/usr/share/doc/ghostscript-doc/Drivers.htm

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=us-ascii">
<title>The interface between Ghostscript and device drivers</title>
<!-- $Id$ -->
<!-- Originally: drivers.txt -->
<link rel="stylesheet" type="text/css" href="gs.css" title="Ghostscript Style">
</head>

<body>
<!-- [1.0 begin visible header] ============================================ -->

<!-- [1.1 begin headline] ================================================== -->

<h1>The interface between Ghostscript and device drivers</h1>

<!-- [1.1 end headline] ==================================================== -->

<!-- [1.2 begin table of contents] ========================================= -->

<h2>Table of contents</h2>

<blockquote><ul>
<li><a href="#Adding_drivers">Adding a driver</a>
<li><a href="#KISS">Keeping things simple</a>
<li><a href="#Structure">Driver structure</a>
<ul>
  <li><a href="#Structure_definition">Structure definition</a>
  <li><a href="#Sophisticated">For sophisticated developers only</a>
</ul>
<li><a href="#coordinates_and_types">Coordinates and types</a>
<ul>
  <li><a href="#Coordinate_system">Coordinate system</a>
  <li><a href="#Color_definition">Color definition</a>
<ul>
  <li><a href="#sep_and_linear_fields">Separable and linear fields</a>
  <li><a href="#Changing_color_info_data">Changing color_info data</a>           
</ul>
<li><a href="#Types">Types</a>
</ul>
<li><a href="#Coding_conventions">Coding conventions</a>
<ul>
  <li><a href="#Allocating_storage">Allocating storage</a>
  <li><a href="#Driver_instance_allocation">Driver instance allocation</a>
</ul>
<li><a href="#Printer_drivers">Printer drivers</a>
<li><a href="#Driver_procedures">Driver procedures</a>
<ul>
  <li><a href="#Life_cycle">Life cycle</a>
  <li><a href="#Open_close">Open, close, sync, copy</a>
  <li><a href="#Color_mapping">Color and alpha mapping</a>
  <li><a href="#Pixel_level_drawing">Pixel-level drawing</a>
<ul>
    <li><a href="#Bitmap_imaging">Bitmap imaging</a>
    <li><a href="#Pixmap_imaging">Pixmap imaging</a>
    <li><a href="#Compositing">Compositing</a>
  [<a href="#S_spec">S</a>, <a href="#T_spec">T</a>, <a href="#F_spec">f</a>,
   <a href="#Compositing_notes">Notes</a>]
</ul>
  <li><a href="#Polygon_level_drawing">Polygon-level drawing</a>
  <li><a href="#Linear_color_drawing">Linear color drawing</a>
  <li><a href="#High_level_drawing">High-level drawing</a>
<ul>
    <li><a href="#Paths">Paths</a>
    <li><a href="#Images">Images</a> [<a href="#Images_notes">Notes</a>]
    <li><a href="#Text">Text</a> [<a href="#Text_notes">Notes</a>]
    <li><a href="#Unicode">Unicode support for high level devices</a>
</ul>
  <li><a href="#Reading_bits_back">Reading bits back</a>
  <li><a href="#Parameters">Parameters</a>
<ul>
    <li><a href="#Default_CRD_parameters">Default color rendering dictionary (CRD) parameters</a>
</ul>
  <li><a href="#External_fonts">External fonts</a>
  <li><a href="#Page_devices">Page devices</a>
  <li><a href="#Miscellaneous">Miscellaneous</a>
</ul>
<li><a href="#Tray">Tray selection</a>
<ul>
  <li><a href="#LeadingEdge">Tray rotation and the LeadingEdge parameter</a>
  <li><a href="#LeadingPage">Interaction between LeadingEdge and PageSize</a>
</ul>
</ul></blockquote>

<!-- [1.2 end table of contents] =========================================== -->

<!-- [1.3 begin hint] ====================================================== -->

<p>For other information, see the <a href="Readme.htm">Ghostscript
overview</a> and the documentation on <a href="Make.htm">how to build
Ghostscript</a>.

<!-- [1.3 end hint] ======================================================== -->

<hr>

<!-- [1.0 end visible header] ============================================== -->

<!-- [2.0 begin contents] ================================================== -->

<h2><a name="Adding_drivers"></a>Adding a driver</h2>

<p>
To add a driver to Ghostscript, first pick a name for your device, say
"<code>smurf</code>".  (Device names must be 1 to 8 characters, begin
with a letter, and consist only of letters, digits, and underscores.  Case
is significant: all current device names are lower case.)  Then all you
need do is edit <code>contrib.mak</code> in two places.

<ol>
<li>The list of devices, in the section headed "Catalog".  Add
<code>smurf</code> to the list.

<li>The section headed "Device drivers".

<p>
Suppose the files containing the smurf driver are called
"<code>joe</code>" and "<code>fred</code>".  Then you should add the
following lines:

<blockquote>
<pre># ------ The SMURF device ------ #

smurf_=&#36;(GLOBJ)joe.&#36;(OBJ) &#36;(GLOBJ)fred.&#36;(OBJ)
&#36;(DD)smurf.dev: &#36;(smurf_)
        &#36;(SETDEV) &#36;(DD)smurf &#36;(smurf_)

&#36;(GLOBJ)joe.&#36;(OBJ) : &#36;(GLSRC)joe.c
	&#36;(GLCC) &#36;(GLO_)joe.&#36;(OBJ) &#36;(C_) &#36;(GLSRC)joe.c

&#36;(GLOBJ)fred.&#36;(OBJ) : &#36;(GLSRC)fred.c
	&#36;(GLCC) &#36;(GLO_)fred.&#36;(OBJ) &#36;(C_) &#36;(GLSRC)fred.c</pre>
</blockquote>

<p>
and whatever <code>joe.c</code> and <code>fred.c</code> depend on.
If the smurf driver also needs special libraries, for instance a library
named "<code>gorf</code>", then the entry should look like this:

<blockquote>
<pre>&#36;(DD)smurf.dev : &#36;(smurf_)
        &#36;(SETDEV) &#36;(DD)smurf &#36;(smurf_)
        &#36;(ADDMOD) &#36;(DD)smurf -lib gorf</pre>
</blockquote>

<p>
If, as will usually be the case, your driver is a printer driver (as
<a href="#Printer_drivers">discussed below</a>), the device entry should
look like this:

<blockquote>
<pre>&#36;(DD)smurf.dev : &#36;(smurf_) &#36;(GLD)page.dev
        &#36;(SETPDEV) &#36;(DD)smurf &#36;(smurf_)</pre>
</blockquote>

<p>
or

<blockquote>
<pre>&#36;(DD)smurf.dev : &#36;(smurf_) &#36;(GLD)page.dev
        &#36;(SETPDEV) &#36;(DD)smurf &#36;(smurf_)
        &#36;(ADDMOD) &#36;(DD)smurf -lib gorf</pre>
</blockquote>

<p>
Note that the space before the :, and the explicit compilation rules for the
.c files, are required for portability,
</ol>

<hr>

<h2><a name="KISS"></a>Keeping things simple</h2>

<p>
If you want to add a simple device (specifically, a monochrome printer), you
probably don't need to read the rest of this document; just use the code in
an existing driver as a guide.  The Epson and Canon BubbleJet drivers <a
href="../base/gdevepsn.c">gdevepsn.c</a> and <a
href="../base/gdevbj10.c">gdevbj10.c</a> are good models for dot-matrix
printers, which require presenting the data for many scan lines at once; the
DeskJet/LaserJet drivers in <a href="../base/gdevdjet.c">gdevdjet.c</a> are
good models for laser printers, which take a single scan line at a time but
support data compression.  For color printers, there are unfortunately no
good models: the two major color inkjet printer drivers, <a
href="../base/gdevcdj.c">gdevcdj.c</a> and <a
href="../base/gdevstc.c">gdevstc.c</a>, are far too complex to read.

<p>
On the other hand, if you're writing a driver for some more esoteric
device, you probably do need at least some of the information in the rest
of this document.  It might be a good idea for you to read it in
conjunction with one of the existing drivers.

<p>
Duplication of code, and sheer volume of code, is a serious maintenance and
distribution problem for Ghostscript.  If your device is similar to an
existing one, try to implement your driver by adding some parameterization
to an existing driver rather than by copying code to create an entirely new
source module.  <a href="../base/gdevepsn.c">gdevepsn.c</a> and <a
href="../base/gdevdjet.c">gdevdjet.c</a> are good examples of this approach.

<hr>

<h2><a name="Structure"></a>Driver structure</h2>

<p>
A device is represented by a structure divided into three parts:

<ul>
<li>procedures that are (normally) shared by all instances of each device;

<li>parameters that are present in all devices but may be different for
each device or instance; and

<li>device-specific parameters that may be different for each instance.
</ul>

<p>
Normally the procedure structure is defined and initialized at compile
time.  A prototype of the parameter structure (including both generic and
device-specific parameters) is defined and initialized at compile time, but
is copied and filled in when an instance of the device is created.  Both of
these structures should be declared as <code>const</code>, but for backward
compatibility reasons the latter is not.

<p>
The <code>gx_device_common</code> macro defines the common structure
elements, with the intent that devices define and export a structure along
the following lines.  Do not fill in the individual generic parameter values
in the usual way for C structures: use the macros defined for this purpose
in <a href="../base/gxdevice.h">gxdevice.h</a> or, if applicable, <a
href="../base/gdevprn.h">gdevprn.h</a>.

<blockquote>
<pre>typedef struct smurf_device_s {
        gx_device_common;
        <b><em>... device-specific parameters ...</em></b>
} smurf_device;
smurf_device gs_smurf_device = {
        <b><em>... macro for generic parameter values ...,</em></b>
        { <b><em>... procedures ...</em></b> },         /* std_procs */
        <b><em>... device-specific parameter values if any ...</em></b>
};</pre>
</blockquote>
<p>
The device structure instance <b>must</b> have the name
<code>gs_smurf_device</code>, where <code>smurf</code> is the device
name used in <code>contrib.mak</code>.  <code>gx_device_common</code>
is a macro consisting only of the element definitions. 
<p>
All the device procedures are called with the device as the first argument.
Since each device type is actually a different structure type, the device
procedures must be declared as taking a <code>gx_device&nbsp;*</code> as
their first argument, and must cast it to
<code>smurf_device&nbsp;*</code> internally.  For example, in the code
for the "memory" device, the first argument to all routines is called
<code>dev</code>, but the routines actually use <code>mdev</code> to
refer to elements of the full structure, using the following standard
initialization statement at the beginning of each procedure:

<blockquote>
<pre>gx_memory_device *const mdev = (gx_device_memory *)dev;</pre>
</blockquote>

<p>
(This is a cheap version of "object-oriented" programming: in C++, for
example, the cast would be unnecessary, and in fact the procedure table
would be constructed by the compiler.)

<h3><a name="Structure_definition"></a>Structure definition</h3>

<p>
You should consult the definition of struct <code>gx_device_s</code> in
<a href="../base/gxdevice.h">gxdevice.h</a> for the complete details of the
generic device structure.  Some of the most important members of this
structure for ordinary drivers are:

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=top>	<td><code>const char *dname;</code>
	<td>&nbsp;&nbsp;&nbsp;&nbsp;
	<td>The device name
<tr valign=top>	<td><code>bool is_open;</code>
	<td>&nbsp;
	<td>True if device has been opened
<tr valign=top>	<td><code>gx_device_color_info color_info;</code>
	<td>&nbsp;
	<td>Color information
<tr valign=top>	<td><code>int width;</code>
	<td>&nbsp;
	<td>Width in pixels
<tr valign=top>	<td><code>int height;</code>
	<td>&nbsp;
	<td>Height in pixels
</table></blockquote>

<p>
The name in the structure (<code>dname</code>) should be the same as the
name in <a href="../base/contrib.mak">contrib.mak</a>.

<h3><a name="Sophisticated"></a>For sophisticated developers only</h3>

<p>
If for any reason you need to change the definition of the basic device
structure, or to add procedures, you must change the following places:

<blockquote><ul>
<li>This document and the <a href="News.htm">news document</a> (if you want
	to keep the documentation up to date).
<li>The definition of <code>gx_device_common</code> and the procedures
	in <a href="../base/gxdevcli.h">gxdevcli.h</a>.
<li>Possibly, the default forwarding procedures declared in
	<a href="../base/gxdevice.h">gxdevice.h</a> and implemented in
	<a href="../base/gdevnfwd.c">gdevnfwd.c</a>.
<li>The device procedure record completion routines in
	<a href="../base/gdevdflt.c">gdevdflt.c</a>.
<li>Possibly, the default device implementation in
	<a href="../base/gdevdflt.c">gdevdflt.c</a>,
	<a href="../base/gdevddrw.c">gdevddrw.c</a>, and
	<a href="../base/gxcmap.c">gxcmap.c</a>.
<li>The bounding box device in <a href="../base/gdevbbox.c">gdevbbox.c</a>
	(probably just adding <code>NULL</code> procedure entries if the
	new procedures don't produce output).
<li>These devices that must have complete (non-defaulted) procedure vectors:
<ul>
<li>The null device in <a href="../base/gdevnfwd.c">gdevnfwd.c</a>.
<li>The command list "device" in <a href="../base/gxclist.c">gxclist.c</a>.
	This is not an actual device; it only defines procedures.
<li>The "memory" devices in <a href="../base/gdevmem.h">gdevmem.h</a> and
	<code>gdevm*.c</code>.
</ul>
<li>The clip list accumulation "device" in
	<a href="../base/gxacpath.c">gxacpath.c</a>.
<li>The clipping "devices" <a href="../base/gxclip.c">gxclip.c</a>,
	<a href="../base/gxclip2.c">gxclip2.c</a>,
	and <a href="../base/gxclipm.c">gxclipm.c</a>.
<li>The pattern accumulation "device" in
	<a href="../base/gxpcmap.c">gxpcmap.c</a>.
<li>The hit detection "device" in <a href="../base/gdevhit.c">gdevhit.c</a>.
<li>The generic printer device macros in
	<a href="../base/gdevprn.h">gdevprn.h</a>.
<li>The generic printer device code in
	<a href="../base/gdevprn.c">gdevprn.c</a>.
<li>The RasterOp source device in
	<a href="../base/gdevrops.c">gdevrops.c</a>.
</ul></blockquote>

<p>
You may also have to change the code for
<code>gx_default_get_params</code> or
<code>gx_default_put_params</code> in <a
href="../base/gsdparam.c">gsdparam.c</a>.

<p>
You should not have to change any of the real devices in the standard
Ghostscript distribution (listed in <a href="../base/devs.mak">devs.mak</a>
and <a href="../base/contrib.mak">contrib.mak</a>) or any of your own
devices, because all of them are supposed to use the macros in <a
href="../base/gxdevice.h">gxdevice.h</a> or <a
href="../base/gdevprn.h">gdevprn.h</a> to define and initialize their state.

<hr>

<h2><a name="coordinates_and_types"></a>Coordinates and types</h2>

<h3><a name="Coordinate_system"></a>Coordinate system</h3>

<p>
Since each driver specifies the initial transformation from user
coordinates to device coordinates, the driver can use any coordinate system
it wants, as long as a device coordinate will fit in an
<code>int</code>.  (This is only an issue on DOS systems, where ints are
only 16 bits.  User coordinates are represented as floats.)  Most current
drivers use a coordinate system with (0,0) in the upper left corner, with
<b><em>X</em></b> increasing to the right and <b><em>Y</em></b> increasing
toward the bottom.  However, there is supposed to be nothing in the rest of
Ghostscript that assumes this, and indeed some drivers use a coordinate
system with (0,0) in the lower left corner.

<p>
Drivers must check (and, if necessary, clip) the coordinate parameters given
to them: they should not assume the coordinates will be in bounds.  The
<code>fit_fill</code> and <code>fit_copy</code> macros in <a
href="../base/gxdevice.h">gxdevice.h</a> are very helpful in doing this.

