>
The height of the page.

4<PDF::Writer#absolute_left_margin>
The absolute horizontal position of the left margin.

4<PDF::Writer#absolute_right_margin>
The absolute horizontal position of the right margin.

4<PDF::Writer#absolute_top_margin>
The absolute vertical position of the top margin.

4<PDF::Writer#absolute_bottom_margin>
The absolute vertical position of the bottom margin.

4<PDF::Writer#margin_height>
The height of the writing space.

4<PDF::Writer#margin_width>
The width of the writing space.

4<PDF::Writer#absolute_x_middle>
The horizontal middle of the page based on the page dimensions.

4<PDF::Writer#absolute_y_middle>
The vertical middle of the page based on the page dimensions.

4<PDF::Writer#margin_x_middle>
The horizontal middle of the page based on the margin dimensions.

4<PDF::Writer#margin_y_middle>
The vertical middle of the page based on the margin dimensions.

.code {{{
x = pdf.page_width - pdf.right_margin # flush right
y = pdf.page_height - pdf.top_margin  # flush top

  # Draw a box at the margin positions.
x = pdf.absolute_left_margin
w = pdf.absolute_right_margin - x
  # or pdf.margin_width
y = pdf.absolute_bottom_margin
h = pdf.absolute_top_margin - y
  # or pdf.margin_height
pdf.rectangle(x, y, w, h).fill
.endcode }}}

2<Internal Writing Pointer>WritingPointer
PDF::Writer keeps an internal writing pointer for use with several different
(mostly text) drawing methods. There are several ways to manipulate this
pointer directly. The writing pointer value is changed automatically with
drawing methods, column support, and starting new pages with #start_new_page.

3<PDF::Writer#y, PDF::Writer#y=>
The vertical position of the writing point. The vertical position is
constrained between the top and bottom margins. Any attempt to set it outside
of those margins will cause the writing point to be placed absolutely at the
margins.

.code {{{
pdf.y             # =&gt; 40
pdf.bottom_margin # =&gt; 36
pdf.y = 30        # =&gt; 36
.endcode }}}

3<PDF::Writer#pointer, PDF::Writer#pointer=>
The vertical position of the writing point. If the vertical position is
outside of the bottom margin, a new page will be created.

.code {{{
pdf.pageset.size  # =&gt; 1
pdf.pointer       # =&gt; 40
pdf.bottom_margin # =&gt; 36
pdf.top_margin    # =&gt; 36
pdf.page_height   # =&gt; 736
pdf.pointer = 30
pdf.pageset.size  # =&gt; 2
pdf.pointer       # =&gt; 700
.endcode }}}

3<PDF::Writer#move_pointer(dy, make_space = false)>
Used to change the vertical position of the writing point. The pointer is
moved <b>down</b> the page by <i>dy<i>. If the pointer is to be moved up, a
negative number must be used.

Moving up the page will not move to the previous page because of limitations
in the way that PDF::Writer works. The writing point will be limited to the
top margin position.

.code {{{
pdf.move_pointer(10)
.endcode }}}

If <i>make_space</i> is true and a new page is required to move the pointer,
then the pointer will be moved down on the new page. This will allow space to
be reserved for graphics. The following will guarantee that there are 100
units of space above the final writing point.

.code {{{
  move_pointer(100, true)
.endcode }}}

2<Pages and Multi-Column Output>ColumnOutput
While it can be said that the PDF document is a canvas, it is a canvas of
multiple pages. PDF::Writer does some intelligent management of the document
to create pages as text flows off of it, but sometimes it will be necessary
or useful to start pages manually. The class also provides facilities to
write multi-column output automatically.

If a margin-aware drawing method reaches the bottom margin, it will request a
new page with <b>#start_new_page</b>.

3<First Page>
When a PDF::Writer document is first created, an initial page is created and
inserted in the document. This page is always accessible with the #first_page
method.

4<PDF::Writer#first_page>
The first page created during startup, useful for adding something to it
later.

3<Starting New Pages>
4<PDF::Writer#start_new_page(force = false)>
Nominally starts a new page and moves the writing pointer is moved back to
the top margin. If multi-column output is on, a new column may be started
instead of a new page. If an actual new page is required, the new page is
added to the page list according to the page insert options (detailed below).
If <b>force</b> is true, then a new page will be created even if multi-column
output is on.

4<PDF::Writer#new_page(insert = false, page = nil, pos = :after)>
Add a new page to the document. This also makes the new page the current
active object. This allows for mandatory page creation regardless of
multi-column output. Note the writing pointer position is not reset with the
use of this method. For most purposes, #start_new_page is preferred.

3<Page Insertion Options>
These options control where in the page set new pages are inserted when using
PDF::Writer#start_new_page. When <b>#insert_mode</b> is on, new pages will be
inserted at the specified page. If a page is inserted <b>:before</b> the
page, then the <b>#insert_position</b> is changed to <b>:after</b> upon
successful insertion of that page. This will ensure that if with a page list
like “1 2 3 4 5 6”, inserting a page between pages 3 and 4 with:

.code {{{
pdf.insert_mode     :on
pdf.insert_page     4
pdf.insert_position :before
 ... # insert pages 7, 8, and 9
.endcode }}}

will result in a page list like “1 2 3 7 8 9 4 5 6” and not “1 2 3 9 8 7 4 5
6.”

When <b>#insert_mode</b> is off, pages are appended to the end of the
document.

4<PDF::Writer#insert_mode(options = {})>
Changes page insert mode. May be called as follows:

.code {{{
pdf.insert_mode         # =&gt; current insert mode
  # The following four affect the insert mode without changing the insert
  # page or insert position.
pdf.insert_mode(:on)    # enables insert mode
pdf.insert_mode(true)   # enables insert mode
pdf.insert_mode(:off)   # disables insert mode
pdf.insert_mode(false)  # disables insert mode

  # Changes the insert mode, the insert page, and the insert position at the
  # same time. This is the same as calling:
  #
  #   pdf.insert_mode(:on)
  #   pdf.insert_page(:last)
  #   pdf.insert_position(:before)
opts = {
  :on       =&gt; true,
  :page     =&gt; :last,
  :position =&gt; :before
}
pdf.insert_mode(opts)
.endcode }}}

4<PDF::Writer#insert_page(page = nil)>
Returns or changes the insert page property. The page must be either
<b>:last</b> or an integer value representing the position in the page
set—this value is completely unrelated to any page numbering scheme that may
be currently in progress.

.code {{{
pdf.insert_page         # =&gt; current insert page
pdf.insert_page(35)     # insert at page 35
pdf.insert_page(:last)  # insert at the last page
.endcode }}}

4<PDF::Writer#insert_position(position = nil)>
Returns or changes the insert position to be before or after the insert page.

.code {{{
pdf.insert_position           # =&gt; current insert position
pdf.insert_position(:before)  # insert before #insert_page
pdf.insert_position(:after)   # insert before #insert_page
.endcode }}}

3<Multi-Column Output>
4<PDF::Writer#start_columns(size = 2, gutter = 10)>
Starts multi-column output. Creates <i>size</i> number of columns with a
<i>gutter</i> of PDF unit space between each column.

If columns are already started, this will return <b>false</b> as only one
level of column definitions may be active at any time.

.code {{{
pdf.start_columns
pdf.start_columns(3)
pdf.start_columns(3, 2)
pdf.start_columns(2, 20)
.endcode }}}

When columns are on, <b>#start_new_page</b> will start a new column unless
the last column is the current writing space, where it will create a new
page.

4<PDF::Writer#stop_columns>
Turns off multi-column output. If we are in the first column, or the lowest
point at which columns were written is higher than the bottom of the page,
then the writing pointer will be placed at the lowest point. Otherwise, a new
page will be started.

4<PDF::Writer#column_width>
Returns the width of the currently active column or 0 if columns are off.

4<PDF::Writer#column_gutter>
Returns the size of the gutter between columns or 0 if columns are off.

4<PDF::Writer#column_number>
Returns the current column number or 0 if columns are off. Column numbers are
1-indexed.

4<PDF::Writer#column_count>
Returns the total number of columns or 0 if columns are off.

4<PDF::Writer#columns?>
Returns <b>true</b> if columns are turned on.

2<Page Numbering>PageNumbering
PDF::Writer supports automatic page numbering. The current page numbering
system does not support anything except standard Arabic numbering (e.g., 1,
2, 3). <b>The page numbering mechanism will be changing in a future version
of PDF::Writer to be far more flexible.</b>

4<PDF::Writer#start_page_numbering(x, y, size, pos = nil, pattern = nil, starting = nil)>
Prepares the placement of page numbers on pages from the current page. Place
them relative to the coordinates <i>(x, y)</i> with <i>pos</i> as the
relative position. <i>pos</i> may be <b>:left</b>, <b>:right</b>, or
<b>:center</b>. The page numbers will be written on each page using
<i>pattern</i>.

When <i>pattern</i> is rendered, <b>&lt;PAGENUM&gt;</b> will be replaced with
the current page number; <b>&lt;TOTALPAGENUM&gt;</b> will be replaced with
the total number of pages in the page numbering scheme. The default
<i>pattern</i> is “&lt;PAGENUM&gt; of &lt;TOTALPAGENUM&gt;”.

If <b>starting</b> is non-nil, this is the first page number. The number of
total pages will be adjusted to account for this.

Each time page numbers are started, a new page number scheme will be started.
The scheme number will be returned for use in other page number methods.

The following code produces a ten page document, numbered from the second
page (labelled ‘1 of 9’) until the eighth page (labelled ‘7 of 9’).

.code {{{
pdf = PDF::Writer.new
pdf.select_font "Helvetica"
                    # page  1, blank
pdf.start_new_page  # page  2, 1 of 9
pdf.start_page_numbering(300, 500, 20, nil, nil, 1)
pdf.start_new_page  # page  3, 2 of 9
pdf.start_new_page  # page  4, 3 of 9
pdf.start_new_page  # page  5, 4 of 9
pdf.start_new_page  # page  6, 5 of 9
pdf.start_new_page  # page  7, 6 of 9
pdf.start_new_page  # page  8, 7 of 9
pdf.stop_page_numbers
pdf.start_new_page  # page  9, blank
pdf.start_new_page  # page 10, blank
.endcode }}}

Multiple page numbering schemes can be used on the same page, as demonstrated
in <b>demo/pagenumber.rb</b> or below.

.code {{{
pdf = PDF::Writer.new
                    # Page  1: blank
sa = pdf.start_page_numbering(5, 60, 9, nil, nil, 1)
pdf.start_new_page  # Page  2: 1 of 2
pdf.start_new_page  # Page  3: 2 of 2
pdf.stop_page_numbering(true, :current, sa)
pdf.start_new_page  # Page  4: blank
sb = pdf.start_page_numbering(5, 50, 9, :center, nil, 10)
pdf.start_new_page  # Page  5: 10 of 12
pdf.start_new_page  # Page  6: 11 of 12
pdf.stop_page_numbering(true, :next, sb)
pdf.start_new_page  # Page  7: 12 of 12
sc = pdf.start_page_numbering(5, 40, 9, nil, nil, 1)
pdf.start_new_page  # Page  8: 1 of 3
pdf.start_new_page  # Page  9: 2 of 3
pdf.start_new_page  # Page 10: 3 of 3
pdf.stop_page_numbering(true, :current, sc)
pdf.start_new_page  # Page 11: blank
sd = pdf.start_page_numbering(5, 30, 9, nil, nil, 1)
pdf.start_new_page  # Page 12: 1 of 6
pdf.start_new_page  # Page 13: 2 of 6
se = pdf.start_page_numbering(5, 20, 9, nil, nil, 5)
sf = pdf.start_page_numbering(5, 10, 9, :right, nil, 1)
pdf.start_new_page  # Page 14: 3 of 6, 5 of 10, 1 of 8
pdf.start_new_page  # Page 15: 4 of 6, 6 of 10, 2 of 8
pdf.start_new_page  # Page 16: 5 of 6, 7 of 10, 3 of 8
pdf.stop_page_numbering(true, :next, sd)
pdf.start_new_page  # Page 17: 6 of 6, 8 of 10, 4 of 8
pdf.start_new_page  # Page 18: 9 of 10, 5 of 8
pdf.stop_page_numbering(true, :next, se)
pdf.stop_page_numbering(false, :current, sf)
pdf.start_new_page  # Page 19: 10 of 10
pdf.start_new_page  # Page 20: blank
.endcode }}}

.eval {{{
pdf.select_font("Helvetica")

data = [
  { "p1" =>  1, "c1" => "blank",  "p2" =>  6, "c2" => "11 of 12",
    "p3" => 11, "c3" => "blank",  "p4" => 16, "c4" => "5 of 6, 7 of 10, 3 of 8", },
  { "p1" =>  2, "c1" => "1 of 2", "p2" =>  7, "c2" => "12 of 12",
    "p3" => 12, "c3" => "1 of 6", "p4" => 17, "c4" => "6 of 6, 8 of 10, 4 of 8" },
  { "p1" =>  3, "c1" => "2 of 2", "p2" =>  8, "c2" => "1 of 3",
    "p3" => 13, "c3" => "2 of 6", "p4" => 18, "c4" => "9 of 10, 5 of 8", },
  { "p1" =>  4, "c1" => "blank",  "p2" =>  9, "c2" => "2 of 3",
    "p3" => 14, "c3" => "3 of 6, 5 of 10, 1 of 8", "p4" => 19, "c4" => "10 of 10", },
  { "p1" =>  5, "c1" => "10 of 12", "p2" => 10, "c2" => "3 of 3",
    "p3" => 15, "c3" => "4 of 6, 6 of 10, 2 of 8", "p4" => 20, "c4" => "blank", },
]

PDF::SimpleTable.new do |tab|
  tab.column_order.replace %w(p1 c1 p2 c2 p3 c3 p4 c4)

  tab.columns["p1"] = PDF::SimpleTable::Column.new("p1") { |col| col.heading = "Page" }
  tab.columns["c1"] = PDF::SimpleTable::Column.new("c1") { |col| col.heading = "Contents" }
  tab.columns["p2"] = PDF::SimpleTable::Column.new("p2") { |col| col.heading = "Page" }
  tab.columns["c2"] = PDF::SimpleTable::Column.new("c2") { |col| col.heading = "Contents" }
  tab.columns["p3"] = PDF::SimpleTable::Column.new("p3") { |col| col.heading = "Page" }
  tab.columns["c3"] = PDF::SimpleTable::Column.new("c3") { |col| col.heading = "Contents" }
  tab.columns["p4"] = PDF::SimpleTable::Column.new("p4") { |col| col.heading = "Page" }
  tab.columns["c4"] = PDF::SimpleTable::Column.new("c4") { |col| col.heading = "Contents" }

  tab.data.replace data

  tab.font_size = 8

  tab.render_on(pdf)
end
.endeval }}}

4<PDF::Writer#stop_page_numbering(stop_total = false, stop_at = :current, scheme = 0)>
Stops page numbering for the provided <b>scheme</b>. Returns <b>false</b> if
page numbering is off. If <b>stop_total</b> is true, then the totaling of
pages for this page numbering scheme will be stopped as well. If
<b>stop_at</b> is <b>:current</b>, then the page numbering will stop at the
current page; otherwise, it will stop at the next page.

4<PDF::Writer#which_page_number(page_num, scheme = 0)>
Given a particular generic page number <b>page_num</b> (numbered sequentially
from the beginning of the page set), return the page number under a
particular page numbering <b>scheme</b>. Returns <b>nil</b> if page numbering
is not turned on.

2<Repeating Elements>
It is common in documents to see items repeated from page to page. These may
be watermarks, headers, footers, or other elements that must appear on
multiple pages—aside from page numbering. PDF::Writer supports this through a
mechanism called “loose content objects.”

3<Loose Content Objects>
Up until this point, a page has been presented as the only canavas available
to write and draw upon. This is a useful fiction, but it <i>is</i> a fiction.
Any contents object may be drawn upon, and an implicit contents object is
created when a new page is created. The methods discussed below create a new
canvas for writing. It has the same physical boundaries as the page itself,
and should only be written to with explicit locations, as #text is not aware
of non-page canvases.

4<PDF::Writer#open_object>
Makes a loose content object. Output will go into this object until it is
closed. This object will not appear until it is included within a page. The
method will return the object reference. To aid in the conceptual grouping of
modifications to a loose content object, this method will yield the opened
object if a block is provided.

4<PDF::Writer#reopen_object(id)>
Makes the object for current content insertion the object specified by
<b>id</b>, which is a value returned by either #open_object or #new_page.

4<PDF::Writer#close_object>
Closes the currently open loose content object, preventing further writing
against the object.

3<Using Loose Content Objects>
Once a loose contents object has been created, it must be added to the
collection of loose objects before it can be seen in the PDF document.
PDF::Writer makes it easy to write these objects to the contents of pages
when the pages are created.

4<PDF::Writer#add_object(id, where = :this_page)>
After a loose content object has been created, it will only show if it has
been added to a page or page(s) with this method. Where the loose content
object will be added is controlled by the <b>where</b> option.

The object will not be added to itself.

.blist {{{
<b>:this_page</b> will add the object just to the current page.
<b>:all_pages</b> will add the object to the current and all following pages.
<b>:even_pages</b> will add the object to following even pages, including the current page if it is an even page.
<b>:odd_pages</b> will add the object to following odd pages, including the current page if it is an odd page.
<b>:all_following_pages</b> will add the object to the next page created and all following pages.
<b>:following_even_pages</b> will add to the next even page created and all following even pages.
<b>:following_odd_pages</b> will add to the next odd page created and all following odd pages.
.endblist }}}

4<PDF::Writer#stop_object(id)>
Stops adding the specified object to pages <i>after</i> this page.

3<Example>
The following example is used to create the heading of this manual.