<h3><a name="Color_definition"></a>Color definition</h3>

<p>
Between the Ghostscript graphics library and the device, colors are
represented in three forms. Color components in a color space (Gray, RGB,
DeviceN, etc.) represented as <code>frac</code> values. Device colorants
are represented as <code>gx_color_value</code> values.  For many
procedures, colors are represented in a type called
<code>gx_color_index</code>.
All three types are described in more detail in <a href="#Types">Types</a>

<p>
The <code>color_info</code> member of the device structure defines the
color and gray-scale capabilities of the device.  Its type is defined as
follows:

<blockquote>
<pre>
/*
 * The enlarged color model information structure: Some of the
 * information that was implicit in the component number in
 * the earlier conventions (component names, polarity, mapping
 * functions) are now explicitly provided.
 *
 * Also included is some information regarding the encoding of
 * color information into gx_color_index. Some of this information
 * was previously gathered indirectly from the mapping
 * functions in the existing code, specifically to speed up the
 * halftoned color rendering operator (see
 * gx_dc_ht_colored_fill_rectangle in gxcht.c). The information
 * is now provided explicitly because such optimizations are
 * more critical when the number of color components is large.
 *
 * Note: no pointers have been added to this structure, so there
 *       is no requirement for a structure descriptor.
 */
typedef struct gx_device_color_info_s {

    /*
     * max_components is the maximum number of components for all
     * color models supported by this device. This does not include
     * any alpha components.
     */
    int max_components;

    /*
     * The number of color components. This does not include any
     * alpha-channel information, which may be integrated into
     * the gx_color_index but is otherwise passed as a separate
     * component.
     */
    int num_components;

    /*
     * Polarity of the components of the color space, either
     * additive or subtractive. This is used to interpret transfer
     * functions and halftone threshold arrays. Possible values
     * are GX_CM_POLARITY_ADDITIVE or GX_CM_POLARITY_SUBTRACTIVE
     */
    gx_color_polarity_t polarity;

    /*
     * The number of bits of gx_color_index actually used. 
     * This must be &lt;= sizeof(gx_color_index), which is usually 64.
     */
    byte depth;

    /*
     * Index of the gray color component, if any. The max_gray and
     * dither_gray values apply to this component only; all other
     * components use the max_color and dither_color values.
     *
     * This will be GX_CINFO_COMP_NO_INDEX if there is no gray 
     * component.
     */
    byte gray_index;

    /*
     * max_gray and max_color are the number of distinct native
     * intensity levels, less 1, for the gray and all other color
     * components, respectively. For nearly all current devices
     * that support both gray and non-gray components, the two
     * parameters have the same value.
     *
     * dither_grays and dither_colors are the number of intensity
     * levels between which halftoning can occur, for the gray and
     * all other color components, respectively. This is
     * essentially redundant information: in all reasonable cases,
     * dither_grays = max_gray + 1 and dither_colors = max_color + 1.
     * These parameters are, however, extensively used in the
     * current code, and thus have been retained.
     *
     * Note that the non-gray values may now be relevant even if
     * num_components == 1. This simplifies the handling of devices
     * with configurable color models which may be set for a single
     * non-gray color model.
     */
    gx_color_value max_gray;	/* # of distinct color levels -1 */
    gx_color_value max_color;

    gx_color_value dither_grays;
    gx_color_value dither_colors;

    /*
     * Information to control super-sampling of objects to support
     * anti-aliasing.
     */
    gx_device_anti_alias_info anti_alias;

    /*
     * Flag to indicate if gx_color_index for this device may be divided
     * into individual fields for each component. This is almost always
     * the case for printers, and is the case for most modern displays
     * as well. When this is the case, halftoning may be performed
     * separately for each component, which greatly simplifies processing
     * when the number of color components is large.
     *
     * If the gx_color_index is separable in this manner, the comp_shift
     * array provides the location of the low-order bit for each
     * component. This may be filled in by the client, but need not be.
     * If it is not provided, it will be calculated based on the values
     * in the max_gray and max_color fields as follows:
     *
     *     comp_shift[num_components - 1] = 0,
     *     comp_shift[i] = comp_shift[i + 1]
     *                      + ( i == gray_index ? ceil(log2(max_gray + 1))
     *                                          : ceil(log2(max_color + 1)) )
     *
     * The comp_mask and comp_bits fields should be left empty by the client.
     * They will be filled in during initialization using the following
     * mechanism:
     *
     *     comp_bits[i] = ( i == gray_index ? ceil(log2(max_gray + 1))
     *                                      : ceil(log2(max_color + 1)) )
     *
     *     comp_mask[i] = (((gx_color_index)1 &lt;&lt; comp_bits[i]) - 1)
     *                       &lt;&lt; comp_shift[i]
     *
     * (For current devices, it is almost always the case that
     * max_gray == max_color, if the color model contains both gray and
     * non-gray components.)
     *
     * If separable_and_linear is not set, the data in the other fields
     * is unpredictable and should be ignored.
     */
    gx_color_enc_sep_lin_t separable_and_linear;
    byte                   comp_shift[GX_DEVICE_COLOR_MAX_COMPONENTS];
    byte                   comp_bits[GX_DEVICE_COLOR_MAX_COMPONENTS];
    gx_color_index         comp_mask[GX_DEVICE_COLOR_MAX_COMPONENTS];
    /*
     * Pointer to name for the process color model.
     */
    const char * cm_name;

} gx_device_color_info;
</pre>
</blockquote>

<p>
Note: See <a href="#Changing_color_info_data">Changing color_info data</a> before changing
any information in the <code>color_info structure</code> for a device.           

<p>
It is recommended that the values for this structure be defined using one
of the standard macros provided for this purpose. This allows for future
changes to be made to the structure without changes being required in the
actual device code.

<p>
The following macros (in <a href="../base/gxdevcli.h">gxdevcli.h</a>) provide
convenient shorthands for initializing this structure for ordinary
black-and-white or color devices:

<blockquote>
<code>#define dci_black_and_white</code> ...<br>
<code>#define dci_color(depth,maxv,dither)</code> ...
</blockquote>

<p>
The <code>#define dci_black_and_white</code> macro defines a
single bit monochrome device (For example: a typical monochrome printer device.)

<p>
The <code>#define dci_color(depth,maxv,dither)</code> macro can be used
to define a 24 bit RGB device or a 4 or 32 bit CMYK device.

<p>
The <code>#define dci_extended_alpha_values</code> macro (in
<a href="../base/gxdevcli.h">gxdevcli.h</a>) 
specifies most of the current fields in the structure. However this macro allows 
only the default setting for the comp_shift, comp_bits, and comp_mask fields 
to be set. Any device which requires a non-default setting for these fields 
has to correctly these fields during the device open procedure.
See 
<a href="#sep_and_linear_fields">Separable and linear fields></a> and
<a href="#Changing_color_info_data">Changing color_info data</a>.

<p>
The idea is that a device has a certain number of gray levels
(<code>max_gray</code>+1) and a certain number of colors
(<code>max_rgb</code>+1) that it can produce directly.  When Ghostscript
wants to render a given color space color value as a device color, it first tests
whether the color is a gray level and if so:

<blockquote>
If <code>max_gray</code> is large (&gt;= 31), Ghostscript asks the
device to approximate the gray level directly.  If the device returns a
valid <code>gx_color_index</code>, Ghostscript uses it.  Otherwise,
Ghostscript assumes that the device can represent
<code>dither_gray</code> distinct gray levels, equally spaced along the
diagonal of the color cube, and uses the two nearest ones to the desired
color for halftoning.
</blockquote>

<p>
If the color is not a gray level:

<blockquote>
If <code>max_rgb</code> is large (&gt;= 31), Ghostscript asks the device
to approximate the color directly.  If the device returns a valid
<code>gx_color_index</code>, Ghostscript uses it.  Otherwise,
Ghostscript assumes that the device can represent

<blockquote>
<code>dither_rgb</code> &times; <code>dither_rgb</code> &times; <code>dither_rgb</code>
</blockquote>

<p>
distinct colors, equally spaced throughout the color cube, and uses two of
the nearest ones to the desired color for halftoning.
</blockquote>

<h4><a name="sep_and_linear_fields"></a>Separable and linear fields</h4>
<p>
The three fields <code>comp_shift</code>, <code>comp_bits</code>, and 
<code>comp_mask</code> are only used if the <code>separable_and_linear</code> 
field is set to <code>GX_CINFO_SEP_LIN</code>. In this situation a <code>gx_color_index</code> 
value must represent a combination created by or'ing bits for each of the devices's 
output colorants. The <code>comp_shift</code> array defines the location 
(shift count) of each colorants bits in the output gx_color_index value. The 
<code>comp_bits</code> array defines the number of bits for each colorant. 
The <code>comp_mask</code> array contains a mask which can be used to isolate 
the bits for each colorant. These fields must be set if the device supports 
more than four colorants.

<h4><a name="Changing_color_info_data"></a>Changing color_info data</h4>

<p> For most devices, the information in the device's <code>color_info</code> 
structure is defined by the various device definition macros and the data remains 
constant during the entire existence of the device. In general the Ghostscript 
graphics assumes that the information is constant. However some devices want 
to modify the data in this structure.

<p>
The device's <code>put_params</code> procedure may change
<code>color_info</code> field values.
After the data has been modified then the 
device should be closed (via a call to <code>gs_closedevice</code>). Closing
the device will erase the current page so these changes should only be made
before anything has been drawn on a page.

<p> The device's <code>open_device</code> procedure may change
<code>color_info</code> field values. These changes should be done before
any other procedures are called.

<p>
The Ghostscript graphics library
uses some of the data in <code>color_info</code> to set the default
procedures for the 
<code>get_color_mapping_procs</code>,
<code>get_color_comp_index</code>,
<code>encode_color</code>, and
<code>decode_color</code> procedures.
These default procedures are set when the
device is originally created. If any changes are made to the
<code>color_info</code> fields then the device's <code>open_device</code>
procedure
has responsibility for insuring that the correct procedures are contained
in the device structure. (For an example, see the display device open procedure
<code>display_open</code> and its subroutine 
<code>display_set_color_format</code> 
(in <a href="../base/gdevdisp.c">gdevdisp</a>).


<h3><a name="Types"></a>Types</h3>

<p>
Here is a brief explanation of the various types that appear as parameters
or results of the drivers.

<dl>
<dt><code>frac</code> (defined in <a href="../base/gxfrac.h">gxfrac.h</a>)
<dd>This is the type used to represent color values for the input to the
color model mapping procedures. It is currently defined as a short.  It has a
range of <code>frac_0</code> to <code>frac_1</code>.
</dl>

<dl>
<dt><code>gx_color_value</code> (defined in
<a href="../base/gxdevice.h">gxdevice.h</a>)
<dd>This is the type used to represent RGB or CMYK color values.  It is
currently equivalent to unsigned short.  However, Ghostscript may use less
than the full range of the type to represent color values:
<code>gx_color_value_bits</code> is the number of bits actually used,
and <code>gx_max_color_value</code> is the maximum value, equal to
(2^<small><sup><code>gx_max_color_value_bits</code></sup></small>)-1.
</dl>

<dl>
<dt><code>gx_device</code> (defined in 
<a href="../base/gxdevice.h">gxdevice.h</a>)
<dd>This is the device structure, as explained above.
</dl>

<dl>
<dt><code>gs_matrix</code> (defined in 
<a href="../base/gsmatrix.h">gsmatrix.h</a>)
<dd>This is a 2-D homogeneous coordinate transformation matrix, used by
many Ghostscript operators.
</dl>

<dl>
<dt><code>gx_color_index</code> (defined in 
<a href="../base/gxcindex.h">gxcindex.h</a>)
<dd>This is meant to be whatever the driver uses to represent a device
color.  For example, it might be an index in a color map, or it might be R,
G, and B values packed into a single integer.  The Ghostscript graphics library
gets <code>gx_color_index</code> values from the device's
<code>encode_color</code> and hands them back as arguments to several other
procedures. If the <code>separable_and_linear</code> field in the device's
<code>color_info</code> structure is not set to
<code>GX_CINFO_SEP_LIN</code> then Ghostscript does not do
any computations with <code>gx_color_index</code> values.

<p>
The special
value <code>gx_no_color_index</code> (defined as
<code>(~(gx_color_index)(0))</code>&nbsp;) means "transparent" for some of
the procedures.

<p>
The size of <code>gx_color_index</code> can be either 32 or 64 bits. The 
choice depends upon the architecture of the CPU and the compiler. The default 
type definition is simply: 

<blockquote><code>
typedef unsigned long gx_color_index;
</code></blockquote>

However if <code>GX_COLOR_INDEX_TYPE</code> is defined, then it is used
as the type for <code>gx_color_index</code>.

<blockquote><code>
typedef GX_COLOR_INDEX_TYPE gx_color_index;
</code></blockquote>

The smaller size (32 bits) may produce more efficient or faster executing
code. The larger size (64 bits) is needed for representing either more
bits per component or more components. An example of the later case is 
a device that supports 8 bit contone colorants using a DeviceCMYK process
color model with its four colorants and also supports additional spot
colorants.

<p>
Currently autoconf attempts to find a 64 bit type definition for the
compiler being used, and if a 64 bit type is found then
<code>GX_COLOR_INDEX_TYPE</code> is set to the type.

<p>
For Microsoft and the MSVC compiler, <code>GX_COLOR_INDEX_TYPE</code> will 
be set to <code>unsigned _int64</code> if <code>USE_LARGE_COLOR_INDEX</code> 
is set to 1 either on the make command line or by editing the definition 
 in <a href="../psi/msvc32.mak">msvc32.mak</a> 
</dl>

<dl>
<dt><code>gs_param_list</code> (defined in <a
href="../base/gsparam.h">gsparam.h</a>) 
  <dd>This is a parameter list, which is used to read and set attributes in a
device.  See the comments in <a href="../base/gsparam.h">gsparam.h</a>, and
the <a href="#Parameters">description of the <code>get_params</code> and
<code>put_params</code> procedures</a> below, for more detail.
</dl>

<dl>
<dt><code>gx_tile_bitmap</code> (defined in
<a href="../base/gxbitmap.h">gxbitmap.h</a>)
<br><code>gx_strip_bitmap</code> (defined in
<a href="../base/gxbitmap.h">gxbitmap.h</a>)
<dd>These structure types represent bitmaps to be used as a tile for
filling a region (rectangle).  <code>gx_tile_bitmap</code> is an 
older, deprecated type lacking <code>shift</code> and 
<code>rep_shift</code>;
<code>gx_strip_bitmap</code> has superseded it, and should be
used in new code.  Here is a copy of the relevant part of the file:

<blockquote>
<pre>
/*
 * Structure for describing stored bitmaps.
 * Bitmaps are stored bit-big-endian (i.e., the 2^7 bit of the first
 * byte corresponds to x=0), as a sequence of bytes (i.e., you can't
 * do word-oriented operations on them if you're on a little-endian
 * platform like the Intel 80x86 or VAX).  Each scan line must start on
 * a (32-bit) word boundary, and hence is padded to a word boundary,
 * although this should rarely be of concern, since the raster and width
 * are specified individually.  The first scan line corresponds to y=0
 * in whatever coordinate system is relevant.
 *
 * For bitmaps used as halftone tiles, we may replicate the tile in
 * X and/or Y, but it is still valuable to know the true tile dimensions
 * (i.e., the dimensions prior to replication).  Requirements:
 *      width % rep_width = 0
 *      height % rep_height = 0
 *
 * For halftones at arbitrary angles, we provide for storing the halftone
 * data as a strip that must be shifted in X for different values of Y.
 * For an ordinary (non-shifted) halftone that has a repetition width of
 * W and a repetition height of H, the pixel at coordinate (X,Y)
 * corresponds to halftone pixel (X mod W, Y mod H), ignoring phase;
 * for a shifted halftone with shift S, the pixel at (X,Y) corresponds
 * to halftone pixel ((X + S * floor(Y/H)) mod W, Y mod H).    In other words,
 * each Y increment of H shifts the strip left by S pixels.
 *
 * As for non-shifted tiles, a strip bitmap may include multiple copies
 * in X or Y to reduce loop overhead.  In this case, we must distinguish:
 *      - The height of an individual strip, which is the same as
 *      the height of the bitmap being replicated (rep_height, H);
 *      - The height of the entire bitmap (size.y).
 * Similarly, we must distinguish:
 *      - The shift per strip (rep_shift, S);
 *      - The shift for the entire bitmap (shift).
 * Note that shift = (rep_shift * size.y / rep_height) mod rep_width,
 * so the shift member of the structure is only an accelerator.  It is,
 * however, an important one, since it indicates whether the overall
 * bitmap requires shifting or not.
 *
 * Note that for shifted tiles, size.y is the size of the stored bitmap
 * (1 or more strips), and NOT the height of the actual tile.  The latter
 * is not stored in the structure at all: it can be computed as H * W /
 * gcd(S, W).
 *
 * If the bitmap consists of a multiple of W / gcd(S, W) copies in Y, the
 * effective shift is zero, reducing it to a tile.  For simplicity, we
 * require that if shift is non-zero, the bitmap height be less than H * W /
 * gcd(S, W).  I.e., we don't allow strip bitmaps that are large enough to
 * include a complete tile but that don't include an integral number of
 * tiles.  Requirements:
 *      rep_shift &lt; rep_width
 *      shift = (rep_shift * (size.y / rep_height)) % rep_width
 *
 * For the benefit of the planar device, we now have a num_planes field.
 * For chunky data this should be set to 1. For planar data, the data pointer
 * points to the first plane of data; subsequent planes of data follow
 * immediately after this as if there were num_planes * height lines of data.
 */
typedef struct gx_strip_bitmap_s {
        byte *data;
        int raster;                     /* bytes per scan line */
        gs_int_point size;              /* width, height */
        gx_bitmap_id id;
        ushort rep_width, rep_height;   /* true size of tile */
        ushort rep_shift;
        ushort shift;
        int num_planes;
} gx_strip_bitmap;</pre>
</blockquote>
</dl>

<hr>

<h2><a name="Coding_conventions"></a>Coding conventions</h2>

<p>
All the driver procedures defined below that return <code>int</code>
results return 0 on success, or an appropriate negative error code in the
case of error conditions.  The error codes are defined in <a
href="../base/gserrors.h">gserrors.h</a>; they correspond directly to the
errors defined in the PostScript language reference manuals.  The most
common ones for drivers are:

<blockquote><dl>
<dt><code>gs_error_invalidfileaccess</code> 
<dd>An attempt to open a file failed.