.code {{{
pdf.open_object do |heading|
  pdf.save_state
  pdf.stroke_color! Color::Black
  pdf.stroke_style! PDF::Writer::StrokeStyle::DEFAULT

  s = 6
  t = "PDF::Writer for Ruby ~ Manual"
  w = pdf.text_width(t, s) / 2.0
  x = pdf.margin_x_middle
  y = pdf.absolute_top_margin
  pdf.add_text(x - w, y, t, s)

  x = pdf.absolute_left_margin
  w = pdf.absolute_right_margin
  y -= (pdf.font_height(s) * 1.01)
  pdf.line(x, y, w, y).stroke

  pdf.restore_state
  pdf.close_object
  pdf.add_object(heading, :next_all_pages)
end
.endcode }}}

2<Active Document Elements>ActiveDocumentElements
Hyperlinks—to locations within the document or on the Internet—can be created
in any document either using the callback forms (&lt;c:alink&gt; and
&lt;c:ilink&gt;) for words, or #add_link and #add_internal_link for any other
location in a document. Before internal links will work, the destination to
which each refers must be created with #add_destination. In fact, a PDF
document will not render if there is an internal link reference to a
nonexistant destination. For those familiar with HTML, this is similar to the
following:

.code {{{
&lt;h1 id="title"&gt;Title of Document&lt;/h1&gt;
...
&lt;a href="#title"&gt;Top&lt;/a&gt;
.endcode }}}

3<PDF::Writer#add_destination(label, style, *params)>
This method is used to create a labelled destination at this point in the
document. The label is a string containing the name which will be used for
links internal to the document (created by &lt;c:ilink&gt; or
#add_internal_link).

The destination style controls how many parameters are required.

4<XYZ, 3: left, top, zoom>
The viewport will be opened at position <b>(left, top)</b> with <b>zoom</b>
percentage. If the values are the string "null", the current parameter values
are unchanged.

4<Fit, 0>
Fit the page to the viewport (horizontal and vertical). There are no
parameters.

4<FitH, 1: top>
Fit the page horizontally to the viewport. The top of the viewport is set to
<b>top</b>.

4<FitV, 1: left>
Fit the page vertically to the viewport. The left of the viewport is set to
<b>left</b>.

4<FitR, 4: left, bottom, right, top>
Fits the page to the provided rectangle with corners at <b>(left, bottom)</b>
and <b>(right, top)</b>.

4<FitB, 0>
Fits the page to the bounding box of the page. There are no parameters.

4<FitBH, 1: top>
Fits the page horizontally to the bounding box of the page. The top of the
viewport is set to <b>top</b>.

4<FitBV, 1: left>
Fits the page vertically to the bounding box of the page. The left of the
viewport is set to <b>left</b>.

3<PDF::Writer#add_internal_link(label, x0, y0, x1, y1)>internal_links
Creates an internal link in the document, ‘label’ is the name of an target
point, and the other settings are the coordinates of the enclosing rectangle
of the clickable area from <b>(x0, y0)</b> to <b>(x1, y1)</b>.

Note that the destination markers are created using the #add_destination
method described below. This the manual way of doing what &lt;c:ilink&gt;
does.

3<PDF::Writer#add_link(uri, x0, y0, x1, y1)>external_links
Creates a clickable rectangular area from <b>(x0, y0)</b> to <b>(x1, y1)</b>
within the current page of the document, which takes the user to the
specified URI when clicked. This is the manual way of doing what
&lt;c:alink&gt; does.

.code {{{
pdf.add_link("http://rubyforge.org/projects/ruby-pdf/", pdf.left_margin, 100,
             pdf.left_margin + 400, 120)
pdf.text("http://rubyforge.org/projects/ruby-pdf/", :font_size => 16,
         :justification => :centre, :left => 50)
pdf.rectangle(50, 100, 400, 20).stroke
.endcode }}}
.eval {{{
pdf.move_pointer(30, true)
pdf.move_pointer(-5)
pdf.text("http://rubyforge.org/projects/ruby-pdf/", :font_size => 16,
         :justification => :centre, :left => 50)
pdf.move_pointer(5)
pdf.add_link("http://rubyforge.org/projects/ruby-pdf/", pdf.left_margin,
             pdf.y, pdf.left_margin + 400, pdf.y + 20)
pdf.rectangle(pdf.left_margin, pdf.y, 400, 20).stroke
.endeval }}}

2<Text Rendering Style>
PDF normally renders text by filling the path described by the font in use.
PDF::Writer provides (through PDF::Writer::Graphics) three methods to change
the way that text is rendered.

3<#text_render_style(style)>
3<#text_render_style!(style)>
3<#text_render_style?>
The first and second methods will set the text rendering style; the second
forces the style to be set in the PDF document even if it’s the same as the
currently known text rendering style. The third method returns the current
text rendering style.

The text rendering styles are integer values zero through seven. The meaning
of “fill” and “stroke” are explained in-depth <r:xref name="PaintingPaths"
label="text" text="later" />. PDF::Writer does not support the use of
clipping paths at this point.

.blist {{{
0: (default) Fill the text. Uses #fill_color for text rendering colour.
1: Stroke the text.
2: Fill, then stroke the text.
3: Neither fill nor stroke the text (e.g., it is invisible).
4: Fill the text and add it to the clipping path.
5: Stroke the text and add it to the clipping path.
6: Fill, then stroke the text and add it to the clipping path.
7: Add the text to the clipping path.
.endblist }}}

.eval {{{
pdf.move_pointer(160, true)
y = pdf.y + 90

pdf.select_font "Times-Roman"

pdf.fill_color    Color::RGB::Grey70
pdf.stroke_color! Color::RGB::Black
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.rectangle(pdf.absolute_left_margin, y + 64, pdf.margin_width, -100).fill_stroke

pdf.fill_color    Color::RGB::Black

(0..7).each do |style|
  pdf.add_text(pdf.absolute_left_margin + 20 + (60 * style), y + 48,
               style.to_s, 12)
end

(0..7).each do |style|
  pdf.text_render_style!(style)
  pdf.add_text(pdf.absolute_left_margin + 5 + (60 * style), y, "Q", 64)
end
.endeval }}}
.newpage
1<Graphics in PDF Documents>GraphicsOps

Graphics primitives are provided by the module PDF::Writer::Graphics,
included into PDF::Writer so that graphics primitives may be accessed from
PDF document objects transparently. All drawing operations return the canvas
so that drawing operations can be chained.

2<Drawing Paths>
PDF graphics drawing operations are different than most drawing libraries. In
most libraries, when a straight line is drawn between point A and point B,
the line is drawn in the current colour, line thickness, and style. PDF
drawing does not draw the line, but instead plots a path. Only when a path is
painted will it be rendered in the document as a visible drawing.

As an illustration, with the following code:
.code {{{
pdf.move_to(70, 70).line_to(100, 100).line_to(40, 40).line_to(70, 70)
.endcode }}}
Nothing will be drawn until it is painted (with a stroke operation in this
case):
.code {{{
pdf.stroke
.endcode }}}

The code for this can be simplified, too.
.code {{{
pdf.move_to(70, 70).line_to(100, 100).line_to(40, 40).close_stroke
.endcode }}}
.eval {{{
pdf.move_pointer(50, true)
pdf.move_to(pdf.margin_x_middle, pdf.y - 20)
pdf.line_to(pdf.margin_x_middle + 30, pdf.y + 10)
pdf.line_to(pdf.margin_x_middle - 30, pdf.y - 10)
pdf.close_stroke
pdf.move_pointer(20, true)
.endeval }}}

Drawing paths may be arbitrarily long and may be comprised of several
sub-paths. Paths are considered sub-paths when a new path is started. Only a
paint operation finishes all drawing paths. Much of the precise behaviour
will become clearer as the graphics operations are introduced over the next
sections.

2<Drawing Colours and Styles>DrawingStyles
PDF documents keep track of two separate drawing colours at any given time:
the stroke colour and the fill colour, used to render the respective paint
operations (see <r:xref name="PaintingPaths" label="title" />). Stroke
operations need not be rendered with solid lines; this is controlled by the
<r:xref name="StrokeStyle" label="title" />.

3<Drawing Colours>
Both stroke and fill colours should be set to instances of either Color::RGB
or Color::CMYK, defined and documented in the color-tools package. This
helper package is briefly documented in “<r:xref name="ColorTools"
label="title" />”.

4<#stroke_color(color)>
4<#fill_color(color)>
4<#stroke_color!(color = nil)>
4<#fill_color!(color = nil)>
4<#stroke_color?>
4<#fill_color?>
These six methods set and report the colour for stroke and fill operations,
respectively. The first two will only change the stroke/fill colour if it is
different than the current stroke/fill colour. The second will force the
current stroke/fill colour to be set or reset, unless a nil value is
provided. The third returns the value of the current stroke/fill colour.

3<Stroke Style>StrokeStyle
Stroke styles must be instances of the PDF::Writer::StrokeStyle class.

4<PDF::Writer::StrokeStyle>StrokeStyleClass
This class represents how lines will be drawn with stroke operations. Line
styles are defined by width, line cap style, join method, miter limit, and
dash patterns.

5<PDF::Writer::StrokeStyle.new(width = 1, options = {})>
Creates the new stroke style with the specified width and options. The
options correspond to the attributes described below. If a block is provided,
the stroke object is yielded to it.

5<PDF::Writer::StrokeStyle#width, PDF::Writer::StrokeStyle#width=>
The thickness of the line in PDF units.

.code {{{
s1 = PDF::Writer::StrokeStyle.new(1)
s2 = PDF::Writer::StrokeStyle.new(2)
s3 = PDF::Writer::StrokeStyle.new(3)
s4 = PDF::Writer::StrokeStyle.new(4)
s5 = PDF::Writer::StrokeStyle.new(5)
.endcode }}}
.eval {{{
s1 = PDF::Writer::StrokeStyle.new(1)
pdf.stroke_style s1
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
s2 = PDF::Writer::StrokeStyle.new(2)
pdf.stroke_style s2
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
s3 = PDF::Writer::StrokeStyle.new(3)
pdf.stroke_style s3
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
s4 = PDF::Writer::StrokeStyle.new(4)
pdf.stroke_style s4
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
s5 = PDF::Writer::StrokeStyle.new(5)
pdf.stroke_style s5
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
pdf.move_pointer(5, true)
.endeval }}}

5<PDF::Writer::StrokeStyle#cap, PDF::Writer::StrokeStyle#cap=>
The type of cap to put on the ends of the line.

.blist {{{
<b>:butt</b> caps squre off the stroke at the endpoint of the path. There is no projection beyond the end of the path.
<b>:round</b> caps draw a semicircular arc with a diameter equal to the line width is drawn around the endpoint and filled in.
<b>:square</b> caps continue the stroke beyond the endpoint of the path for a distance equal to half the line width and is squared off.
.endblist }}}

.code {{{ s5-butt
s5.cap = :butt
.endcode }}}
.eval {{{ s5-butt
s5 = PDF::Writer::StrokeStyle.new(5, :cap => :butt)
pdf.stroke_style s5
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
pdf.move_pointer(5, true)
.endeval }}}
.code {{{ s5-round
s5.cap = :round
.endcode }}}
.eval {{{ s5-round
s5 = PDF::Writer::StrokeStyle.new(5, :cap => :round)
pdf.stroke_style s5
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
pdf.move_pointer(5, true)
.endeval }}}
.code {{{ s5-square
s5.cap = :square
.endcode }}}
.eval {{{ s5-square
s5 = PDF::Writer::StrokeStyle.new(5, :cap => :square)
s5.cap = :square
pdf.stroke_style s5
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
pdf.move_pointer(5, true)
.endeval }}}
  
If this is unspecified, the cap style is unchanged.

5<PDF::Writer::StrokeStyle#join, PDF::Writer::StrokeStyle#join=>
How two lines join together.

.blist {{{
<b>:miter</b> joins indicate that the outer edges of the strokes for the two segments are extended until they meet at an angle, as in a picture frame. If the segments meet at too sharp an angle (as defined by the #miter_limit), a bevel join is used instead.
<b>:round</b> joins draw an arc of a circle with a diameter equal to the line width is drawn around the point where the two segments meet, connecting the outer edges of the strokes for the two segments. This pie-slice shaped figure is filled in, producing a rounded corner.
<b>:bevel</b> joins finish the two segments with butt caps and the the resulting notch beyond the ends of the segments is filled with a triangle, forming a flattened edge on the join.
.endblist }}}

.code {{{ s5-miter
s5.join = :miter
.endcode }}}
.eval {{{ s5-miter
s5 = PDF::Writer::StrokeStyle.new(5, :join => :miter)
pdf.stroke_style s5
pdf.move_pointer(40, true)
pdf.move_to(100, pdf.y).line_to(100, pdf.y + 30)
pdf.line_to(400, pdf.y + 30).stroke
pdf.move_pointer(5, true)
.endeval }}}
.code {{{ s5-round
s5.join = :round
.endcode }}}
.eval {{{ s5-round
s5 = PDF::Writer::StrokeStyle.new(5, :join => :round)
pdf.stroke_style s5
pdf.move_pointer(40, true)
pdf.move_to(100, pdf.y).line_to(100, pdf.y + 30)
pdf.line_to(400, pdf.y + 30).stroke
pdf.move_pointer(5, true)
.endeval }}}
.code {{{ s5-bevel
s5.join = :bevel
.endcode }}}
.eval {{{ s5-bevel
s5 = PDF::Writer::StrokeStyle.new(5, :join => :bevel)
pdf.stroke_style s5
pdf.move_pointer(40, true)
pdf.move_to(100, pdf.y).line_to(100, pdf.y + 30)
pdf.line_to(400, pdf.y + 30).stroke
pdf.move_pointer(5, true)
.endeval }}}

If this is unspecified, the join style is unchanged.

5<PDF::Writer::StrokeStyle#miter_limit, PDF::Writer::StrokeStyle#miter_limit=>
When two line segments meet and <tt>:miter</tt> joins have been specified,
the miter may extend far beyond the thickness of the line stroking the path.
The #miter_limit imposes a maximum ratio miter length to line width at which
point the join will be converted from a miter to a bevel. Adobe points out
that the ratio is directly related to the angle between the segments in user
space. With [p] representing the angle at which the segments meet:

.code {{{
     miter_length / line_width == 1 / (sin ([p] / 2))
.endcode }}}

A miter limit of 1.414 converts miters to bevels for [p] less than 90
degrees, a limit of 2.0 converts them for [p] less than 60 degrees, and a
limit of 10.0 converts them for [p] less than approximately 11.5 degrees. 

5<PDF::Writer::StrokeStyle#dash, PDF::Writer::StrokeStyle#dash=>
Controls the pattern of dashes and gaps used to stroke paths. This value must
either be nil, or a hash with <b>:pattern</b> and <b>:phase</b>

The <b>:pattern</b> is an array of numbers specifying the lengths (in PDF
userspace units) of alternating dashes and gaps. The array is processed
cyclically, so that <b>{ :pattern =&gt; [3] }</b> represents three units on,
three units off, and <b>{ :pattern =&gt; [2, 1] }</b> represents two units
on, one unit off. These are shown against a solid line for comparison.

.eval {{{ :pattern
sp1 = PDF::Writer::StrokeStyle.new(5, :dash => PDF::Writer::StrokeStyle::SOLID_LINE)
pdf.stroke_style sp1
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke

sp2 = PDF::Writer::StrokeStyle.new(5, :dash => { :pattern => [3] })
pdf.stroke_style sp2
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke

sp3 = PDF::Writer::StrokeStyle.new(5, :dash => { :pattern => [2, 1] })
pdf.stroke_style sp3
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
pdf.move_pointer(5, true)
.endeval }}}

The <b>:phase</b> is an integer offset to the <b>:pattern</b> where the
drawing of the stroke begins. This means that <b>{ :pattern</b> =&gt; [3],
:phase =&gt; 1 }</b> represents be two units on and followed by a repeating
pattern of three units off, three units on. A dash setting of <b>{ :pattern
=&gt; [2, 1], :phase =&gt; 2 }</b> repeats an off unit followed by two units
on. These are shown against a solid line for comparison.

.eval {{{ :phase
sp1 = PDF::Writer::StrokeStyle.new(5, :dash => PDF::Writer::StrokeStyle::SOLID_LINE)
pdf.stroke_style sp1
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke

sp2 = PDF::Writer::StrokeStyle.new(5, :dash => { :pattern => [3], :phase => 1 })
pdf.stroke_style sp2
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke

sp3 = PDF::Writer::StrokeStyle.new(5, :dash => { :pattern => [2, 1], :phase => 2 })
pdf.stroke_style sp3
pdf.move_pointer(8, true)
pdf.line(100, pdf.y, 400, pdf.y).stroke
pdf.move_pointer(5, true)
.endeval }}}

StrokeStyle#dash may be set to StrokeStyle::SOLID_LINE, returning drawing to
a solid line.

Dashed lines wrap around curves and corners just as solid stroked lines do,
with normal cap and join handling with no consideration of the dash pattern.
A path with several subpaths treats each sub-path independently; the complete
dash pattern is restarted at the beginning of each sub-path.

4<#stroke_style(style)>
4<#stroke_style!(style = nil)>
4<#stroke_style?(style)>
These methods set and report the style for stroke operations. The first two
will only change the stroke style if it is different than the current stroke
style. The second will force the current stroke style to be set or reset,
unless a nil value is provided. The third returns the value of the current
stroke style.

The default drawing style (a one unit solid line) can be set with the
constant StrokeStyle::DEFAULT.

2<Cubic Bézier Curves>
PDF drawings can only derive curved images (including ellipses and circles)
through a concept known as <c:alink
uri="http://astronomy.swin.edu.au/~pbourke/curves/bezier/">cubic Bézier
curves</c:alink>. These draw curved lines using four control points. (A
demonstration applet can be found at the Princeton <c:alink
uri="http://www.cs.princeton.edu/~min/cs426/jar/bezier.html">computer science
department</c:alink>).