<dt><code>gs_error_ioerror</code> 
<dd>An error occurred in reading or writing a file.

<dt><code>gs_error_limitcheck</code> 
 <dd>An otherwise valid parameter value was too large for the
implementation.

<dt><code>gs_error_rangecheck</code> 
<dd>A parameter was outside the valid range.

<dt><code>gs_error_VMerror</code> 
<dd>An attempt to allocate memory failed.  (If this happens, the procedure
should release all memory it allocated before it returns.)
</dl></blockquote>

<p>
If a driver does return an error, rather than a simple return statement it
should use the <code>return_error</code> macro defined in <a
href="../base/gx.h">gx.h</a>, which is automatically included by <a
href="../base/gdevprn.h">gdevprn.h</a> but not by <a
href="../base/gserrors.h">gserrors.h</a>.  For example

<blockquote>
<code> return_error(gs_error_VMerror);
</code></blockquote>

<h3><a name="Allocating_storage"></a>Allocating storage</h3>

<p>
While most drivers (especially printer drivers) follow a very similar
template, there is one important coding convention that is not obvious from
reading the code for existing drivers: driver procedures must not use
<code>malloc</code> to allocate any storage that stays around after the
procedure returns. Instead, they must use <code>gs_malloc</code> and
<code>gs_free</code>, which have slightly different calling conventions.
(The prototypes for these are in <a href="../base/gsmemory.h">gsmemory.h</a>, 
which is included in <a href="../base/gx.h">gx.h</a>, which is included in <a
href="../base/gdevprn.h">gdevprn.h</a>.) This is necessary so that
Ghostscript can clean up all allocated memory before exiting, which is
essential in environments that provide only single-address-space
multi-tasking (some versions of Microsoft Windows). 

<blockquote>
<pre>char *gs_malloc(uint num_elements, uint element_size,
  const char *client_name);</pre>
</blockquote>

<p>
Like <code>calloc</code>, but unlike <code>malloc</code>,
<code>gs_malloc</code> takes an element count and an element size. For
structures, <code>num_elements</code> is 1 andi
<code>element_size</code> is <code>sizeof</code> the structure; for
byte arrays, <code>num_elements</code> is the number of bytes and
<code>element_size</code> is 1. Unlike <code>calloc</code>,
<code>gs_malloc</code> does <b>not</b> clear the block of storage. 

<p>
The <code>client_name</code> is used for tracing and debugging.  It must
be a real string, not <code>NULL</code>.  Normally it is the name of the
procedure in which the call occurs.

<blockquote>
<pre>void gs_free(char *data, uint num_elements, uint element_size,
  const char *client_name);</pre>
</blockquote>

<p>
Unlike <code>free</code>, <code>gs_free</code> demands that
<code>num_elements</code> and element_size be supplied. It also
requires a client name, like <code>gs_malloc</code>. 

<h3><a name="Driver_instance_allocation"></a>Driver instance allocation</h3>

<p>
All driver instances allocated by Ghostscript's standard allocator must
point to a "structure descriptor" that tells the garbage collector how to
trace pointers in the structure. For drivers registered in the normal way
(using the makefile approach described above), no special care is needed as
long as instances are created only by calling the
<code>gs_copydevice</code> procedure defined in <a
href="../base/gsdevice.h">gsdevice.h</a>. If you have a need to define
devices that are not registered in this way, you must fill in the stype
member in any dynamically allocated instances with a pointer to the same
structure descriptor used to allocate the instance. For more information
about structure descriptors, see <a href="../base/gsmemory.h">gsmemory.h</a>
and <a href="../base/gsstruct.h">gsstruct.h</a>. 

<hr>

<h2><a name="Printer_drivers"></a>Printer drivers</h2>

<p>
Printer drivers (which include drivers that write some kind of raster file)
are especially simple to implement.  
The printer driver must implement a <code>print_page</code> or
<code>print_page_copies</code> procedure.  There are macros in <a
href="../base/gdevprn.h">gdevprn.h</a> that generate the device structure for
such devices, of which the simplest is <code>prn_device</code>; for an
example, see <a href="../base/gdevbj10.c">gdevbj10.c</a>.  If you are writing
a printer driver, we suggest you start by reading <a
href="../base/gdevprn.h">gdevprn.h</a> and the <a
href="#Color_mapping">subsection on "Color mapping"</a> below; you may be
able to ignore all the rest of the driver procedures.

<p>
The <code>print_page</code> procedures are defined as follows:

<blockquote>
<pre>int (*print_page)(gx_device_printer *, FILE *)
int (*print_page_copies)(gx_device_printer *, FILE *, int)</pre>
</blockquote>

<p>
This procedure must read out the rendered image from the device and write
whatever is appropriate to the file.  To read back one or more scan lines
of the image, the <code>print_page</code> procedure must call one of the
following procedures:

<blockquote>
<pre>int gdev_prn_copy_scan_lines(gx_device_printer *pdev, int y, byte *str,
    uint size)</pre>
</blockquote>

<p>
For this procedure, <code>str</code> is where the data should be copied to, and <code>size</code> is
the size of the buffer starting at <code>str</code>.  This procedure returns the number
of scan lines copied, or &lt;0 for an error.  <code>str</code> need not be aligned.

<blockquote>
<pre>int gdev_prn_get_bits(gx_device_printer *pdev, int y, byte *str,
  byte **actual_data)</pre>
</blockquote>

<p>
This procedure reads out exactly one scan line.  If the scan line is
available in the correct format already, <code>*actual_data</code> is
set to point to it; otherwise, the scan line is copied to the buffer
starting at <code>str</code>, and <code>*actual_data</code> is set to
<code>str</code>.  This saves a copying step most of the time.
<code>str</code> need not be aligned; however, if
<code>*actual_data</code> is set to point to an existing scan line, it
will be aligned.  (See the description of the <code>get_bits</code>
procedure below for more details.)

<p>
In either case, each row of the image is stored in the form described in
the comment under <code>gx_tile_bitmap</code> above; each pixel takes
the number of bits specified as <code>color_info.depth</code> in the
device structure, and holds values returned by the device's
<code>encode_color</code> procedure.

<p>
The <code>print_page</code> procedure can determine the number of bytes
required to hold a scan line by calling:

<blockquote>
<pre>uint gdev_prn_raster(gx_device_printer *)</pre>
</blockquote>

<p>
For a very simple concrete example, we suggest reading the code in
<code>bit_print_page</code> in <a href="../base/gdevbit.c">gdevbit.c</a>.

<p>
If the device provides <code>print_page</code>, Ghostscript will call
<code>print_page</code> the requisite number of times to print the
desired number of copies; if the device provides
<code>print_page_copies</code>, Ghostscript will call
<code>print_page_copies</code> once per page, passing it the desired
number of copies.

<hr>

<h2><a name="Driver_procedures"></a>Driver procedures</h2>

<p>
Most of the procedures that a driver may implement are optional.  If a
device doesn't supply an optional procedure <code>WXYZ</code>, the entry
in the procedure structure may be either <code>gx_default_WXYZ</code>,
for instance <code>gx_default_tile_rectangle</code>, or
<code>NULL</code> or 0.  (The device procedure must also call the
<code>gx_default_</code> procedure if it doesn't implement the function
for particular values of the arguments.)  Since C compilers supply 0 as the
value for omitted structure elements, this convention means that statically
initialized procedure structures continue to work even if new (optional)
members are added.

<h3><a name="Life_cycle"></a>Life cycle</h3>

<p>
A device instance begins life in a closed state.  In this state, no output
operations will occur.  Only the following procedures may be called:

<blockquote><code>
open_device<br>
finish_copydevice<br>
get_initial_matrix<br>
get_params<br>
put_params<br>
get_hardware_params
</code></blockquote>

<p>
When <code>setdevice</code> installs a device instance in the graphics
state, it checks whether the instance is closed or open.  If the instance
is closed, <code>setdevice</code> calls the open routine, and then sets
the state to open.

<p>
There is no user-accessible operation to close a device instance.  This is
not an oversight -- it is required in order to enforce the following
invariant:

<blockquote>
If a device instance is the current device in <em>any</em> graphics state,
it must be open (have <code>is_open</code> set to true).
</blockquote>

<p>
Device instances are only closed when they are about to
be freed, which occurs in three situations:

<ul>
<li>When a <code>restore</code> occurs, if the instance was created since
the corresponding <code>save</code> and is in a VM being restored.  I.e.,
if the instance was created in local VM since a <code>save</code>, it
will always be closed and freed by the corresponding
<code>restore</code>; if it was created in global VM, it will only be
closed by the outermost <code>restore</code>, regardless of the save
level at the time the instance was created.

<li>By the garbage collector, if the instance is no longer accessible.

<li>When Ghostscript exits (terminates).
</ul>

<h3><a name="Open_close"></a>Open, close, sync, copy</h3>

<dl>
<dt><code>int (*open_device)(gx_device *)</code> <b><em>[OPTIONAL]</em></b>
<dd>Open the device: do any initialization associated with making the device
instance valid. This must be done before any output to the device. The
default implementation does nothing. <b>NOTE</b>: Clients should never call
a device's <code>open_device</code> procedure directly: they should
always call <code>gs_opendevice</code> instead. 
</dl>

<dl>
<dt><code>int (*finish_copydevice)(gx_device *dev, const gx_device
*from_dev)</code> <b><em>[OPTIONAL]</em></b> <dd>Perform any cleanup
required after <code>copydevice</code> has created a new device instance
by copying <code>from_dev</code>. If the copy operation should not be
allowed, this procedure should return an error; the copy will be freed. The
default implementation allows copying the device prototype, but does not
allow copying device instances, because instances may contain internal
pointers that should not be shared between copies, and there is no way to
determine this from outside the device. <b>NOTE</b>: Clients should never
call a device's <code>finish_copydevice</code> procedure: this procedure 
is only intended for use by <code>gs_copydevice[2]</code>. 
</dl>

<dl>
<dt><code>void (*get_initial_matrix)(gx_device *, gs_matrix *)</code> <b><em>[OPTIONAL]</em></b> 
<dd>Construct the initial transformation matrix mapping user coordinates
(nominally 1/72 inch per unit) to device coordinates.  The default
procedure computes this from width, height, and
[<code>xy</code>]<code>_pixels_per_inch</code> on the assumption that
the origin is in the upper left corner, that is
<blockquote>
<code>xx</code> = <code>x_pixels_per_inch</code>/72, <code>xy</code> = 0,<br>
<code>yx = 0, yy = -y_pixels_per_inch</code>/72,<br>
<code>tx = 0, ty = height</code>.
</blockquote>
</dl>

<dl>
<dt><code>int (*sync_output)(gx_device *)</code> <b><em>[OPTIONAL]</em></b>
<dd>Synchronize the device.  If any output to the device has been
buffered, send or write it now.  Note that this may be called several times
in the process of constructing a page, so printer drivers should <b>not</b>
implement this by printing the page.  The default implementation does
nothing.
</dl>

<dl>
<dt><code>int (*output_page)(gx_device *, int num_copies, int flush)</code> <b><em>[OPTIONAL]</em></b>
<dd>Output a fully composed page to the device.  The
<code>num_copies</code> argument is the number of copies that should be
produced for a hardcopy device.  (This may be ignored if the driver has
some other way to specify the number of copies.)  The <code>flush</code>
argument is true for <code>showpage</code>, false for
<code>copypage</code>.  The default definition just calls
<code>sync_output</code>.  Printer drivers should implement this by
printing and ejecting the page.
</dl>

<dl>
<dt><code>int (*close_device)(gx_device *)</code> <b><em>[OPTIONAL]</em></b>
<dd>Close the device: release any associated resources. After this, output
to the device is no longer allowed. The default implementation does
nothing.  <b>NOTE</b>: Clients should never call a device's
<code>close_device</code> procedure directly: they should always call
<code>gs_closedevice</code> instead. 
</dl>
<h3><a name="Color_mapping"></a>Color and alpha mapping</h3>

<p>
Note that code in the Ghostscript library may cache the results of calling
one or more of the color mapping procedures.  If the result returned by any
of these procedures would change (other than as a result of a change made by
the driver's <code>put_params</code> procedure), the driver must call
<code>gx_device_decache_colors(dev)</code>.

<p>
The <code>map_rgb_color</code>, <code>map_color_rgb</code>, and 
<code>map_cmyk_color</code> are obsolete. They have been left
in the device procedure list for backward compatibility. See the
<code>encode_color</code> and <code>decode_color</code> procedures
below. To insure that older device drivers are changed to use the new
<code>encode_color</code> and <code>decode_color</code> 
procedures,
the parameters for the older procedures have been changed to
match the new procedures.  To minimize changes in devices that have
already been written, the map_rgb_color and map_cmyk_color routines
are used as the default value for the encode_color routine.  The
map_cmyk_color routine is used if the number of components is four.
The map_rgb_color routine is used if the number of components is one
or three. This works okay for RGB and CMYK process color model devices.
However this does not work properly for gray devices. The encode_color
routine for a gray device is only passed one component. Thus the
map_rgb_color routine must be modified to only use a single input (instead
of three).  (See the encode_color and decode_color routines below.)


<p>
Colors can be specified to the Ghostscript graphics library in a variety
of forms.  For example, there are a wide variety of color spaces that can
be used such as Gray, RGB, CMYK, DeviceN, Separation, Indexed, CIEbasedABC,
etc.  The graphics library converts the various input color space
values into four base color spaces: Gray, RGB, CMYK, and DeviceN. The
DeviceN color space allows for specifying values for individual device
colorants or spot colors.

<p>
Colors are converted by the device in a two step process. The first step
is to convert a color in one of the base color spaces (Gray, RGB, CMYK,
or DeviceN) into values for each device colorant.  This transformation is
done via a set of procedures provided by the device.  These procedures are
provided by the <code>get_color_mapping_procs</code> device procedure.

<p>
Between the first and second steps, the graphics library applies transfer
functions to the device colorants. Where needed, the output of the results
after the transfer functions is used by the graphics library for halftoning.

<p>
In the second step, the device procedure <code>encode_color</code> is
used to convert the transfer function results into a
<code>gx_color_index</code> value.
The <code>gx_color_index</code> values are passed to specify colors
to various routines.
The choice of the encoding for a <code>gx_color_index</code> is
up to the device. Common choices are indexes into a color palette or
several integers packed together into a single value. The manner of this
encoding is usually opaque to the graphics library. The only exception to this
statement occurs when halftoning 5 or more colorants. In this case the
graphics library assumes that if a colorant values is zero then the
bits associated with the colorant in the <code>gx_color_index</code>
value are zero.