The distinctive feature of these curves is that the curve is guaranteed to be
completely contained by the four sided polygon formed from the control
points. The example below demonstrates a Bézier curve with its control points
and three sides of the polygon marked. The curve is tangent to the line
between the control points at either end of the curve.

.code {{{
pdf.curve(200, pdf.y + 40, 250, pdf.y + 5, 350, pdf.y, 400, pdf.y + 45).stroke
.endcode }}}
.eval {{{
pdf.move_pointer(50, true)
pdf.stroke_style PDF::Writer::StrokeStyle.new(2)
pdf.curve(200, pdf.y + 40, 250, pdf.y + 5, 350, pdf.y, 400, pdf.y + 45).stroke
pdf.stroke_style PDF::Writer::StrokeStyle.new(0.5)
pdf.circle_at(200, pdf.y + 40, 5).stroke
pdf.circle_at(250, pdf.y + 5, 5).stroke
pdf.circle_at(350, pdf.y, 5).stroke
pdf.circle_at(400, pdf.y + 45, 5).stroke
stroke = PDF::Writer::StrokeStyle.new(0.5, :dash => { :pattern => [10, 15] })
pdf.stroke_style(stroke)
pdf.line(200, pdf.y + 40, 250, pdf.y + 5).stroke
pdf.line(250, pdf.y + 5, 350, pdf.y).stroke
pdf.line(350, pdf.y, 400, pdf.y + 45).stroke
.endeval }}}

2<Drawing Path Primitives>
These primitives control the drawing path that can later be painted. PDF
documents have a single drawing point for drawing operations. The operators
described here work with this drawing point and move the drawing point to a
new position.

4<#move_to(x, y)>
Moves the drawing point to <b>(x, y)</b>. This will start a new sub-path.

4<#line_to(x, y)>
Continues the current sub-path in a straight line from the drawing point to
the new <b>(x, y)</b> coordiante. The drawing point is left at <b>(x, y)</b>.

4<#curve_to(x0, y0, x1, y1, x2, y2)>
This draws a Bézier curve with the drawing point <b>(dpx, dpy)</b> and
<b>(x2, y2)</b> as the terminal points. <b>(x0, y0)</b> and <b>(x1, y1)</b>
are the control points on the curve. The drawing pointer is set to <b>(x2,
y2)</b> at the end of the curve.

4<#scurve_to(x0, y0, x1, y1)>
This draws a Bézier spline with the drawing point <b>(dpx, dpy)</b> and
<b>(x1, y1)</b> as the terminal points. <b>(dpx, dpy)</b> and <b>(x0, y0)</b>
are the control points on the curve. The drawing pointer is set to <b>(x1,
y1)</b> at the end of the curve.

4<#ecurve_to(x0, y0, x1, y1)>
This draws a Bézier spline with the drawing point <b>(dpx, dpy)</b> and
<b>(x1, y1)</b> as the terminal points. <b>(x0, y0)</b> and <b>(x1, y1)</b>
are the control points on the curve. The drawing pointer is set to <b>(x1,
y1)</b> at the end of the curve.

4<#rectangle(x, y, w, h = w)>
Draws a rectangle from <b>(x, y)</b> to <b>(x + w, y + h)</b>. The drawing
pointer is set to <b>(x + w, y + h)</b> when complete. This is a basic PDF
drawing primitive that continues the current drawing sub-path.

4<#close>
Closes the current sub-path by appending a straight line segment from the
drawing point to the starting point of the path. If the path is already
closed, this is does nothing. This operator terminates the current sub-path.

2<Shape Operators>
These operators build on the drawing path primitives to provide complete
shapes for drawing. These will always generate new sub-paths.

4<#line(x0, y0, x1, y1)>
Draws a straight line from <b>(x0, y0)</b> to <b>(x1, y1)</b> and leaves the
drawing pointer at <b>(x1, y1)</b>.

.code {{{
pdf.line(100, pdf.y - 5, 400, pdf.y + 5)
.endcode }}}
.eval {{{
pdf.move_pointer(20, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.line(100, pdf.y - 15, 400, pdf.y - 5).stroke
.endeval }}}

4<#curve(x0, y0, x1, y1, x2, y2, x3, y3)>
This draws a Bézier curve with <b>(x0, x1)</b> and <b>(x3, y3)</b> as the
terminal points. <b>(x1, y1)</b> and <b>(x2, y2)</b> are the control points
on the curve. The drawing pointer is set to <b>(x3, y3)</b> at the end of the
curve.

.code {{{
pdf.curve(100, pdf.y - 5, 200, pdf.y + 5, 300, pdf.y - 5, 400, pdf.y + 5
.endcode }}}
.eval {{{
pdf.move_pointer(20, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.curve(100, pdf.y - 15, 200, pdf.y - 5, 300, pdf.y - 15, 400, pdf.y - 5)
pdf.stroke
.endeval }}}

4<#scurve(x0, y0, x1, y1, x2, y2)>
This draws a Bézier spline with <b>(x0, y0)</b> and <b>(x2, y2)</b> as the
terminal points. <b>(x0, y0)</b> and <b>(x1, y1)</b> are the control points
on the curve. The drawing pointer is set to <b>(x2, y2)</b> at the end of the
curve.

.code {{{
pdf.scurve(100, pdf.y - 5, 300, pdf.y + 5, 400, pdf.y - 5)
.endcode }}}
.eval {{{
pdf.move_pointer(20, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.scurve(100, pdf.y - 15, 300, pdf.y - 5, 400, pdf.y - 15).stroke
.endeval }}}

4<#ecurve(x0, y0, x1, y1, x2, y2)>
This draws a Bézier spline with <b>(x0, y0)</b> and <b>(x2, y2)</b> as the
terminal points. <b>(x1, y1)</b> and <b>(x2, y2)</b> are the control points
on the curve. The drawing pointer is set to <b>(x2, y2)</b> at the end of the
curve.

.code {{{
pdf.ecurve(100, pdf.y - 5, 200, pdf.y + 5, 400, pdf.y - 5)
.endcode }}}
.eval {{{
pdf.move_pointer(20, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.ecurve(100, pdf.y - 15, 200, pdf.y - 5, 400, pdf.y - 15)
pdf.stroke
.endeval }}}

4<#circle_at(x, y, r)>
Draws a circle of radius <b>r</b> with centre point at <b>(x, y)</b>. The
drawing pointer is moved to <b>(x, y)</b> when finished.

.code {{{
pdf.circle_at(250, pdf.y, 10)
.endcode }}}
.eval {{{
pdf.move_pointer(30, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.circle_at(250, pdf.y + 15, 10)
pdf.stroke
.endeval }}}

4<#ellipse_at(x, y, r1, r2 = r1)>
Draws an ellipse of horizontal radius <b>r1</b> and vertical radius <b>r2</b>
with centre point at <b>(x, y)</b>. The drawing pointer is moved to <b>(x,
y)</b> when finished.

.code {{{
pdf.ellipse_at(250, pdf.y, 20, 10)
.endcode }}}
.eval {{{
pdf.move_pointer(30, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.ellipse_at(250, pdf.y + 15, 20, 10)
pdf.stroke
.endeval }}}

4<#ellipse2_at(x, y, r1, r2 = r1, start = 0, stop = 360, segments = 8)>
Draws an ellipse with centre point at <tt>(x, y)</tt>. The horizontal radius
is <b>r1</b> and the vertical radius is <b>r2</b>. A partial ellipse may be
drawn by specifying the starting and finishing angles in degrees. The ellipse
is drawn in segments composed of cubic Bézier curves. It is not recommended
that this be set to a value smalelr than 4, and any value less than 2 will be
treated as 2. The drawing pointer is moved to <b>(x, y)</b> when finished.
This ellipse is more “natural” at smaller sizes.

.code {{{
pdf.ellipse2_at(250, pdf.y, 20, 10)
.endcode }}}
.eval {{{
pdf.move_pointer(30, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.ellipse2_at(250, pdf.y + 15, 20, 10).stroke
.endeval }}}

.code {{{
pdf.ellipse2_at(250, pdf.y, 20, 10, 135, -45)
.endcode }}}
.eval {{{
pdf.move_pointer(30, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.ellipse2_at(250, pdf.y + 15, 20, 10, 135, -45).stroke
.endeval }}}

.code {{{
ss = PDF::Writer::StrokeStyle.new(2)
ss.cap = :round

pdf.stroke_style  ss
pdf.fill_color! Color::RGB::BlueViolet
pdf.translate_axis(300, pdf.y + 25)
pdf.circle_at(0, 0, 38).fill
pdf.rotate_axis(10)
pdf.stroke_color! Color::RGB::Grey30
pdf.ellipse2_at(0, 0, 81, 20, 65.5, -245.5).stroke
pdf.ellipse2_at(0, 0, 85, 22, 67.5, -247.5).stroke
pdf.ellipse2_at(0, 0, 89, 25, 70.5, -250.5).stroke
pdf.restore_state
.endcode }}}
.eval {{{
pdf.move_pointer(60, true)
ss = PDF::Writer::StrokeStyle.new(2)
ss.cap = :round

pdf.stroke_style  ss
pdf.fill_color! Color::RGB::BlueViolet
pdf.translate_axis(300, pdf.y + 25)
pdf.circle_at(0, 0, 38).fill
pdf.rotate_axis(10)
pdf.stroke_color! Color::RGB::Grey30
pdf.ellipse2_at(0, 0, 81, 20, 65.5, -245.5).stroke
pdf.ellipse2_at(0, 0, 85, 22, 67.5, -247.5).stroke
pdf.ellipse2_at(0, 0, 89, 25, 70.5, -250.5).stroke
.endeval }}}

4<#segment_at(x, y, r1, r2 = r1, start = 0, stop = 359.99, segments = 8)>
Draws an ellipse segment. This is a closed partial ellipse with lines from
the starting point to the centre point and the ending point and the centre
point. The drawing pointer is moved to <b>(x, y)</b> when finished.