<dl>
<dt><code>int get_color_comp_index(const gx_device&nbsp;* dev, const char * pname,
int name_size, int src_index)</code> <b><em>[OPTIONAL]</em></b> 
<dd>This procedure returns the device colorant number of the given name.
The possible return values are -1, 0 to
<code>GX_DEVICE_COLOR_MAX_COMPONENTS - 1</code>, or
<code>GX_DEVICE_COLOR_MAX_COMPONENTS</code>. A value of -1 indicates that
the specified name is not a colorant for the device. A value of 0 to
<code>GX_DEVICE_COLOR_MAX_COMPONENTS - 1</code> indicates the colorant number
of the given name. A value of <code>GX_DEVICE_COLOR_MAX_COMPONENTS</code>
indicates that the given name is a valid colorant name for the device but the
colorant is not currently being used. This is used for implementing names
which are in SeparationColorNames but not in SeparationOrder.

<p>
The default procedure returns results based upon process color model
of DeviceGray, DeviceRGB, or DeviceCMYK selected by
<code>color_info.num_components</code>. This procedure must be
defined if another process color model is used by the device or spot colors are
supported by the device.
</dd>
</dl>

<dl>
<dt><code>const gx_cm_color_map_procs&nbsp;* get_color_mapping_procs(const
gx_device&nbsp;* dev)</code> <b><em>[OPTIONAL]</em></b> 
<dd>This procedure returns a list of three procedures. These procedures
are used to translate values in either Gray, RGB, or CMYK color spaces
into device colorant values. A separate procedure is not required for the
DeviceN and Separation color spaces since these already represent
device colorants.

<p>
The default procedure returns a list of procedures based upon
<code>color_info.num_components</code>.  These procedures are appropriate
for DeviceGray, DeviceRGB, or DeviceCMYK process color model devices. A
procedure must be defined if another process color model is used by the
device or spot colors are to be supported.
</dd>
</dl>

<dl>
<dt><code>gx_color_index (*encode_color)(gx_device&nbsp;* dev,
gx_color_value&nbsp;* cv)</code> <b><em>[OPTIONAL]</em></b>
<dd>Map a set of device color values into a <code>gx_color_index</code>
value. The range of legal values of the
arguments is 0 to <code>gx_max_color_value</code>.  The default procedure
packs bits into a <code>gx_color_index</code> value based upon the
values in <code>color_info.depth</code> and
<code>color_info.num_components</code>.

<p>
Note that the <code>encode_color</code> procedure
must not return <code>gx_no_color_index</code> (all 1s).
</dl>

<dl>
<dt><code>int (*decode_color)(gx_device&nbsp;*, gx_color_index&nbsp;color,
gx_color_value&nbsp;*&nbsp;CV)</code> <b><em>[OPTIONAL]</em></b> 
<dd>This is the inverse of the <code>encode_color</code> procedure.
Map a <code>gx_color_index</code> value to color values.  The default
procedure unpacks bits from the <code>gx_color_index</code> value based upon
the values in <code>color_info.depth</code> and
<code>color_info.num_components</code>.
</dl>

<dl>
<dt><code>gx_color_index (*map_rgb_alpha_color)(gx_device&nbsp;*,
gx_color_value&nbsp;red, gx_color_value&nbsp;green,
gx_color_value&nbsp;blue, gx_color_value&nbsp;alpha)</code> <b><em>[OPTIONAL]</em></b>
<dd>Map a RGB color and an opacity value to a device color.  The range of
legal values of the RGB and alpha arguments is 0 to
<code>gx_max_color_value</code>; <code>alpha</code> = 0 means
transparent, <code>alpha</code> = <code>gx_max_color_value</code>
means fully opaque.  The default is to use the
<code>encode_color</code> procedure and ignore alpha.

<p>
Note that if a driver implements <code>map_rgb_alpha_color</code>, it
must also implement <code>encode_color</code>, and must implement them
in such a way that
<code>map_rgb_alpha_color(dev,&nbsp;r,&nbsp;g,&nbsp;b,&nbsp;gx_max_color_value)</code> 
returns the same value as
<code>encode_color(dev,&nbsp;CV)</code>. 
</dl>

<dl>
<dt><code>int (*map_color_rgb_alpha)(gx_device&nbsp;*,
gx_color_index&nbsp;color, gx_color_value&nbsp;rgba[4])</code>
<b><em>[OPTIONAL]</em></b>
<dd>Map a device color code to RGB and alpha values.  The default
implementation calls <code>map_color_rgb</code> and fills in
<code>gx_max_color_value</code> for alpha.

<p>
Note that if a driver implements <code>map_color_rgb_alpha</code>, it
must also implement <code>decode_color</code>, and must implement them
in such a way that the first 3 values returned by
<code>map_color_rgb_alpha</code> are the same as the values returned by
<code>decode_color</code>.

<p>
Note that only RGB devices currently support variable opacity; alpha is ignored 
on other devices. The PDF 1.4 transparency features are supported on all devices. 
</dl>

<dl>
<dt><code>typedef&nbsp;enum&nbsp;{&nbsp;go_text,
go_graphics&nbsp;}&nbsp;graphic_object_type;&nbsp;int
(*get_alpha_bits)(gx_device&nbsp;*dev,
graphic_object_type&nbsp;type)</code> <b><em>[OPTIONAL] [OBSOLETE]</em></b>
<dd>This procedure is no longer used: it is replaced by the
color_info.anti_alias member of the driver structure.  However, it still
appears in the driver procedure vector for backward compatibility.  It
should never be called, and drivers should not implement it.
</dl>

<dl>
<dt><code>void (*update_spot_equivalent_colors)(gx_device&nbsp;*,
const gs_state *)</code>
<b><em>[OPTIONAL]</em></b>
<dd>This routine provides a method for the device to gather an equivalent
color for spot colorants. This routine is called when a Separation or DeviceN
color space is installed.  See comments at the start of
<a href="../base/gsequivc.c">gsequivc.c</a>. Note: This procedure is only needed
for devices that support spot colorants and also need to have an equivalent
color for simulating the appearance of the spot colorants.
</dl>

<h3><a name="Pixel_level_drawing"></a>Pixel-level drawing</h3>

<p>
This group of drawing operations specifies data at the pixel level.  All
drawing operations use device coordinates and device color values.

<dl>
<dt><code>int (*fill_rectangle)(gx_device&nbsp;*, int&nbsp;x,
int&nbsp;y, int&nbsp;width, int&nbsp;height,
gx_color_index&nbsp;color)</code>
<dd>Fill a rectangle with a color.  The set of pixels filled is {(px,py) |
x &lt;= px &lt; x + width and y &lt;= py &lt; y + height}.  In other words,
the point <em>(x,y)</em> is included in the rectangle, as are
<em>(x+w-1,y)</em>, <em>(x,y+h-1)</em>, and <em>(x+w-1,y+h-1)</em>, but
<b><em>not</em></b> <em>(x+w,y)</em>, <em>(x,y+h)</em>, or
<em>(x+w,y+h)</em>.  If <code>width</code>&nbsp;&lt;=&nbsp;0 or
height&nbsp;&lt;=&nbsp;0, <code>fill_rectangle</code> should return 0
without drawing anything.

<p>
Note that <code>fill_rectangle</code> is the only non-optional procedure
in the driver interface.
</dl>

<h4><a name="Bitmap_imaging"></a>Bitmap imaging</h4>

<p>
Bitmap (or pixmap) images are stored in memory in a nearly standard way.
The first byte corresponds to <em>(0,0)</em> in the image coordinate
system: bits (or polybit color values) are packed into it left to right.
There may be padding at the end of each scan line: the distance from one
scan line to the next is always passed as an explicit argument.

<dl>
<dt><code>int (*copy_mono)(gx_device&nbsp;*,
const&nbsp;unsigned&nbsp;char&nbsp;*data, int&nbsp;data_x, int&nbsp;raster,
gx_bitmap_id&nbsp;id, int&nbsp;x, int&nbsp;y, int&nbsp;width,
int&nbsp;height, gx_color_index&nbsp;color0,
gx_color_index&nbsp;color1)</code> <b><em>[OPTIONAL]</em></b>
<dd>Copy a monochrome image (similar to the PostScript image operator).
Each scan line is raster bytes wide.  Copying begins at
(<code>data_x</code>,0) and transfers a rectangle of the given width and
height to the device at device coordinate <em>(x,y)</em>.  (If the transfer
should start at some non-zero y value in the data, the caller can adjust
the data address by the appropriate multiple of the raster.)  The copying
operation writes device color <code>color0</code> at each 0-bit, and
<code>color1</code> at each 1-bit: if <code>color0</code> or
<code>color1</code> is <code>gx_no_color_index</code>, the device
pixel is unaffected if the image bit is 0 or 1 respectively.  If
<code>id</code> is different from <code>gx_no_bitmap_id</code>, it
identifies the bitmap contents unambiguously; a call with the same
<code>id</code> will always have the same <code>data</code>,
<code>raster</code>, and data contents.

<p>
This operation, with
<code>color0</code>&nbsp;=&nbsp;<code>gx_no_color_index</code>, is
the workhorse for text display in Ghostscript, so implementing it
efficiently is very important.
</dl>

<dl>
<dt><code>int (*tile_rectangle)(gx_device&nbsp;*,
const&nbsp;gx_tile_bitmap&nbsp;*tile, int&nbsp;x, int&nbsp;y,
int&nbsp;width, int&nbsp;height, gx_color_index&nbsp;color0,
gx_color_index&nbsp;color1, int&nbsp;phase_x, int&nbsp;phase_y)</code>
<b><em>[OPTIONAL] [OBSOLETE]</em></b>
<dd>This procedure is still supported, but has been superseded by
<code>strip_tile_rectangle</code>.  New drivers should implement
<code>strip_tile_rectangle</code>; if they cannot cope with non-zero
shift values, they should test for this explicitly and call the default
implementation (<code>gx_default_strip_tile_rectangle</code>) if
shift&nbsp;!=&nbsp;0.  Clients should call
<code>strip_tile_rectangle</code>, not <code>tile_rectangle</code>.
</dl>

<dl>
<dt><code>int (*strip_tile_rectangle)(gx_device&nbsp;*,
const&nbsp;gx_strip_bitmap&nbsp;*tile, int&nbsp;x, int&nbsp;y,
int&nbsp;width, int&nbsp;height, gx_color_index&nbsp;color0,
gx_color_index&nbsp;color1, int&nbsp;phase_x, int&nbsp;phase_y)</code>
<b><em>[OPTIONAL]</em></b>
<dd>Tile a rectangle.  Tiling consists of doing multiple
<code>copy_mono</code> operations to fill the rectangle with copies of
the tile.  The tiles are aligned with the device coordinate system, to
avoid "seams".  Specifically, the (<code>phase_x</code>,
<code>phase_y</code>) point of the tile is aligned with the origin of
the device coordinate system.  (Note that this is backwards from the
PostScript definition of halftone phase.)  <code>phase_x</code> and
<code>phase_y</code> are guaranteed to be in the range
<em>[0..</em><code>tile-&gt;width</code><em>)</em> and
<em>[0..</em><code>tile-&gt;height</code><em>)</em> respectively.

<p>
If <code>color0</code> and <code>color1</code> are both
<code>gx_no_color_index</code>, then the tile is a color pixmap, not a
bitmap: see the next section.

<p>
This operation is the workhorse for halftone filling in Ghostscript, so
implementing it efficiently for solid tiles (that is, where either
<code>color0</code> and <code>color1</code> are both
<code>gx_no_color_index</code>, for colored halftones, or neither one is
<code>gx_no_color_index</code>, for monochrome halftones) is very
important.
</dl>

<h4><a name="Pixmap_imaging"></a>Pixmap imaging</h4>

<p>
Pixmaps are just like bitmaps, except that each pixel occupies more than
one bit.  All the bits for each pixel are grouped together (this is
sometimes called "chunky" or "Z" format).  For <code>copy_color</code>,
the number of bits per pixel is given by the
<code>color_info.depth</code> parameter in the device structure: the
legal values are 1, 2, 4, 8, 16, 24, 32, 40, 48, 56, or 64.  The pixel
values are device color codes (that is, whatever it is that
<code>encode_color</code> returns).

<dl>
<dt><code>int (*copy_color)(gx_device&nbsp;*,
const&nbsp;unsigned&nbsp;char&nbsp;*data, int&nbsp;data_x, int&nbsp;raster,
gx_bitmap_id&nbsp;id, int&nbsp;x, int&nbsp;y, int&nbsp;width,
int&nbsp;height)</code> <b><em>[OPTIONAL]</em></b>
<dd>Copy a color image with multiple bits per pixel.  The raster is in
bytes, but <code>x</code> and <code>width</code> are in pixels, not
bits.  If <code>id</code> is different from
<code>gx_no_bitmap_id</code>, it identifies the bitmap contents
unambiguously; a call with the same <code>id</code> will always have the
same <code>data</code>, <code>raster</code>, and data contents.

<p>
We do not provide a separate procedure for tiling with a pixmap; instead,
<code>tile_rectangle</code> can also take colored tiles.  This is
indicated by the <code>color0</code> and <code>color1</code>
arguments' both being <code>gx_no_color_index</code>.  In this case, as
for <code>copy_color</code>, the <code>raster</code> and
<code>height</code> in the "bitmap" are interpreted as for real bitmaps,
but the <code>x</code> and <code>width</code> are in pixels, not
bits.
</dl>

<h4><a name="Compositing"></a>Compositing</h4>

<p>
In addition to direct writing of opaque pixels, devices must also support
compositing.  Currently two kinds of compositing are defined
(<code>RasterOp</code> and alpha-based), but more may be added in the
future.

<blockquote>
<b><em>THIS AREA OF THE INTERFACE IS SOMEWHAT UNSTABLE: USE AT YOUR OWN
RISK.</em></b>
</blockquote>

<dl>
<dt><code>int (*copy_alpha)(gx_device&nbsp;*dev,
const&nbsp;unsigned&nbsp;char&nbsp;*data, int&nbsp;data_x, int&nbsp;raster,
gx_bitmap_id&nbsp;id, int&nbsp;x, int&nbsp;y, int&nbsp;width,
int&nbsp;height, gx_color_index&nbsp;color, int&nbsp;depth)</code>
<b><em>[OPTIONAL]</em></b>
<dd>This procedure is somewhat misnamed: it was added to the interface
before we really understood alpha channel and compositing.

<p>
Fill a given region with a given color modified by an individual alpha
value for each pixel.  For each pixel, this is equivalent to
alpha-compositing with a source pixel whose alpha value is obtained from
the pixmap (<code>data</code>, <code>data_x</code>, and
<code>raster</code>) and whose color is the given color (which has
<b><em>not</em></b> been premultiplied by the alpha value), using the Sover
rule.  <code>depth</code>, the number of bits per alpha value, is either
2 or 4, and in any case is always a value returned by a previous call on
the <code>get_alpha_bits</code> procedure.  Note that if
<code>get_alpha_bits</code> always returns 1, this procedure will never
be called.
</dl>

<dl>
<dt><code>int (*create_compositor)(dev_t&nbsp;*dev,
gx_device_t&nbsp;**pcdev, const&nbsp;gs_composite_t&nbsp;*pcte,
const&nbsp;gs_imager_state&nbsp;*pis, gs_memory_t&nbsp;*memory)</code>
<b><em>[OPTIONAL]</em></b> 
<dd>Create a new device (called a "compositing device" or "compositor")
that will composite data written to it with the device's existing data,
according to the compositing function defined by <code>*pcte</code>.
Devices will normally implement this in one of the following standard ways:

<ul>
<li>Devices that don't do any imaging and don't forward any imaging
operations (for example, the null device, the hit detection device, and the
clipping list accumulation device) simply return themselves, which
effectively ignores the compositing function.

<li>"Leaf" devices that do imaging and have no special optimizations for
compositing (for example, some memory devices) ask the
<code>gs_composite_t</code> to create a default compositor. 