.code {{{
pdf.segment_at(250, pdf.y, 20, 10, 45, -45)
.endcode }}}
.eval {{{
pdf.move_pointer(30, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.segment_at(250, pdf.y + 15, 20, 10, 45, -45)
pdf.stroke
.endeval }}}

4<#polygon(points)>
Draws a polygon using <i>points</i>. This is an array of
PDF::Writer::PolygonPoint objects or an array that can be converted to an
array of PolygonPoint objects. A PolygonPoint is a simple object that
contains <b>(x, y)</b> coordinates for the point and the way that the line
will be connected to the previous point. The connnector may be either
<b>:curve</b>, <b>:scurve</b>, <b>:ecurve</b>, or <b>:line</b>.

The connector is ignored on the first point provided to the polygon; that is
the starting point.

The drawing point is left at the last <b>(x, y)</b> point position.

5<:curve>
A <b>:curve</b> connector indicates that there are three points following
this point, to be provided as parameters to <b>#curve_to</b>.

.code {{{
points = [ [ 10, 10 ],          # starting point
           [ 20, 20, :curve ],  # first control point
           [ 30, 30 ],          # second control point
           [ 40, 40 ],          # ending point of the curve.
         ]
.endcode }}}

5<:scurve>
An <b>:scurve</b> connector indicates that there are two points following
this point, to be provided as parameters to <b>#scurve_to</b>.

.code {{{
points = [ [ 10, 10 ],          # starting point and first control point
           [ 20, 20, :scurve ], # second control point
           [ 30, 30 ],          # ending point of the curve.
         ]
.endcode }}}

5<:ecurve>
An <b>:ecurve</b> connector indicates that there are two points following
this point, to be provided as parameters to <b>#ecurve_to</b>.

.code {{{
points = [ [ 10, 10 ],          # starting point
           [ 20, 20, :ecurve ], # first control point
           [ 30, 30 ],          # ending point and second control point
         ]
.endcode }}}

5<:line>
A <b>:line</b> connector draws a line between the previous point and this
point.

.code {{{
points = [ [ 10, 10 ],          # starting point
           [ 20, 20, :line ],   # ending point
         ]
.endcode }}}

5<Examples>
All of these examples use :line connectors.

.code {{{
pdata = [[200, 10], [400, 20], [300, 50], [150, 40]]
pdf.polygon(pdata)
.endcode }}}
.eval {{{
pdata = [[200, 10], [400, 20], [300, 50], [150, 40]]
pdf.move_pointer(60, true)
pdata.collect! { |e| [e[0], e[1] + pdf.y] }
pdf.polygon(pdata).stroke
.endeval }}}

.code {{{
pdf.polygon(pdata).fill
.endcode }}}
.eval {{{
pdata = [[200, 10], [400, 20], [300, 50], [150, 40]]
pdf.move_pointer(60, true)
pdata.collect! { |e| [e[0], e[1] + pdf.y] }
pdf.polygon(pdata).fill
.endeval }}}

.code {{{
pdf.fill_color(Color::RGB.from_fraction(0.9, 0.8, 0.7))
pdf.polygon(pdata).fill
.endcode }}}
.eval {{{
pdata = [[200, 10], [400, 20], [300, 50], [150, 40]]
pdf.move_pointer(60, true)
pdata.collect! { |e| [e[0], e[1] + pdf.y] }
pdf.fill_color(Color::RGB.from_fraction(0.9, 0.8, 0.7))
pdf.polygon(pdata).fill
.endeval }}}

4<#rounded_rectangle(x, y, w, h, r)>
Draws a rectangle with rounded corners. The drawing point is left at the
<b>(x + w, y + h)</b>. The radius should be smaller than the width and height.

.code {{{
pdf.rounded_rectangle(100, pdf.y, 300, 30, 5)
.endcode }}}
.eval {{{
pdf.move_pointer(40, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.rounded_rectangle(100, pdf.y + 35, 100, 30, 10)
pdf.stroke
pdf.move_pointer(5, true)
.endeval }}}

4<#star(cx, cy, length, rays = 5)>
Draws a star centered at <b>(cx, cy)</b> with the specified number of
<b>rays</b>, where each ray is <b>length</b> units long. Stars with an odd
number of rays should have the top ray pointing toward the top of the
document. If the number of rays is less than four, it will be converted to
four.

.code {{{
pdf.star(166, pdf.y, 20, 4).stroke
pdf.star(232, pdf.y, 20, 5).stroke
pdf.star(298, pdf.y, 20, 6).stroke
pdf.star(364, pdf.y, 20, 7).stroke
pdf.star(430, pdf.y, 20, 8).stroke
.endcode }}}
.eval {{{
pdf.move_pointer(40, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
pdf.star(166, pdf.y + 15, 20, 4).stroke
pdf.star(232, pdf.y + 15, 20, 5).stroke
pdf.star(298, pdf.y + 15, 20, 6).stroke
pdf.star(364, pdf.y + 15, 20, 7).stroke
pdf.star(430, pdf.y + 15, 20, 8).stroke
pdf.move_pointer(5, true)
.endeval }}}

2<Painting Paths>PaintingPaths
Drawing paths become visible in PDF documents when they are painted. In
painting a path, the path may be <i>stroke</i>d or <i>fill</i>ed.

3<Stroking Operations>
Painting a path with a <i>stroke</i> draws a line along the current path. The
line follows each segment in the path, centered on the segment with sides
parallel to it. Each segment of the line uses the current line width and dash
pattern. Lines that meet use the current join style; ends of lines
unconnected to other lines use the current line cap style.

When two unconnected segments meet or intersect in the same space, they will
not be drawn as a closed segment; an explicit #close operator is recommended
in this case. If paths are manually closed, according to the Adobe PDF
Reference Manual version 1.6, the result may be a “messy corner, because line
caps are applied instead of a line join.”

There is the special case of single-point paths; when a path is either a
closed path that is exactly one point in size or of two or more points at the
same coordinates, the #stroke operator only paints the path if <b>:round</b>
line caps are specified, producing a filled circle centered at the single
point. This is considered a “degenerate” sub-path. An open degenerate
sub-path is not drawn.

3<Filling Operations>
<i>Fill</i> paint operations colour the entire region enclosed by the current
path with the current fill colour. If there are several disconnected
subpaths, open subpaths are implicitly closed and the overall enclosed area
is painted—the space is “joined” before painting.

Degenerate subpaths (the same as described above) will result in the painting
of the single device pixel lying under that point—this is device dependent
and not useful. An open degenerate sub-path is not drawn.

Filling simple paths is intuitively clear: parts of the document enclosed in
the path will be painted in the current fill colour. Complex paths—those that
intersect themselves or have one or more sub-paths enclosing other
sub-paths—require additional considerations for knowing what portions of the
canvas are inside a path. PDF offers two different rules: <i>nonzero winding
number</i> and the <i>even-odd</i>.

4<Nonzero Winding Number Fill Rule>
This is the standard fill behaviour. In this rule, a given point is inside a
path by conceptually drawing a line from the point to beyond the outer edge
of the path. Using an initial value of zero, at every place where the path
crosses the ray from left to right, one will be added to the value; where it
crosses from right to left, one will be subtracted from the value. If the
resulting value is zero, the point is outside of the path. In all other
cases, the path is inside. This can be expressed in Ruby as:
.code {{{
crossings.inject(0) { |ii, xd| xd.left_right? ? ii + 1 : ii - 1 } != 0
.endcode }}}

For the purposes of applying this rule, if a ray concides with or is tangent
to a path segment, a different ray is chosen. Consider the two six-pointed
stars below. The first is drawn so that all paths head in the same direction:
.code {{{
pdf.move_to(x0, y1).line_to(x2, y1).line_to(x1, y3).close
pdf.move_to(x1, y0).line_to(x2, y2).line_to(x0, y2).close
.endcode }}}

The second is drawn differently:
.code {{{
pdf.move_to(x0, y1).line_to(x2, y1).line_to(x1, y3).close
pdf.move_to(x0, y2).line_to(x2, y2).line_to(x1, y0).close
.endcode }}}

In the first, the figure is solid. In the second, because of the nonzero
winding number fill rule, it is hollow.