<li>Leaf devices that can implement some kinds of compositing operation 
efficiently (for example, monobit memory devices and RasterOp) inspect the
type and values of <code>*pcte</code> to determine whether it specifies 
such an operation: if so, they create a specialized compositor, and if not,
they ask the <code>gs_composite_t</code> to create a default compositor. 
</ul>

<p>
Other kinds of forwarding devices, which don't fall into any of these
categories, require special treatment.  In principle, what they do is ask
their target to create a compositor, and then create and return a copy of
themselves with the target's new compositor as the target of the copy.
There is a possible default implementation of this approach: if the
original device was <b>D</b> with target <b>T</b>, and <b>T</b> creates a
compositor <b>C</b>, then the default implementation creates a device
<b>F</b> that for each operation temporarily changes <b>D</b>'s target to
<b>C</b>, forwards the operation to <b>D</b>, and then changes <b>D</b>'s
target back to <b>T</b>.  However, the Ghostscript library currently only
creates a compositor with an imaging forwarding device as target in a few
specialized situations (banding, and bounding box computation), and these
are handled as special cases.

<p>
Note that the compositor may have a different color space, color
representation, or bit depth from the device to which it is compositing.
For example, alpha-compositing devices use standard-format chunky color
even if the underlying device doesn't.

<p>
Closing a compositor frees all of its storage, including the compositor
itself.  However, since the <code>create_compositor</code> call may
return the same device, clients must check for this case, and only call the
close procedure if a separate device was created.
</dl>

<p>
<font size="+1">
<b><em>[strip_]copy_rop WILL BE SUPERSEDED BY COMPOSITORS</em></b>
</font>

<dl>
<dt><code>int (*copy_rop)(gx_device&nbsp;*dev,
const&nbsp;byte&nbsp;*sdata, int&nbsp;sourcex, uint&nbsp;sraster,
gx_bitmap_id&nbsp;id, const&nbsp;gx_color_index&nbsp;*scolors,
const&nbsp;gx_tile_bitmap&nbsp;*texture,
const&nbsp;gx_color_index&nbsp;*tcolors, int&nbsp;x, int&nbsp;y,
int&nbsp;width, int&nbsp;height, int&nbsp;phase_x, int&nbsp;phase_y,
int&nbsp;command)</code> <b><em>[OPTIONAL]</em></b>
<dd>This procedure is still supported, but has been superseded by
<code>strip_copy_rop</code>.  New drivers should implement
<code>strip_copy_rop</code>; if they cannot cope with non-zero shift
values in the texture, they should test for this explicitly and call the
default implementation (<code>gx_default_strip_copy_rop</code>) if
shift&nbsp;!=&nbsp;0.  Clients should call <code>strip_copy_rop</code>,
not <code>copy_rop</code>.
</dl>

<dl>
<dt><code>int (*strip_copy_rop)(gx_device&nbsp;*dev,
const&nbsp;byte&nbsp;*sdata, int&nbsp;sourcex, uint&nbsp;sraster,
gx_bitmap_id&nbsp;id, const&nbsp;gx_color_index&nbsp;*scolors,
const&nbsp;gx_strip_bitmap&nbsp;*texture,
const&nbsp;gx_color_index&nbsp;*tcolors, int&nbsp;x, int&nbsp;y,
int&nbsp;width, int&nbsp;height, int&nbsp;phase_x, int&nbsp;phase_y,
int&nbsp;command)</code> <b><em>[OPTIONAL]</em></b>
<dd>Combine an optional source image <b>S</b> (as for
<code>copy_mono</code> or <code>copy_color</code>) and an optional
texture <b>T</b> (a tile, as for <code>tile_rectangle</code>) with the
existing bitmap or pixmap <b>D</b> held by the driver, pixel by pixel,
using any 3-input Boolean operation as modified by "transparency" flags:
schematically, set <b>D&nbsp;=&nbsp;f(D,S,T)</b>, computing <b>f</b> in RGB
space rather than using actual device pixel values.  <b>S</b> and <b>T</b>
may each (independently) be a solid color, a bitmap with "foreground" and
"background" colors, or a pixmap.  This is a complex (and currently rather
slow) operation.  The arguments are as follows:

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=top>	<td><code>dev</code>
	<td>&nbsp;
	<td>the device, as for all driver procedures
<tr valign=top>	<td><code>sdata</code>, <code>sourcex</code>, <code>sraster</code>, <code>id</code>, <code>scolors</code>
	<td>&nbsp;
	<td>specify <b>S</b>, <a href="#S_spec">see below</a>
<tr valign=top>	<td><code>texture</code>, <code>tcolors</code>
	<td>&nbsp;
	<td>specify <b>T</b>, <a href="#T_spec">see below</a>
<tr valign=top>	<td><code>x</code>, <code>y</code>, <code>width</code>, <code>height</code>
	<td>&nbsp;
	<td>as for the other copy and fill procedures
<tr valign=top>	<td><code>phase_x</code>, <code>phase_y</code>
	<td>&nbsp;
	<td>part of <b>T</b> specification, <a href="#T_spec">see below</a>
<tr valign=top>	<td><code>command</code>
	<td>&nbsp;
	<td><a href="#F_spec">see below</a>
</table></blockquote>
</dl>

<h5><a name="S_spec"></a>The source specification S</h5>

<p>
As noted above, the source <b>S</b> may be a solid color, a bitmap, or a
pixmap.  If <b>S</b> is a solid color:

<ul>
<li><code>sdata</code>, <code>sourcex</code>,
<code>sraster</code>, and <code>id</code> are irrelevant.

<li><code>scolors</code> points to two <code>gx_color_index</code>
values; <code>scolors[0]</code> = <code>scolors[1]</code> = the
color.
</ul>

<p>
If <b>S</b> is a bitmap:

<ul>
<li><code>sdata</code>, <code>sourcex</code>,
<code>sraster</code>, and <code>id</code> arguments are as for
<code>copy_mono</code> or <code>copy_color</code>
(<code>data</code>, <code>data_x</code>, <code>raster</code>,
<code>id</code>), and specify a source bitmap.

<li><code>scolors</code> points to two <code>gx_color_index</code>
values; <code>scolors[0]</code> is the background color (the color
corresponding to 0-bits in the bitmap), <code>scolors[1]</code> is the
foreground color (the color corresponding to 1-bits in the bitmap).
</ul>

<p>
If <b>S</b> is a pixmap:

<ul>
<li><code>sdata</code>, <code>sourcex</code>,
<code>sraster</code>, and <code>id</code> arguments are as for
<code>copy_mono</code> or <code>copy_color</code>
(<code>data</code>, <code>data_x</code>, <code>raster</code>,
<code>id</code>), and specify a source pixmap whose depth is the same as
the depth of the destination.

<li><code>scolors</code> is <code>NULL</code>.
</ul>

<p>
Note that if the source is a bitmap with background=0 and foreground=1, and
the destination is 1 bit deep, then the source can be treated as a pixmap
(scolors=<code>NULL</code>).

<h5><a name="T_spec"></a>The texture specification T</h5>

<p>
Similar to the source, the texture <b>T</b> may be a solid color, a bitmap,
or a pixmap.  If <b>T</b> is a solid color:

<ul>
<li>The texture pointer is irrelevant.

<li><code>tcolors</code> points to two <code>gx_color_index</code>
values; <code>tcolors[0]</code> = <code>tcolors[1]</code> = the
color.
</ul>

<p>
If <b>T</b> is a bitmap:

<ul>
<li>The texture argument points to a <code>gx_tile_bitmap</code>, as for
the <code>tile_rectangle</code> procedure.  Similarly,
<code>phase_x</code> and <code>phase_y</code> specify the offset of
the texture relative to the device coordinate system origin, again as for
<code>tile_rectangle</code>.  The tile is a bitmap (1 bit per pixel).

<li><code>tcolors</code> points to two <code>gx_color_index</code>
values; <code>tcolors[0]</code> is the background color (the color
corresponding to 0-bits in the bitmap), <code>tcolors[1]</code> is the
foreground color (the color corresponding to 1-bits in the bitmap).
</ul>

<p>
If <b>T</b> is a pixmap:

<ul>
<li>The texture argument points to a <code>gx_tile_bitmap</code> whose
depth is the same as the depth of the destination.

<li>tcolors is <code>NULL</code>.
</ul>

<p>
Again, if the texture is a bitmap with background=0 and foreground=1, and
the destination depth is 1, the texture bitmap can be treated as a pixmap
(tcolors=<code>NULL</code>).

<p>
Note that while a source bitmap or pixmap has the same width and height as
the destination, a texture bitmap or pixmap has its own width and height
specified in the <code>gx_tile_bitmap</code> structure, and is
replicated or clipped as needed.

<h5><a name="F_spec"></a>The function specification f</h5>

<p>
"Command" indicates the raster operation and transparency as follows:

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=bottom>
	<th>Bits
	<td>&nbsp;
	<td>&nbsp;
<tr valign=top>	<td>7-0
	<td>&nbsp;
	<td>raster op
<tr valign=top>	<td>8
	<td>&nbsp;
	<td>0 if source opaque, 1 if source transparent
<tr valign=top>	<td>9
	<td>&nbsp;
	<td>0 if texture opaque, 1 if texture transparent
<tr valign=top>	<td>10
	<td>&nbsp;
	<td>1 if pdf transparency is in use, 0 otherwise. This makes no
difference to the rendering, but forces the raster operation to be considered
non-idempotent by internal routines.
<tr valign=top>	<td>11
	<td>&nbsp;
	<td>1 if the target of this operation is a specific plane, rather
than all planes. The plane in question is given by bits 13 upwards. This
is only used by the planar device.
<tr valign=top>	<td>12-
	<td>&nbsp;
	<td>If bit 11 = 1, then bits 12 upwards give the plane number to
operate on. Otherwise, should be set to 0.
</table></blockquote>

<p>In general most devices should just check to see that bits they do not
handle (11 and above typically) are zero, and should jump to the default
implementation, or return an error otherwise.

<p>
The raster operation follows the Microsoft and H-P specification.  It is an
8-element truth table that specifies the output value for each of the
possible 2&times;2&times;2 input values as follows:

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=bottom>
	<th>Bit
	<td>&nbsp;
	<th>Texture
	<td>&nbsp;
	<th>Source
	<td>&nbsp;
	<th>Destination
<tr>	<td colspan=7><hr>
<tr valign=top>	<td align=center>7
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>1
<tr valign=top>	<td align=center>6
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>0
<tr valign=top>	<td align=center>5
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>1
<tr valign=top>	<td align=center>4
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>0
<tr valign=top>	<td align=center>3
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>1
<tr valign=top>	<td align=center>2
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>1
	<td>&nbsp;
	<td align=center>0
<tr valign=top>	<td align=center>1
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>1
<tr valign=top>	<td align=center>0
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>0
	<td>&nbsp;
	<td align=center>0
</table></blockquote>

<p>
Transparency affects the output in the following way. A source or texture 
pixel is considered transparent if its value is all 1s (for instance, 1 for 
bitmaps, <tt>0xffffff</tt> for 24-bit RGB pixmaps) <b><em>and</em></b> the
corresponding transparency bit is set in the command. For each pixel, the
result of the Boolean operation is written into the destination iff neither
the source nor the texture pixel is transparent. (Note that the HP
RasterOp specification, on which this is based, specifies that if the
source and texture are both all 1s and the command specifies transparent
source and opaque texture, the result <b><em>should</em></b> be written in
the output. We think this is an error in the documentation.) 

<h5><a name="Compositing_notes"></a>Notes</h5>

<p>
<code>copy_rop</code> is defined to operate on pixels in RGB space,
again following the HP and Microsoft specification. For devices that
don't use RGB (or gray-scale with black = 0, white = all 1s) as their
native color representation, the implementation of <code>copy_rop</code>
must convert to RGB or gray space, do the operation, and convert back (or
do the equivalent of this). Here are the <code>copy_rop</code>
equivalents of the most important previous imaging calls. We assume the
declaration: 

<blockquote><code>
static const gx_color_index white2[2] = { 1, 1 };
</code></blockquote>

<p>
Note that <code>rop3_S</code> may be replaced by any other Boolean operation.
For monobit devices, we assume that black = 1.

<blockquote>
<pre>/* For all devices: */
(*fill_rectangle)(dev, x, y, w, h, color) ==&gt;

        { gx_color_index colors[2];
          colors[0] = colors[1] = color;
          (*dev_proc(dev, copy_rop))(dev, NULL, 0, 0, gx_no_bitmap_id, colors,
                                     NULL, colors /*irrelevant*/,
                                     x, y, w, h, 0, 0, rop3_S);
        }

/* For black-and-white devices only: */
(*copy_mono)(dev, base, sourcex, sraster, id,
             x, y, w, h, (gx_color_index)0, (gx_color_index)1) ==&gt;

        (*dev_proc(dev, copy_rop))(dev, base, sourcex, sraster, id, NULL,
                                   NULL, white2 /*irrelevant*/,
                                   x, y, w, h, 0, 0, rop3_S);

/* For color devices, where neither color0 nor color1 is gx_no_color_index: */
(*copy_mono)(dev, base, sourcex, sraster, id,
             x, y, w, h, color0, color1) ==&gt;

        { gx_color_index colors[2];
          colors[0] = color0, colors[1] = color1;
          (*dev_proc(dev, copy_rop))(dev, base, sourcex, sraster, id, colors,
                                     NULL, white2 /*irrelevant*/,
                                     x, y, w, h, 0, 0, rop3_S);
        }

/* For black-and-white devices only: */
(*copy_mono)(dev, base, sourcex, sraster, id,
             x, y, w, h, gx_no_color_index, (gx_color_index)1) ==&gt;

        (*dev_proc(dev, copy_rop))(dev, base, sourcex, sraster, id, NULL,
                                   NULL, white2 /*irrelevant*/,
                                   x, y, w, h, 0, 0,
                                   rop3_S | lop_S_transparent);

/* For all devices: */
(*copy_color)(dev, base, sourcex, sraster, id,
              x, y, w, h) ==&gt; [same as first copy_mono above]

/* For black-and-white devices only: */
(*tile_rectangle)(dev, tile, x, y, w, h,
                  (gx_color_index)0, (gx_color_index)1, px, py) ==&gt;

        (*dev_proc(dev, copy_rop))(dev, NULL, 0, 0, gx_no_bitmap_id,
                                   white2 /*irrelevant*/,
                                   tile, NULL,
                                   x, y, w, h, px, py, rop3_T)
</pre></blockquote>

<h3><a name="Polygon_level_drawing"></a>Polygon-level drawing</h3>

<p>
In addition to the pixel-level drawing operations that take integer device
coordinates and pure device colors, the driver interface includes
higher-level operations that draw polygons using fixed-point coordinates,
possibly halftoned colors, and possibly a non-default logical operation.

<p>
The <code>fill_</code>* drawing operations all use the center-of-pixel
rule: a pixel is colored iff its center falls within the polygonal region
being filled.  If a pixel center <em>(X+0.5,Y+0.5)</em> falls exactly on
the boundary, the pixel is filled iff the boundary is horizontal and the
filled region is above it, or the boundary is not horizontal and the filled
region is to the right of it.

<dl>
<dt><code>int (*fill_trapezoid)(gx_device&nbsp;*dev, const&nbsp;
gs_fixed_edge&nbsp;*left, const&nbsp;gs_fixed_edge&nbsp;*right,
fixed&nbsp;ybot, fixed&nbsp;ytop, bool&nbsp;swap_axes, 
const&nbsp;gx_drawing_color&nbsp;*pdcolor,
gs_logical_operation_t&nbsp;lop)</code> <b><em>[OPTIONAL]</em></b> 
<dd>Fill a trapezoid. The bottom and top edges are parallel to the x
axis, and are defined by <code>ybot</code> and <code>ytop</code>,
respectively.  The left and right edges are defined by <code>left</code>
and <code>right</code>.  Both of these represent lines (<code>gs_fixed_edge</code>
is defined in <a href="../base/gxdevcli.h">gxdevcli.h</a> and consists
of <code>gs_fixed_point</code> <code>start</code> and <code>end</code> points).
The y coordinates of these lines need not have any specific relation to
<code>ybot</code> and <code>ytop</code>. The routine is defined this way so
that the filling algorithm can subdivide edges and still guarantee
that the exact same pixels will be filled. If
<code>swap_axes</code> is set, the meanings of X and Y are
interchanged. 
</dd>

<dt><code>int (*fill_parallelogram)(gx_device&nbsp;*dev,
fixed&nbsp;px, fixed&nbsp;py, fixed&nbsp;ax, fixed&nbsp;ay, fixed&nbsp;bx,
fixed&nbsp;by, const&nbsp;gx_drawing_color&nbsp;*pdcolor,
gs_logical_operation_t&nbsp;lop)</code> <b><em>[OPTIONAL]</em></b>
<dd>Fill a parallelogram whose corners are <em>(px,py)</em>,
<em>(px+ax,py+ay)</em>, <em>(px+bx,py+by)</em>, and
<em>(px+ax+bx,py+ay+by)</em>.  There are no constraints on the values of
any of the parameters, so the parallelogram may have any orientation
relative to the coordinate axes.