.eval {{{
pdf.move_pointer(60, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
y0 = pdf.y + 50 # top
y1 = y0 - 10    # top-middle
y2 = y1 - 20    # bottom-middle
y3 = y2 - 10    # bottom

x0 = 140
x1 = 160
x2 = 180

pdf.move_to(x0, y1).line_to(x2, y1).line_to(x1, y3).close
pdf.move_to(x1, y0).line_to(x2, y2).line_to(x0, y2).close
pdf.fill

x0 += 100
x1 += 100
x2 += 100

pdf.move_to(x0, y1).line_to(x2, y1).line_to(x1, y3).close
pdf.move_to(x0, y2).line_to(x2, y2).line_to(x1, y0).close
pdf.fill
.endeval }}}

4<Even-Odd Fill Rule>
The even-odd fill rule uses the same imaginary ray concept as the nonzero
winding number. Instead of depending on the direction of the intersections of
path with the ray, it depends on the total number of intersections. If that
total is odd, the point is inside the path for painting purposes. If the
total is even, the point is outside the path. Shown below are the same two
figures used in the nonzero winding number rule, but this time they are
filled using the even-odd rule. They appear the same.

.eval {{{
pdf.move_pointer(60, true)
pdf.stroke_style  PDF::Writer::StrokeStyle::DEFAULT
pdf.stroke_color  Color::RGB::Black
y0 = pdf.y + 50 # top
y1 = y0 - 10    # top-middle
y2 = y1 - 20    # bottom-middle
y3 = y2 - 10    # bottom

x0 = 140
x1 = 160
x2 = 180

pdf.move_to(x0, y1).line_to(x2, y1).line_to(x1, y3).close
pdf.move_to(x1, y0).line_to(x2, y2).line_to(x0, y2).close
pdf.fill(:even_odd)

x0 += 100
x1 += 100
x2 += 100

pdf.move_to(x0, y1).line_to(x2, y1).line_to(x1, y3).close
pdf.move_to(x0, y2).line_to(x2, y2).line_to(x1, y0).close
pdf.fill(:even_odd)
.endeval }}}

4<Fill Methods>

5<#stroke>
Terminate the path object and draw it with the current stroke style.

5<#close_stroke>
Close the current path by appending a straight line segment from the drawing
point to the starting point of the path and then stroke it. This is the same
as “pdf.close.stroke”.

5<#fill(rule = nil)>
Fills the path according to the rules noted above, using the nonzero winding
number rule <i>unless</i> the rule parameter has the value <b>:even_odd</b>.
Uses the current fill colour.

5<#close_fill(rule = nil)>
Close the current path by appending a straight line segment from the drawing
point to the starting point of the path and then fill it. This is the same
as “pdf.close.fill”.

5<#fill_stroke(rule = nil)>
Fills and strokes the path. This is the same as constructing two identical
path objects and calling #fill on the first and #stroke on the second. Paths
filled and stroked in this manner are treated as if they were one object for
PDF transparency purposes (the PDF transparency model is not yet supported by
PDF::Writer).

5<#close_fill_stroke(rule = nil)>
Closes (as per #close), fills (as per #fill), and strokes (as per #stroke)
the drawing path. As with #fill_stroke, it constructs a single object that
acts like two paths.

2<Images>PDFImages
PDF::Writer supports the insertion of images into documents. These images
must be either JPEG (either RGB or CMYK colour spaces) or PNG format; other
format images must be converted to JPEG or PNG before they can be used in a
document.

Images are inserted into PDF documents as one pixel per PDF userspace unit,
making a 320×240 image approximately 4½”×3¼” (113mm×85mm) in size.

3<#add_image(image, x, y, width = nil, height = nil, image_info = nil, link = nil)>
Adds an image to the PDF document. The image must be a String that represents
the binary contents of a PNG or JPEG file <i>or</i> it must be a PDF image
that has previously been inserted by one of the image insertion methods
(including this one).

The loaded image will be placed with its <i>lower left-hand corner</i> at
<b>(x, y)</b> on the current page. The size of the image will be determined
by the <b>width</b> and <b>height</b> parameters. If neither width nor height
are provided, the image will be inserted at its “natural” size.

If width is specified, but height is not specified, then the image will be
scaled according to the width of the image. If height is specified, but width
is not, the image will be scaled according to the height of the image. In
both cases, the aspect ratio—the ratio of width to height—will be maintained.
If both width and height are specified, the set values will be used.

If the width and/or height values are negative, the image will be flipped
horizontally or vertically, as appropriate.

The <b>image_info</b> parameter must either be unspecified, nil, or a valid
instance of PDF::Writer::Graphics::ImageInfo (see “<r:xref name="ImageInfo"
label="title" />”).

The <b>link</b> parameter allows the image to be made as a clickable link to
internal or external targets. It must be specified as a hash with two
options:

.blist {{{
<b>:type</b> is the type of the link, either <b>:internal</b> or <b>:external</b>. This affects the interpretation of <b>:target</b>.
<b>:target</b> is the destination of the link. For <b>:type =&gt; :internal</b> links, this is an internal destination. For <b>:type =&gt; :external</b> links, this is an URI. See <r:xref name="internal_links" label="text" text="#add_internal_link" /> and <r:xref name="external_links" label="text" text="#add_link" /> for more information.
.endblist }}}

The image object will be returned so that it can be used again in the future.

3<#add_image_from_file(image, x, y, width = nil, height = nil, link = nil)>
Adds an image to the PDF document from a file. The <b>image</b> parameter may
be an IO-like object (it must respond to #read and return the full image data
in a single read) or a filename. If <i>open-uri</i> is active, the filename
may also be an URI to a remote image.

In all other ways, it behaves just like #add_image.

3<#image(image, options = {})>
This third method for inserting images into a PDF document holds much the
same relationship to #add_image and #add_image_from_file as #text holds to
the text insertion methods #add_text and #add_text_wrap. This method will
insert an image in the document positioned vertically relative to the current
internal writing pointer. The horizontal position and image size are
controlled by the <i>options</i> passed to the method. If the image is taller
than the available space on the page and the image is not to be resized to
the available space (see below), a new page will be created and the image
will be inserted there.

4<:pad>
The image will be inserted into the document with this number of blank
userspace units on all sides. The default is 5 units.

4<:width>
The desired width of the image in userspace units. The image will be resized
to this width using the aspect ratio of the image. The image will not be
allowed to be larger than the amount of space available for the image as
controlled by the document margins and the padding. So, if a page is 400
units wide with 50 unit margins, that leaves an available size of 300 units.
With the default padding of 5 units, the inserted picture will be no larger
than 290 units wide, even if <b>:width</b> is set larger.

4<:resize>
How to automatically resize the image. This may be either <b>:width</b>,
<b>:full</b>, or a numeric value.

.blist {{{
<b>:resize =&gt; :width</b> will resize the image to be as wide as the writing area. The image height will be resized in accordance with the aspect ratio of the image. If the resized image is taller than will fit in the remaining space on the page, a new page will be created and the image will be inserted at the top of the next page.
<b>:resize =&gt; :full</b> will resize the image to be as wide as the writing area. The image height will be resized in accordinace with the aspect ratio of the image. If the resized image is taller than will fit in the remaining space on the page, the image will be resized again to be as tall as the remaining space. The width will be correspondingly shrunk.
<b>:resize =&gt; <i>number</i></b> will resize the image by a factor of the provided number. <b>:resize =&gt; 2</b> will double the size of the image and <b>:resize &gt; 0.5</b> will halve the size of the image. If the resized image is wider than the maximum width, it will be resized again to be no wider than that, with a proportional height.
.endblist }}}

This option is incompatible with the <b>:width</b> option. The behaviour when
both are specified is undefined.

4<:justification>
The horizontal placement of the image. It may be :center (the centre of the
image is placed in the centre of the margins and padding space), :right (the
right edge of the image is flush with the right margin and padding space), or
:left (the left edge of the image is flush with the left margin and padding
space). The default justification is :left.

4<:border>
The image may optionally be drawn with a border. If this parameter is
<b>true</b>, the default border will be drawn, which is a 50% grey solid line
border. The parameter may also be a hash with two values.

.blist {{{
<b>:color</b> specifies the colour of the border.
<b>:style</b> specifies the stroke style of the border.
.endblist }}}

4<:link>
The image may be made a clickable link to internal or external targets. This
option must be specified as a hash with two options:

.blist {{{
<b>:type</b> is the type of the link, either <b>:internal</b> or <b>:external</b>. This affects the interpretation of <b>:target</b>.
<b>:target</b> is the destination of the link. For <b>:type =&gt; :internal</b> links, this is an internal destination. For <b>:type =&gt; :external</b> links, this is an URI. See <r:xref name="internal_links" label="text" text="#add_internal_link" /> and <r:xref name="external_links" label="text" text="#add_link" /> for more information.
.endblist }}}

3<Limitations>
Because #text and #image comprise a <i>simple</i> layout engine, the
limitations of these methods must be understood. The image will be inserted
and text will be inserted after it. There is no option to place the image and
then wrap text around it as one might with OpenOffice or a proper destkop
publishing system. The one example of anything similar in this manual (page
<r:xref name="PDFCoordinateSpace" label="page" />) was done manually.