<dt><code>int (*fill_triangle)(gx_device&nbsp;*dev, fixed&nbsp;px,
fixed&nbsp;py, fixed&nbsp;ax, fixed&nbsp;ay, fixed&nbsp;bx, fixed&nbsp;by,
const&nbsp;gx_drawing_color&nbsp;*pdcolor,
gs_logical_operation_t&nbsp;lop)</code> <b><em>[OPTIONAL]</em></b>
<dd>Fill a triangle whose corners are <em>(px,py)</em>,
<em>(px+ax,py+ay)</em>, and <em>(px+bx,py+by)</em>.

<dt><code>int (*draw_thin_line)(gx_device&nbsp;*dev,
fixed&nbsp;fx0, fixed&nbsp;fy0, fixed&nbsp;fx1, fixed&nbsp;fy1,
const&nbsp;gx_drawing_color&nbsp;*pdcolor,
gs_logical_operation_t&nbsp;lop)</code> <b><em>[OPTIONAL]</em></b>
<dd>Draw a one-pixel-wide line from <em>(fx0,fy0)</em> to
<em>(fx1,fy1)</em>.

<dt><code>int (*draw_line)(gx_device&nbsp;*dev, int&nbsp;x0, int&nbsp;y0,
int&nbsp;x1, int&nbsp;y1, gx_color_index&nbsp;color)</code>
<b><em>[OPTIONAL] [OBSOLETE]</em></b>
<dd>This procedure is no longer used: it is replaced by the draw_thin_line
procedure.  However, still appears in the driver procedure vector for
backward compatibility.  It should never be called, and drivers should not
implement it.
</dl>

<h3><a name="Linear_color_drawing"></a>Linear color drawing</h3>

<p>
Linear color functions allow fast high quality rendering of 
shadings on continuous tone devices. They implement filling simple areas
with a lineary varying color. These functions are not called if the device applies halftones,
or uses a non-separable or a non-linear color model.

<dl>
<dt><code> int (*fill_linear_color_triangle)
  (dev_t *dev, const gs_fill_attributes *fa,
	const gs_fixed_point *p0, const gs_fixed_point *p1,
	const gs_fixed_point *p2,
	const frac31 *c0, const frac31 *c1, const frac31 *c2)
</code>
<b><em>[OPTIONAL]</em></b>
<dd>This function is the highest level one within the linear color function group.
It fills a triangle with a linearly varying color.
Arguments specify 3 points in the device space - vertices of a triangle, and their colors.
The colors are represented as vectors of positive fractional numbers, each of which
represents a color component value in the interval <code>[0,1]</code>.
The number of components in a vector in the number of color
components in the device (process) color model.
<dd>
The implementation fills entire triangle.
The filling rule is same as for <a href="#Polygon_level_drawing">Polygon-level drawing</a>.
The color of each pixel within the triangle is computed as a linear interpolation
of vertex colors.
<dd>
The implementation may reject the request if the area or the color appears too complex
for filling in a single action. For doing that the implementation returns 0 and must not 
paint any pixel. In this case the graphics library will perform a subdivision of the area
into smaller triangles and call the function again with smaller areas.
<dd>
<b><em>Important note :</em></b> Do not try to decompose the area within
the implementation of <code> fill_linear_color_triangle</code>, because
it can break the plane coverage contiguity and cause a dropout.
Instead request that the graphics library should perform the decomposition.
The graphics libary is smart enough to do that properly.
<dd>
<b><em>Important note :</em></b>
The implementation must handle a special case, when only 2 colors are specified.
It happens if <code>p2</code> is <code>NULL</code>. 
This means that the color does not depend on the X coordinate,
i.e. it forms a linear gradient along the Y axis.
The implementation must not reject (return 0) such cases.
<dd>
<b><em>Important note :</em></b>The device color component
value 1 may be represented with several hexadecimal values :
<code>0x7FFF0000</code>, <code>0x7FFFF000</code>, <code>0x7FFFFF00</code>, etc.,
because the precision here exceeds the color precision of the device.
To convert a <code>frac31</code> value into a device color component value,
fist drop (ignore) the sign bit, then drop least significant bits -
so many ones as you need to fit the device color precision.
<dd>
<b><em>Important note :</em></b> The <code>fa</code> argument may contain
the <code>swap_axes</code> bit set. In this case the implementation must swap (transpose)
<code>X</code> and <code>Y</code> axes.
<dd>
<b><em>Important note :</em></b> The implementation must not paint outside the
clipping rectangle specified in the <code>fa</code> argument.
If <code>fa-&gt;swap_axes</code> is true, the clipping rectangle is transposed.
<dd>
See <code> gx_default_fill_linear_color_triangle </code>
in <code>gdevddrw.c</code> for sample code.
</dl>


<dl>
<dt><code> int (*fill_linear_color_trapezoid)
  (dev_t *dev, const gs_fill_attributes *fa,
	const gs_fixed_point *p0, const gs_fixed_point *p1,
	const gs_fixed_point *p2, const gs_fixed_point *p3,
	const frac31 *c0, const frac31 *c1,
	const frac31 *c2, const frac31 *c2)
</code>
<b><em>[OPTIONAL]</em></b>
<dd>This function is a lower level one within the linear color function group.
The default implementation of <code> fill_linear_color_triangle </code>
calls this function 1-2 times per triangle. Besides that,
this function may be called by the graphics library for other special cases,
when a decomposition into triangles appears undesirable.
<dd>
While the prototype can specify a bilinear color,
we assume that the implementation handles linear colors only.
This means that the implementation can ignore any of <code> c0, c1, c2, c3 </code>.
The graphics library takes a special care of the color linearity 
when calling this function. The reason for passing all 4 color arguments
is to avoid color precision problems.
<dd>
Similarly to <code> fill_linear_color_triangle </code>,
this function may be called with only 2 colors, and may reject areas as being too complex.
All those important notes are applicable here.
<dd>
Sample code may be found in in <code>gxdtfill.h</code>; be aware it's rather complicated.
A linear color function is generated from it as <code> gx_fill_trapezoid_ns_lc </code>
with the following template parameters :

<pre>
#define LINEAR_COLOR 1 
#define EDGE_TYPE gs_linear_color_edge
#define FILL_ATTRS const gs_fill_attributes *
#define CONTIGUOUS_FILL 0
#define SWAP_AXES 0
#define FILL_DIRECT 1
</pre>
See the helplers <code>init_gradient</code>,
<code>step_gradient</code> (defined in in <code>gdevddrw.c</code>), how to manage colors.
See <code>check_gradient_overflow</code>
(defined in in <code>gdevddrw.c</code>), as an example of an area
that can't be painted in a single action due to 64-bits fixed overflows.

</dl>

<dl>
<dt><code> int (*fill_linear_color_scanline)
      (dev_t *dev, const gs_fill_attributes *fa,
	int i, int j, int w,
	const frac31 *c0,
	const int32_t *c0_f,
	const int32_t *cg_num,
	int32_t cg_den)
</code>
<b><em>[OPTIONAL]</em></b>
<dd>This function is the lowest level one within the linear color function group.
It implements filling a scanline with a linearly varying color.
The default implementation for <code> fill_linear_color_trapezoid </code>
calls this function, and there are no other calls to it from the graphics libary.
Thus if the device implements <code> fill_linear_color_triangle </code> and
<code> fill_linear_color_trapezoid </code> by own means,
this function may be left unimplemented.
<dd>
<code>i</code> and <code>j</code> specify device coordinates (indices)
of the starting pixel of the scanline, <code>w</code> specifies the 
width of the scanline, i.e. the number of pixels to be painted to the right from
the starting pixel, including the starting pixel.
<dd>
<code>c0</code> specifies the color for the starting pixel
as a vector of fraction values, each of which represents 
a color value in the interval <code>[0,1]</code>.
<dd>
<code>c0_f</code> specify a fraction part of the color for the starting pixel.
See the formula below about using it.
<dd>
<code>cg_num</code> specify a numerator for the color gradient -
a vector of values in <code>[-1,1]</code>, each of which correspond to a color component.
<dd>
<code>cg_den</code> specify the denominator for the color gradient -
a value in <code>[-1,1]</code>.
<dd><p>
The color for the pixel <code>[i + k, j]</code> to be computed like this :
<pre><code>
         (double)(c0[n] + (c0_f[n] + cg_num[n] * k) / cg_den) / (1 ^ 31 - 1)
</code></pre>
<dd>where <code>0 &lt;= k &lt;= w </code>, and <code>n</code> is a device color component index.

<dd>
<b><em>Important note :</em></b> The <code>fa</code> argument may contain
the <code>swap_axes</code> bit set. In this case the implementation must swap (transpose)
<code>X</code> and <code>Y</code> axes.
<dd>
<b><em>Important note :</em></b> The implementation must not paint outside the
clipping rectangle specified in the <code>fa</code> argument.
If <code>fa-&gt;swap_axes</code> is true, the clipping rectangle is transposed.
<dd>
See <code> gx_default_fill_linear_color_scanline</code>
in <code>gdevdsha.c</code> as a sample code.
</dl>


<h3><a name="High_level_drawing"></a>High-level drawing</h3>

<p>
In addition to the lower-level drawing operations described above, the
driver interface provides a set of high-level operations.  Normally these
will have their default implementation, which converts the high-level
operation to the low-level ones just described; however, drivers that
generate high-level output formats such as pdfwrite, or communicate with devices
that have firmware for higher-level operations such as polygon fills, may
implement these high-level operations directly.  For more details, please
consult the source code, specifically:

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=top>	<th align=left>Header
	<td>&nbsp;&nbsp;&nbsp;
	<th align=left>Defines
<tr valign=top>	<td><a href="../base/gxpaint.h">gxpaint.h</a>
	<td>&nbsp;
	<td><code>gx_fill_params</code>, <code>gx_stroke_params</code>
<tr valign=top>	<td><a href="../base/gxfixed.h">gxfixed.h</a>
	<td>&nbsp;
	<td><code>fixed</code>, <code>gs_fixed_point</code> (used by
	    <code>gx_*_params</code>)
<tr valign=top>	<td><a href="../base/gxistate.h">gxistate.h</a>
	<td>&nbsp;
	<td><code>gs_imager_state</code> (used by <code>gx_*_params</code>)
<tr valign=top>	<td><a href="../base/gxline.h">gxline.h</a>
	<td>&nbsp;
	<td><code>gx_line_params</code> (used by <code>gs_imager_state</code>)
<tr valign=top>	<td><a href="../base/gslparam.h">gslparam.h</a>
	<td>&nbsp;
	<td>line cap/join values (used by <code>gx_line_params</code>)
<tr valign=top>	<td><a href="../base/gxmatrix.h">gxmatrix.h</a>
	<td>&nbsp;
	<td><code>gs_matrix_fixed</code> (used by <code>gs_imager_state</code>)
<tr valign=top>	<td><a href="../base/gspath.h">gspath.h</a>, <a href="../base/gxpath.h">gxpath.h</a>, <a href="../base/gzpath.h">gzpath.h</a>
	<td>&nbsp;
	<td><code>gx_path</code>
<tr valign=top>	<td><a href="../base/gxcpath.h">gxcpath.h</a>, <a href="../base/gzcpath.h">gzcpath.h</a>
	<td>&nbsp;
	<td><code>gx_clip_path</code>
</table></blockquote>

<p>
For a minimal example of how to implement the high-level drawing operations,
see <a href="../base/gdevtrac.c">gdevtrac.c</a>.

<h4><a name="Paths"></a>Paths</h4>

<dl>
<dt><code>int (*fill_path)(gx_device&nbsp;*dev,
const&nbsp;gs_imager_state&nbsp;*pis, gx_path&nbsp;*ppath,
const&nbsp;gx_fill_params&nbsp;*params,
const&nbsp;gx_drawing_color&nbsp;*pdcolor,
const&nbsp;gx_clip_path&nbsp;*pcpath)</code> <b><em>[OPTIONAL]</em></b>
<dd>Fill the given path, clipped by the given clip path, according to the
given parameters, with the given color.  The clip path pointer may be
<code>NULL</code>, meaning do not clip.
<dd>
The implementation must paint the path with the specified device color,
which may be either a pure color, or a pattern. If the device can't
handle non-pure colors, it should check the color type and
call the default implementation gx_default_fill_path for cases
which it can't handle. The default implementation will perform
a subdivision of the area to be painted, and will
call other device virtual functions (such as fill_linear_color_triangle)
with simpler areas.

</dl>

<dl>
<dt><code>int (*stroke_path)(gx_device&nbsp;*dev,
const&nbsp;gs_imager_state&nbsp;*pis, gx_path&nbsp;*ppath,
const&nbsp;gx_stroke_params&nbsp;*params,
const&nbsp;gx_drawing_color&nbsp;*pdcolor,
const&nbsp;gx_clip_path&nbsp;*pcpath)</code> <b><em>[OPTIONAL]</em></b>
<dd>Stroke the given path, clipped by the given clip path, according to the
given parameters, with the given color.  The clip path pointer may be
<code>NULL</code>, meaning not to clip.
</dl>

<dl>
<dt><code>int (*fill_mask)(gx_device&nbsp;*dev,
const&nbsp;byte&nbsp;*data, int&nbsp;data_x, int&nbsp;raster,
gx_bitmap_id&nbsp;id, int&nbsp;x, int&nbsp;y, int&nbsp;width,
int&nbsp;height, const&nbsp;gx_drawing_color&nbsp;*pdcolor, int&nbsp;depth,
int&nbsp;command, const&nbsp;gx_clip_path&nbsp;*pcpath)</code>
<b><em>[OPTIONAL]</em></b>
<dd>Color the 1-bits in the given mask (or according to the alpha values,
if <code>depth</code>&nbsp;&gt;&nbsp;1), clipped by the given clip path,
with the given color and logical operation.  The clip path pointer may be
<code>NULL</code>, meaning do not clip.  The parameters
<code>data</code>, ..., <code>height</code> are as for
<code>copy_mono</code>; depth is as for <code>copy_alpha</code>;
command is as for <code>copy_rop</code>.
</dl>

<h4><a name="Images"></a>Images</h4>

<p>
Similar to the high-level interface for fill and stroke graphics, a high-level
interface exists for bitmap images.  The procedures in this part of the
interface are optional.

<p>
Bitmap images come in a variety of types, corresponding closely (but not
precisely) to the PostScript ImageTypes.  The generic or common part of all
bitmap images is defined by:

<blockquote>
<pre>typedef struct {
	const gx_image_type_t *type;
        gs_matrix ImageMatrix;
} gs_image_common_t;</pre>
</blockquote>

<p>
Bitmap images that supply data (all image types except
<code>image_type_from_device</code> (2)) are defined by:

<blockquote>
<pre>#define gs_image_max_components 5
typedef struct {
        &lt;&lt; gs_image_common_t &gt;&gt;
        int Width;
        int Height;
        int BitsPerComponent;
        float Decode[gs_image_max_components * 2];
        bool Interpolate;
} gs_data_image_t;</pre>
</blockquote>

<p>
Images that supply pixel (as opposed to mask) data are defined by:

<blockquote>
<pre>typedef enum {
	/* Single plane, chunky pixels. */
	gs_image_format_chunky = 0,
	/* num_components planes, chunky components. */
	gs_image_format_component_planar = 1,
	/* BitsPerComponent * num_components planes, 1 bit per plane */
	gs_image_format_bit_planar = 2
} gs_image_format_t;
typedef struct {
        &lt;&lt; gs_data_image_t &gt;&gt;
        const gs_color_space *ColorSpace;
        bool CombineWithColor;
} gs_pixel_image_t;</pre>
</blockquote>

<p>
Ordinary PostScript Level 1 or Level 2 (<code>ImageType</code> 1) images
are defined by:

<blockquote>
<pre>typedef enum {
	/* No alpha. */
	gs_image_alpha_none = 0,
	/* Alpha precedes color components. */
	gs_image_alpha_first,
	/* Alpha follows color components. */
	gs_image_alpha_last
} gs_image_alpha_t;
typedef struct {
        &lt;&lt; gs_pixel_image_t &gt;&gt;
        bool ImageMask;
        bool adjust;
	gs_image_alpha_t Alpha;
} gs_image1_t;
typedef gs_image1_t gs_image_t;</pre>
</blockquote>

<p>
Of course, standard PostScript images don't have an alpha component.  For
more details, consult the source code in <a
href="../base/gsiparam.h">gsiparam.h</a> and <code>gsiparm*.h</code>,
which define parameters for an image.

<p>
The <code>begin[_typed_]image</code> driver procedures create image
enumeration structures.  The common part of these structures consists of:

<blockquote>
<pre>typedef struct gx_image_enum_common_s {
        const gx_image_type_t *image_type;
	const gx_image_enum_procs_t *procs;
	gx_device *dev;
	gs_id id;
        int num_planes;
        int plane_depths[gs_image_max_planes];  /* [num_planes] */
	int plane_widths[gs_image_max_planes]	/* [num_planes] */
} gx_image_enum_common_t;</pre>
</blockquote>

<p>
where <code>procs</code> consists of:

<blockquote>
<pre>typedef struct gx_image_enum_procs_s {

        /*
         * Pass the next batch of data for processing.
         */
#define image_enum_proc_plane_data(proc)\
  int proc(gx_device *dev,\
    gx_image_enum_common_t *info, const gx_image_plane_t *planes,\
    int height)

        image_enum_proc_plane_data((*plane_data));

        /*
         * End processing an image, freeing the enumerator.
         */
#define image_enum_proc_end_image(proc)\
  int proc(gx_device *dev,\
    gx_image_enum_common_t *info, bool draw_last)

        image_enum_proc_end_image((*end_image));

	/*
	 * Flush any intermediate buffers to the target device.
	 * We need this for situations where two images interact
	 * (currently, only the mask and the data of ImageType 3).
	 * This procedure is optional (may be 0).
	 */
#define image_enum_proc_flush(proc)\
  int proc(gx_image_enum_common_t *info)

	image_enum_proc_flush((*flush));

} gx_image_enum_procs_t;</pre>
</blockquote>

<p> In other words, <code>begin[_typed]_image</code> sets up an
enumeration structure that contains the procedures that will process the
image data, together with all variables needed to maintain the state of the
process.  Since this is somewhat tricky to get right, if you plan to create
one of your own you should probably read an existing implementation of
<code>begin[_typed]_image</code>, such as the one in <a
href="../base/gdevbbox.c">gdevbbox.c</a> or <a
href="../base/gdevps.c">gdevps.c</a>.

<p>
The data passed at each call of <code>image_plane_data</code> consists of
one or more planes, as appropriate for the type of image.
<code>begin[_typed]_image</code> must initialize the
<code>plane_depths</code> array in the enumeration structure with the
depths (bits per element) of the planes.  The array of
<code>gx_image_plane_t</code> structures passed to each call of
<code>image_plane_data</code> then defines where the data are stored, as
follows:

<blockquote>
<pre>typedef struct gx_image_plane_s {
  const byte *data;
  int data_x;
  uint raster;
} gx_image_plane_t;</pre>
</blockquote>

<dl>
<dt><code>int (*begin_image)(gx_device&nbsp;*dev,
const&nbsp;gs_imager_state&nbsp;*pis, const&nbsp;gs_image_t&nbsp;*pim,
gs_image_format_t&nbsp;format, gs_int_rect&nbsp;*prect,
const&nbsp;gx_drawing_color&nbsp;*pdcolor,
const&nbsp;gx_clip_path&nbsp;*pcpath, gs_memory_t&nbsp;*memory,
gx_image_enum_common_t&nbsp;**pinfo)</code> <b><em>[OPTIONAL]</em></b>
<dd>Begin the transmission of an image.  Zero or more calls of
<code>image_plane_data</code> will follow, and then a call of
<code>end_image</code>.  The parameters of <code>begin_image</code>
are as follows:

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=top>	<td><code>pis</code>
	<td>&nbsp;&nbsp;&nbsp;
	<td>pointer to an imager state.  The only relevant elements of the
	    imager state are the CTM (coordinate transformation matrix),
	    the logical operation (<code>RasterOp</code> or
	    transparency), and the color rendering information.
<tr valign=top>	<td><code>pim</code>
	<td>&nbsp;
	<td>pointer to the <code>gs_image_t</code> structure that
	    defines the image parameters
<tr valign=top>	<td><code>format</code>
	<td>&nbsp;
	<td>defines how pixels are represented for
	    <code>image_plane_data</code>.  See the description of
	    <code>image_plane_data</code> below
<tr valign=top>	<td><code>prect</code>
	<td>&nbsp;
	<td>if not <code>NULL</code>, defines a subrectangle of the
	    image; only the data for this subrectangle will be passed to
	    <code>image_plane_data</code>, and only this subrectangle should
	    be drawn
<tr valign=top>	<td><code>pdcolor</code>
	<td>&nbsp;
	<td>defines a drawing color, only needed for masks or if
	    <code>CombineWithColor</code> is true
<tr valign=top>	<td><code>pcpath</code>
	<td>&nbsp;
	<td>if not <code>NULL</code>, defines an optional clipping path
<tr valign=top>	<td><code>memory</code>
	<td>&nbsp;
	<td>defines the allocator to be used for allocating bookkeeping
	    information
<tr valign=top>	<td><code>pinfo</code>
	<td>&nbsp;
	<td>the implementation should return a pointer to its state
	    structure here
</table></blockquote>

<p>
<code>begin_image</code> is expected to allocate a structure for its
bookkeeping needs, using the allocator defined by the memory parameter, and
return it in <code>*pinfo</code>.  <code>begin_image</code> should not assume that
the structures in <code>*pim</code>, <code>*prect</code>, or
<code>*pdcolor</code> will survive the call on
<code>begin_image</code> (except for the color space in
<code>*pim-&gt;ColorSpace</code>): it should copy any necessary parts of
them into its own bookkeeping structure.  It may, however, assume that
<code>*pis</code>, <code>*pcpath</code>, and of course
<code>*memory</code> will live at least until <code>end_image</code>
is called.

<p>
<code>begin_image</code> returns 0 normally, or 1 if the image does not
need any data.  In the latter case, <code>begin_image</code> does not
allocate an enumeration structure.
</dl>

<dl>
<dt><code>int (*begin_typed_image)(gx_device&nbsp;*dev,
const&nbsp;gs_imager_state&nbsp;*pis, const&nbsp;gs_matrix&nbsp;*pmat,
const&nbsp;gs_image_common_t&nbsp;*pim, gs_int_rect&nbsp;*prect,
const&nbsp;gx_drawing_color&nbsp;*pdcolor,
const&nbsp;gx_clip_path&nbsp;*pcpath, gs_memory_t&nbsp;*memory,
gx_image_enum_common_t&nbsp;**pinfo)</code> <b><em>[OPTIONAL]</em></b>
<dd>This has the same function as <code>begin_image</code>, except
<ul>
<li>The image may be of any <code>ImageType</code>, not only
<code>image_type_simple</code> (1);

<li>The image format is included in the image structure, not supplied as a
separate argument;

<li>The optional <code>pmat</code> argument provides a matrix that
substitutes for the one in the imager state;

<li>For mask images, if <code>pmat</code> is not <code>NULL</code>
and the color is pure, <code>pis</code> may be <code>NULL</code>.
</ul>
</dl>

<p>
The actual transmission of data uses the procedures in the enumeration
structure, not driver procedures, since the handling of the data usually
depends on the image type and parameters rather than the device.  These
procedures are specified as follows.

<dl>
<dt><code>int (*image_plane_data)(gx_device&nbsp;*dev,
gx_image_enum_common_t&nbsp;*info,
const&nbsp;gx_image_plane_t&nbsp;*planes, int&nbsp;height)</code>
<dd>This call provides more of the image source data: specifically,
<code>height</code> rows, with <code>Width</code> pixels supplied for
each row.

<p>
The data for each row are packed big-endian within each byte, as for
<code>copy_color</code>.  The <code>data_x</code> (starting X position
within the row) and <code>raster</code> (number of bytes per row) are
specified separately for each plane, and may include some padding at the
beginning or end of each row.  Note that for non-mask images, the input data
may be in any color space and may have any number of bits per component (1,
2, 4, 8, 12); currently mask images always have 1 bit per component, but in
the future, they might allow multiple bits of alpha.  Note also that each
call of <code>image_plane_data</code> passes complete pixels: for example, for
a chunky image with 24 bits per pixel, each call of
<code>image_plane_data</code> passes 3N bytes of data (specifically,
3&nbsp;&times;&nbsp;Width&nbsp;&times;&nbsp;height).

<p>
The interpretation of planes depends on the <code>format</code> member of
the <code>gs_image[_common]_t</code> structure:

<ul>
<li>If the format is <code>gs_image_format_chunky</code>,
<code>planes[0].data</code> points to data in "chunky" format, in which
the components follow each other (for instance, RGBRGBRGB....)

<li>If the format is <code>gs_image_format_component_planar</code>,
<code>planes[0&nbsp;..&nbsp;N-1].data</code> point to data for the
<b><em>N</em></b> components (for example, <b><em>N</em></b>=3 for RGB
data); each plane contains samples for a single component, for instance,
RR..., GG..., BB....  Note that the planes are divided by component, not by
bit: for example, for 24-bit RGB data, <b><em>N</em></b>=3, with 8-bit
values in each plane of data.

<li>If the format is <code>gs_image_format_bit_planar</code>,
<code>planes[0&nbsp;..&nbsp;N*B-1].data</code> point to data for the
<b><em>N</em></b> components of <b><em>B</em></b> bits each (for example,
<b><em>N</em></b>=3 and <b><em>B</em></b>=4 for RGB data with 4 bits per
component); each plane contains samples for a single bit, for instance, R0
R1 R2 R3 G0 G1 G2 G3 B0 B1 B2 B3.  Note that the most significant bit of
each plane comes first.
</ul>

<p>
If, as a result of this call, <code>image_plane_data</code> has been called with all
the data for the (sub-)image, it returns 1; otherwise, it returns 0 or an
error code as usual.

<p>
<code>image_plane_data</code>, unlike most other procedures that take bitmaps as
arguments, does not require the data to be aligned in any way.

<p>
Note that for some image types, different planes may have different
numbers of bits per pixel, as defined in the <code>plane_depths</code> array.
</dl>

<dl>
<dt><code>int (*end_image)(gx_device&nbsp;*dev, void&nbsp;*info,
bool&nbsp;draw_last)</code>
<dd>Finish processing an image, either because all data have been supplied
or because the caller has decided to abandon this image.
<code>end_image</code> may be called at any time after
<code>begin_image</code>.  It should free the info structure and any
subsidiary structures.  If <code>draw_last</code> is true, it should
finish drawing any buffered lines of the image.
</dl>

<h5><a name="Images_notes"></a>Notes</h5>

<p>
While there will almost never be more than one image enumeration in
progress -- that is, after a <code>begin_image</code>,
<code>end_image</code> will almost always be called before the next
<code>begin_image</code> -- driver code should not rely on this
property; in particular, it should store all information regarding the
image in the info structure, not in the driver structure.

<p>
Note that if <code>begin_[typed_]image</code> saves its parameters in
the info structure, it can decide on each call whether to use its own
algorithms or to use the default implementation.  (It may need to call
<code>gx_default_begin</code>/<code>end_image</code> partway
through.)  [A later revision of this document may include an example here.]

<h4><a name="Text"></a>Text</h4>

<p>
The third high-level interface handles text.  As for images, the interface
is based on creating an enumerator which then may execute the operation in
multiple steps.  As for the other high-level interfaces, the procedures are
optional.

<dl>
<dt><code>int (*text_begin)(gx_device&nbsp;*dev,
gs_imager_state&nbsp;*pis, const&nbsp;gs_text_params_t&nbsp;*text,
gs_font&nbsp;*font, gx_path&nbsp;*path,
const&nbsp;gx_device_color&nbsp;*pdcolor,
const&nbsp;gx_clip_path&nbsp;*pcpath, gs_memory_t&nbsp;*memory,
gs_text_enum_t&nbsp;**ppte)</code> <b><em>[OPTIONAL]</em></b>

<dd>
Begin processing text, by creating a state structure and storing it in
<code>*ppte</code>.  The parameters of <code>text_begin</code> are as
follows:
</dl>

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=top>	<td><code>dev</code>
	<td>&nbsp;&nbsp;&nbsp;
	<td>The usual pointer to the device.
<tr valign=top>	<td><code>pis</code>
	<td>&nbsp;&nbsp;&nbsp;
	<td>A pointer to an imager state.  All elements may be relevant,
	    depending on how the text is rendered.	
<tr valign=top>	<td><code>text</code>
	<td>&nbsp;
	<td>A pointer to the structure that defines the text operation
	    and parameters.  See <a href="../base/gstext.h">gstext.h</a> for details.
<tr valign=top>	<td><code>font</code>
	<td>&nbsp;
	<td>Defines the font for drawing.
<tr valign=top>	<td><code>path</code>
	<td>&nbsp;
	<td>Defines the path where the character outline will be appended
	    (if the text operation includes <code>TEXT_DO_...PATH</code>),
	    and whose current point indicates where drawing should occur
	    and will be updated by the string width (unless the text
	    operation includes <code>TEXT_DO_NONE</code>).
<tr valign=top>	<td><code>pdcolor</code>
	<td>&nbsp;
	<td>Defines the drawing color for the text.  Only relevant if
	    the text operation includes <code>TEXT_DO_DRAW</code>.
<tr valign=top>	<td><code>pcpath</code>
	<td>&nbsp;
	<td>If not <code>NULL</code>, defines an optional clipping path.
	    Only relevant if the text operation includes
	    <code>TEXT_DO_DRAW</code>.
<tr valign=top>	<td><code>memory</code>
	<td>&nbsp;
	<td>Defines the allocator to be used for allocating bookkeeping
	    information.
<tr valign=top>	<td><code>ppte</code>
	<td>&nbsp;
	<td>The implementation should return a pointer to its state
	    structure here.
</table></blockquote>

<p>
<code>text_begin</code> must allocate a structure for its bookkeeping
needs, using the allocator defined by the <code>memory</code> parameter,
and return it in <code>*ppte</code>.  <code>text_begin</code> may
assume that the structures passed as parameters will survive until text
processing is complete.

<p>
Clients should not call the driver <code>text_begin</code> procedure
directly.  Instead, they should call <code>gx_device_text_begin</code>,
which takes the same parameters and also initializes certain common elements
of the text enumeration structure, or <code>gs_text_begin</code>, which
takes many of the parameters from a graphics state structure.  For details,
see <a href="../base/gstext.h">gstext.h</a>.

<p>
The actual processing of text uses the procedures in the enumeration
structure, not driver procedures, since the handling of the text may depend
on the font and parameters rather than the device.  Text processing may also
require the client to take action between characters, either because the
client requested it (<code>TEXT_INTERVENE</code> in the operation) or
because rendering a character requires suspending text processing to call an
external package such as the PostScript interpreter.  (It is a deliberate
design decision to handle this by returning to the client, rather than
calling out of the text renderer, in order to avoid potentially unknown
stack requirements.)  Specifically, the client must call the following
procedures, which in turn call the procedures in the text enumerator.

<dl>
<dt><code>int gs_text_process(gs_text_enum_t&nbsp;*pte)</code>
<dd>Continue processing text.  This procedure may return 0 or a negative
error code as usual, or one of the following values (see
<a href="../base/gstext.h">gstext.h</a> for details).

<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=top>	<td><code>TEXT_PROCESS_RENDER</code>
	<td>The client must cause the current character to be rendered.
	    This currently only is used for PostScript Type 0-4 fonts
	    and their CID-keyed relatives.
<tr valign=top>	<td><code>TEXT_PROCESS_INTERVENE</code>
	<td>The client has asked to intervene between characters.
	    This is used for <code>cshow</code> and <code>kshow</code>.
</table></blockquote>
</dl>

<dl>
<dt><code>int gs_text_release(gs_text_enum_t&nbsp;*pte,
client_name_t&nbsp;cname)</code> <dd>Finish processing text and release
all associated structures.  Clients must call this procedure after
<code>gs_text_process</code> returns 0 or an error, and may call it at
any time.
</dl>

<p>
There are numerous other procedures that clients may call during text
processing.  See <a href="../base/gstext.h">gstext.h</a> for details.