2<Coordinate Axis Transformations>CTMatrix
As described earlier in “<r:xref name="PDFCoordinateSpace" label="title" />,”
the standard PDF coordinate space has the base coordinate (0, 0) in the lower
left-hand corner of the canvas. PDF::Writer’s layout engine depends on this
behaviour, but the axis can be transformed in several ways. It can be
translated, rotated, scaled, and skewed.

Axis transformations are cumulative (matrix multiplicative; see the
description of #transform_matrix for a full explanation) within a graphics
state stack level (see the next section). It is recommended that axis
transformations are performed within a #save_state/#restore_state
combination.

3<translate_axis(x, y)>
Translates the coordinate space so that <b>(x, y)</b> under the old
coordinate space is now <b>(0, 0)</b>. The <i>direction</i> of the old
coordinate space is unchanged (y values are increasing). The example code
below will make the middle of the page the <b>(0, 0)</b> position.

.code {{{
pdf.translate_axis(pdf.margin_x_middle, pdf.margin_y_middle)
.endcode }}}

3<rotate_axis(angle)>
The entire coordinate space is rotated by the specified angle in a
counter-clockwise direction.

<b>NOTE:</b> As of PDF::Writer 1.1.0, angle rotations are now
counter-clockwise, not clockwise as in earlier versions. This is a necessary
incompatible change to make transformations more compatible with other vector
graphics systems.

3<scale_axis(x = 1, y = 1)>
Changes the size of PDF userspace units. This changes the scale of all
operations in PDF documents. With <b>#scale_axis(0.5, 1)</b>, the x-axis is
modified so that PDF userspace units are 1/144” in width, instead of the
normal 1/72” width. Note that PDF::Writer takes no notice of scale changes
for purposes of margins.

If the scale value is negative, the coordinate axis is reversed and graphics
or text will be drawn in reverse.

3<skew_axis(xangle = 0, yangle = 0)>
Rotates the x and y axes independently of one another, creating a skewed
appearance.

3<transform_matrix(a, b, c, d, e, f)>
Transforms the coordinate axis with the transformation vector interpreted as
a coordinate transformation matrix:

.eval {{{
  def bracket(pdf, x, y, width, height, size)
    y -= (height / 1.75)
    tw = pdf.text_width("[", size * 4)
    pdf.add_text(x - tw, y, "[", size * 4)
    tw = pdf.text_width("]", size * 3)
    pdf.add_text(x + width - tw, y, "]", size * 4)
    x + width + tw
  end

  rowh = pdf.font_height(14)

  pdf.move_pointer(rowh * 5, true)

  mw  = pdf.text_width("MMMM", 14)
  _x  = pdf.margin_x_middle - (mw / 2.0)
  _y  = pdf.y + (rowh * 3)

  pdf.add_text(_x, _y + rowh, "a  c  e", 14)
  pdf.add_text(_x, _y,        "b  d  f", 14)
  pdf.add_text(_x, _y - rowh, "0  0  1", 14)
  bracket(pdf, _x, _y, mw, rowh, 14)

  pdf.move_pointer(rowh)
.endeval }}}

The matrix transforms the new coordinates (x1, y1) to the old coordinates
(x0, y0) through the dot product:

.eval {{{
  def bracket(pdf, x, y, width, height, size)
    y -= (height / 1.75)
    tw = pdf.text_width("[", size * 4)
    pdf.add_text(x - tw, y, "[", size * 4)
    tw = pdf.text_width("]", size * 3)
    pdf.add_text(x + width - tw, y, "]", size * 4)
    x + width + tw
  end

  rowh = pdf.font_height(14)
  mtxw = pdf.text_width("MM", 14)

  pdf.move_pointer(rowh * 5, true)
  _y  = pdf.y + (rowh * 3)

  _x = 100
  pdf.add_text(_x, _y + rowh, "x0", 14)
  pdf.add_text(_x, _y, "y0", 14)
  pdf.add_text(_x, _y - rowh, " 1", 14)
  _x = bracket(pdf, _x, _y, mtxw, rowh, 14)

  pdf.add_text(_x, _y - (rowh / 1.5), '=', 42)

  mtxw  = pdf.text_width("MMMM", 14)
  _x += mtxw

  pdf.add_text(_x, _y + rowh, "a  c  e", 14)
  pdf.add_text(_x, _y,        "b  d  f", 14)
  pdf.add_text(_x, _y - rowh, "0  0  1", 14)
  _x = bracket(pdf, _x, _y, mtxw, rowh, 14)

  pdf.circle_at(_x - 3, _y + 3, 3).fill

  mtxw = pdf.text_width("MM", 14)
  _x += mtxw

  pdf.add_text(_x, _y + rowh, "x1", 14)
  pdf.add_text(_x, _y, "y1", 14)
  pdf.add_text(_x, _y - rowh, " 1", 14)
  bracket(pdf, _x, _y, mtxw, rowh, 14)

  pdf.move_pointer(rowh)
.endeval }}}

In practice, the six variables (a—f) be represented as a six-element vector:
[ a b c d e f ].

.blist {{{
Axis translation uses [ 1 0 0 1 x y ] where x and y are the new (0,0) coordinates in the old axis system.
Scaling uses [ sx 0 0 sy 0 0 ] where sx and sy are the scaling factors.
Rotation uses [ cos(a) sin(a) -sin(a) cos(a) 0 0 ] where a is the angle, measured in radians.
X axis skewing uses [ 1 0 tan(a) 1 0 0 ] where a is the angle, measured in radians.
Y axis skewing uses [ 1 tan(a) 0 1 0 0 ] where a is the angle, measured in radians.
.endblist }}}

2<Graphics State>
PDF allows for multiple independent levels of graphics state to be saved in a
last-in first-out stack. Graphics states may not remain open over page
boundaries, and must be balanced; the use of the two methods below will
ensure that this is managed automatically.

3<PDF::Writer#save_state>
This operation saves the current graphics state on a stack. This stack keeps
various parameters about the graphics state.

.blist {{{
The stroke style (line width, line cap style, line join style, miter limit, and dash pattern), page <r:xref name="StrokeStyleClass" label="page" />.
The coordinate axis transformation matrix, page <r:xref name="CTMatrix" label="page" />.
The colour rendering intent (not yet supported by PDF::Writer).
The flatness tolerance (not yet supported by PDF::Writer).
Extended graphics state dictionaries (not yet supported by PDF::Writer).
.endblist }}}

3<PDF::Writer#restore_state>
Restores the current graphics state.

2<PDF::Writer::Graphics::ImageInfo>ImageInfo
This class obtains metadata information about an image. It is a modified
version of ImageSize by <c:alink uri="mailto:keisuke@rccn.com">Keisuke
Minami</c:alink>, which can be found at <c:alink
uri="http://www.rubycgi.org/tools/index.en.htm">http://www.rubycgi.org/tools/index.en.htm</c:alink>.
It is available under the standard PDF::Writer licence but this class is also
available under the GNU General Public Licence, version 2 or later.

It supports GIF (GIF87a and GIF89a), PNG (Portable Network Graphics), JPEG,
BMP (Windows or OS/2 Bitmaps), PPM, PBM, PGM, TIFF, XBM (X Bitmap), XPM (X
Pixmap), PSD (PhotoShop), PCX, and SWF (Flash) image formats.

3<ImageInfo.new(data, format = nil)>
Creates the ImageInfo object from the provided image data. If <b>format</b>
is specified, the image will be treated as the desired format. Otherwise, it
will be automatically discovered (this is preferred).

3<ImageInfo#format>
Returns the format of the image.

3<ImageInfo#height>
3<ImageInfo#width>
Returns the height and width of the image.

3<ImageInfo#bits>
Returns the colour bit depth of the image, if supported.

3<ImageInfo#channels>
Returns the number of colour channels, if supported.

3<ImageInfo#info>
Returns other information the image may know about itself.
.newpage
1<PDF::Writer Document Operations>WriterDocOps

There are several operations for manipulating the generated document as a
whole or metadata about the document.

2<Whole Document Operations>
These operations act on, or return information about, the entire document.

3<PDF::Writer#size>
Returns the number of PDF objects in the document.

3<PDF::Writer#compressed, PDF::Writer#compressed=, PDF::Writer#compressed?>
Indicates or sets document compression. If the value is true, the document
will be compressed on writing using the “deflate” compression method. If the
‘zlib’ library cannot be loaded, compression settings will be ignored.

<b>NOTE:</b> This value should be set as early as possible during the
document generation, or only some sections of the PDF will be compressed.

3<PDF::Writer#media_box(x0, y0, x1, y1)>
3<PDF::Writer#trim_box(x0, y0, x1, y1)>
The #trim_box and the #media_box are both expressed in default (unscaled)
user space units. The #media_box defines the boundaries of the physical
medium on which the page is intended to be displayed or printed. As of this
version, PDF::writer supports only one media box for the e