<h5><a name="Text_notes"></a>Notes</h5>

<p>
Note that unlike many other optional procedures, the default implementation
of <code>text_begin</code> cannot simply return: like the default
implementation of <code>begin[_typed]_image</code>, it must create and
return an enumerator.  Furthermore, the implementation of the
<code>process</code> procedure (in the enumerator structure, called by
<code>gs_text_process</code>) cannot simply return without doing
anything, even if it doesn't want to draw anything on the output.  See the
comments in <a href="../base/gxtext.h">gxtext.h</a> for details.

<h4><a name="Unicode"></a>Unicode support for high level devices</h4>

<p>
<p>Implementing a new high level device, one may need to translate <code>Postscript</code>
character codes into <code>Unicode</code>. This can be done pretty simply.

<p>For translating a <code>Postscript</code> text you need to inplement the device
virtual function <code>text_begin</code>. It should create a new instance of
<code>gs_text_enum_t</code> in the heap (let its pointer be <code>pte</code>),
and assign a special function to <code>gs_text_enum_t::procs.process</code>.
The function will receive <code>pte</code>. It should take the top level font from
<code>pte-&gt;orig_font</code>,
and iterate with <code>font-&gt;procs.next_char_glyph(pte, ..., &amp;glyph)</code>.
The last argument receives a <code>gs_glyph</code> value, which encodes a
<code>Postscript</code> character name or CID (and also stores it into
<code>pte-&gt;returned.current_glyph</code>).
Then obtain the current subfont with <code>gs_text_current_font(pte)</code>
(it can differ from the font)
and call <code>subfont-&gt;procs.decode_glyph(subfont, glyph)</code>.
The return value will be an <code>Unicode</code> code, or <code>GS_NO_CHAR</code>
if the glyph can't be translated to Unicode.

<h3><a name="Reading_bits_back"></a>Reading bits back</h3>

<dl>
<dt><code>int (*get_bits_rectangle)(gx_device&nbsp;*dev,
const&nbsp;gs_int_rect&nbsp;*prect, gs_get_bits_params_t&nbsp;*params,
gs_int_rect&nbsp;**unread)</code> <b><em>[OPTIONAL]</em></b>

<dd>
Read a rectangle of bits back from the device.  The <code>params</code>
structure consists of:

<table cellpadding=0 cellspacing=0>
<tr valign=top>	<td><code>options</code>
	<td>&nbsp;
	<td>the allowable formats for returning the data
<tr valign=top>	<td><code>data[32]</code>
	<td>&nbsp;
	<td>pointers to the returned data
<tr valign=top>	<td><code>x_offset</code>
	<td>&nbsp;
	<td>the X offset of the first returned pixel in data
<tr valign=top>	<td><code>raster</code>
	<td>&nbsp;
	<td>the distance between scan lines in the returned data
</table>

<p>
<code>options</code> is a bit mask specifying what formats the client is
willing to accept.  (If the client has more flexibility, the implementation
may be able to return the data more efficiently, by avoiding representation
conversions.)  The options are divided into groups.

<blockquote><dl>
<dt><b><em>alignment</em></b>
<dd>Specifies whether the returned data must be aligned in the normal
manner for bitmaps, or whether unaligned data are acceptable.

<dt><b><em>pointer or copy</em></b>
<dd>Specifies whether the data may be copied into storage provided by the
client and/or returned as pointers to existing storage. (Note that if
copying is not allowed, it is much more likely that the implementation will
return an error, since this requires that the client accept the data in the
implementation's internal format.)

<dt><b><em>X offset</em></b>
<dd>Specifies whether the returned data must have a specific X offset
(usually zero, but possibly other values to avoid skew at some later stage
of processing) or whether it may have any X offset (which may avoid skew in
the <code>get_bits_rectangle</code> operation itself).

<dt><b><em>raster</em></b>
<dd>Specifies whether the raster (distance between returned scan lines)
must have its standard value, must have some other specific value, or may
have any value.  The standard value for the raster is the device width
padded out to the alignment modulus when using pointers, or the minimum
raster to accommodate the X offset + width when copying (padded out to the
alignment modulus if standard alignment is required).

<dt><b><em>format</em></b>
<dd>Specifies whether the data are returned in chunky (all components of a
single pixel together), component-planar (each component has its own scan
lines), or bit-planar (each bit has its own scan lines) format.

<dt><b><em>color space</em></b>
<dd>Specifies whether the data are returned as native device pixels, or in
a standard color space.  Currently the only supported standard space is
RGB.

<dt><b><em>standard component depth</em></b>
<dd>Specifies the number of bits per component if the data are returned in
the standard color space.  (Native device pixels use
<code>dev</code>-&gt;<code>color_info.depth</code> bits per pixel.)

<dt><b><em>alpha</em></b>
<dd>Specifies whether alpha channel information should be returned as the
first component, the last component, or not at all.  Note that for devices
that have no alpha capability, the returned alpha values will be all 1s.
</dl></blockquote>

<p>
The client may set more than one option in each of the above groups; the
implementation will choose one of the selected options in each group to
determine the actual form of the returned data, and will update
<code>params[].options</code> to indicate the form.  The returned
<code>params[].options</code> will normally have only one option set per
group.

<p>
For further details on <code>params</code>, see <a
href="../base/gxgetbit.h">gxgetbit.h</a>.  For further details on
<code>options</code>, see <a href="../base/gxbitfmt.h">gxbitfmt.h</a>.

<p>
Define w = <code>prect</code>-&gt;q.x - <code>prect</code>-&gt;p.x, h
= <code>prect</code>-&gt;q.y - <code>prect</code>-&gt;p.y.  If the
bits cannot be read back (for example, from a printer), return
<code>gs_error_unknownerror</code>; if raster bytes is not enough space
to hold <code>offset_x</code> + w pixels, or if the source rectangle
goes outside the device dimensions (p.x &lt; 0 || p.y &lt; 0 || q.x &gt;
<code>dev</code>-&gt;width || q.y &gt; <code>dev</code>-&gt;height),
return <code>gs_error_rangecheck</code>; if any regions could not be
read, return <code>gs_error_ioerror</code> if unpainted is
<code>NULL</code>, otherwise the number of rectangles (see below);
otherwise return 0.

<p>
The caller supplies a buffer of <code>raster</code>&nbsp;&times;&nbsp;h
bytes starting at <code>data[0]</code> for the returned data in chunky
format, or <b><em>N</em></b> buffers of
<code>raster</code>&nbsp;&times;&nbsp;h bytes starting at
<code>data[0]</code> through
<code>data[</code><b><em>N-1</em></b><code>]</code> in planar format
where <b><em>N</em></b> is the number of components or bits.  The contents
of the bits beyond the last valid bit in each scan line (as defined by w)
are unpredictable.  data need not be aligned in any way.  If
<code>x_offset</code> is non-zero, the bits before the first valid bit
in each scan line are undefined.  If the implementation returns pointers to
the data, it stores them into <code>data[0]</code> or
<code>data[</code><b><em>0..N-1</em></b><code>]</code>.

<p>
If not all the source data are available (for example, because the source
was a partially obscured window and backing store was not available or not
used), or if the rectangle does not fall completely within the device's
coordinate system, any unread bits are undefined, and the value returned
depends on whether unread is <code>NULL</code>.  If unread is
<code>NULL</code>, return <code>gs_error_ioerror</code>; in this case,
some bits may or may not have been read.  If unread is not
<code>NULL</code>, allocate (using <code>dev</code>-&gt;memory) and
fill in a list of rectangles that could not be read, store the pointer to
the list in <code>*unread</code>, and return the number of rectangles; in
this case, all bits not listed in the rectangle list have been read back
properly.  The list is not sorted in any particular order, but the
rectangles do not overlap.  Note that the rectangle list may cover a
superset of the region actually obscured: for example, a lazy implementation
could return a single rectangle that was the bounding box of the region.
</dl>

<dl>
<dt><code>int (*get_bits)(gx_device&nbsp;*dev, int&nbsp;y,
byte&nbsp;*data, byte&nbsp;**actual_data)</code>
<b><em>[OPTIONAL]</em></b>
<dd>Read scan line <code>y</code> of bits back from the device into the
area starting at data.  This call is functionally equivalent to

<blockquote>
<pre>(*get_bits_rectangle)
  (dev, {0, y, dev-&gt;width, y+1},
   {(GB_ALIGN_ANY | (GB_RETURN_COPY | GB_RETURN_POINTER) | GB_OFFSET_0 |
     GB_RASTER_STANDARD | GB_FORMAT_CHUNKY | GB_COLORS_NATIVE |
     GB_ALPHA_NONE),
    {data}})</pre></blockquote>

<p>
with the returned value of
<code>params</code>-&gt;<code>data[0]</code> stored in
<code>*actual_data</code>, and will in fact be implemented this way if
the device defines a <code>get_bits_rectangle</code> procedure and does
not define one for <code>get_bits</code>.  (If
<code>actual_data</code> is <code>NULL</code>,
<code>GB_RETURN_POINTER</code> is omitted from the options.)
</dl>

<h3><a name="Parameters"></a>Parameters</h3>

<p>
Devices may have an open-ended set of parameters, which are simply pairs
consisting of a name and a value.  The value may be of various types:
integer (int or long), boolean, float, string, name, <code>NULL</code>,
array of integer, array of float, or arrays or dictionaries of mixed types.
For example, the <code>Name</code> of a device is a string; the
<code>Margins</code> of a device is an array of two floats.  See
<a href="../base/gsparam.h">gsparam.h</a> for more details.

<p>
If a device has parameters other than the ones applicable to all devices
(or, in the case of printer devices, all printer devices), it must provide
<code>get_params</code> and <code>put_params</code> procedures.  If
your device has parameters beyond those of a straightforward display or
printer, we strongly advise using the <code>_get_params</code> and
<code>_put_params</code> procedures in an existing device (for example,
<a href="../base/gdevcdj.c">gdevcdj.c</a> or <a
href="../base/gdevbit.c">gdevbit.c</a>) as a model for your own code.

<dl>
<dt><code>int (*get_params)(gx_device&nbsp;*dev,
gs_param_list&nbsp;*plist)</code> <b><em>[OPTIONAL]</em></b>
<dd>Read the parameters of the device into the parameter list at
<code>plist</code>, using the <code>param_write_*</code>
macros or procedures defined in <a href="../base/gsparam.h">gsparam.h</a>.
</dl>

<dl>
<dt><code>int (*get_hardware_params)(gx_device&nbsp;*dev,
gs_param_list&nbsp;*plist)</code> <b><em>[OPTIONAL]</em></b>
<dd>Read the hardware-related parameters of the device into the parameter
list at plist.  These are any parameters whose values are under control of
external forces rather than the program -- for example, front panel
switches, paper jam or tray empty sensors, etc.  If a parameter involves
significant delay or hardware action, the driver should only determine the
value of the parameter if it is "requested" by the
<code>gs_param_list</code> [<code>param_requested</code>(plist,
<code>key_name</code>)].  This function may cause the asynchronous
rendering pipeline (if enabled) to be drained, so it should be used
sparingly.
</dl>

<dl>
<dt><code>int (*put_params)(gx_device&nbsp;*dev,
gs_param_list&nbsp;*plist)</code> <b><em>[OPTIONAL]</em></b>
<dd>Set the parameters of the device from the parameter list at
<code>plist</code>, using the <code>param_read_</code>*
macros/procedures defined in <a href="../base/gsparam.h">gsparam.h</a>.  All
<code>put_params</code> procedures must use a "two-phase commit"
algorithm; see <a href="../base/gsparam.h">gsparam.h</a> for details.
</dl>

<h4><a name="Default_CRD_parameters"></a>Default color rendering
dictionary (CRD) parameters</h4>

<p>
Drivers that want to provide one or more default CIE color rendering
dictionaries (CRDs) can do so through <code>get_params</code>.  To do
this, they create the CRD in the usual way (normally using the
<code>gs_cie_render1_build</code> and <code>_initialize</code>
procedures defined in <a href="../base/gscrd.h">gscrd.h</a>), and then write
it as a parameter using <code>param_write_cie_render1</code> defined in
<a href="../base/gscrdp.h">gscrdp.h</a>.  However, the TransformPQR procedure
requires special handling.  If the CRD uses a TransformPQR procedure
different from the default (identity), the driver must do the following:

<ul>
<li>The TransformPQR element of the CRD must include a
<code>proc_name</code>, and optionally <code>proc_data</code>.  The
<code>proc_name</code> is an arbitrary name chosen by the driver to
designate the particular TransformPQR function.  It must not be the same as
any device parameter name; we strongly suggest it include the device name,
for instance, "<code>bitTPQRDefault</code>".

<li>For each such named TransformPQR procedure, the driver's
<code>get_param</code> procedure must provide a parameter of the same
name.  The parameter value must be a string whose bytes are the actual
procedure address.
</ul>

<p>
For a complete example, see the <code>bit_get_params</code> procedure in
<a href="../base/gdevbit.c">gdevbit.c</a>.  Note that it is essential that
the driver return the CRD or the procedure address only if specifically
requested (<code>param_requested(...)</code> &gt; 0); otherwise, errors
will occur.

<h4><a name="Device parameters affecting interpretation"></a>Device parameters affecting interpretation</h4>

<p>
Some parameters have been defined for high level device drivers which affect
the operation of the interpreter. These are documented here so that other devices
requiring the same behaviour can use these parameters.

<blockquote><dl>
<dt><b><em>/HighLevelDevice</em></b>
<dd>True if the device is a high level device. Currently this controls haltone emission
during setpagedevice. Normally setpagdevice resets the halftone to a default value, which is
unfortunate for high-level devices such as ps2write and pdfwrite, as they are unable to tell
that this is caused by setpagdevice rather than a halftone set by the input file. In order to prevent 
spurious default halftones being embedded in the output, if /HighLevelDevice is present and
true in the device paramters, then the default halftone will not be set during setpagedevice.
</dd>

<dt><b><em>/AllowIncrementalCFF</em></b>
<dd>Pdfwrite relies on font processing occuring in a particular order, which 
may not happen if CFF fonts are downloaded incrementally. Defining this 
parameter to true will prevent incremental CFF downloading (may raise an error
during processing).
</dd>

<dt><b><em>/AllowPSRepeatFuncs</em></b>
<dd>Pdfwrite emits functions as type 4, and as a result can't convert PostScript
functions using the repeat operator into PDF functions. Defining this parameter
as true will cause such functions to raise an error during processing.
</dd>

<dt><b><em>/IsDistiller</em></b>
<dd>Defining this parameter as true will result in the operators relating to 
'distillerparams' being defined (setdistillerparams/currentdistillerparams).
Some PostScript files behave differently if these operators are present (e.g.
rotating the page) so this parameter may be true even if the device is not
strictly a Distiller. For example ps2write defines this parameter to be
true.
</dd>

<dt><b><em>/PreserveSMask</em></b>
<dd>If this parameter is true then the PDF interpreter will not convert SMask
(soft mask, ie transparent) images into opaque images. This should be set to true
for devices which can handle transparency (e.g. pdfwrite)
</dd>

<dt><b><em>/PreserveTrMode</em></b>
<dd>If this parameter is true then the PDF interpreter will not handle Text
Rendering modes by degenerating into a sequence of text operations, but will
instead set the Tr mode, and emit the text once. This value should be true
for devices which can handle PDF text rendering modes directly.
</dd>

<dt><b><em>/WantsToUnicode</em></b>
<dd>In general, Unicode values are not of interest to rendering devices, but
for high level devices, they can be extremely valuable. If this parameter is
defined as true then ToUnicode CMaps and GlyphName2Unicode tables will be
processed and stored.
</dd>
</dl></blockquote>

<h3><a name="External_fonts"></a>External fonts</h3>

<p>
Drivers may include the ability to display text.  More precisely, they may
supply a set of procedures that in turn implement some font and text
handling capabilities, described in <a href="Xfonts.htm">a separate
document</a>.  The link between the two is the driver procedure that
supplies the font and text procedures:

<dl>
<dt><code>xfont_procs&nbsp;*(*get_xfont_procs)(gx_device&nbsp;*dev)</code> <b><em>[OPTIONAL]</em></b>
<dd>Return a structure of procedures for handling external fonts and text
display.  A <code>NULL</code> value means that this driver doesn't
provide this capability.
</dl>

<p>
For technical reasons, a second procedure is also needed:

<dl>
<dt><code>gx_device&nbsp;*(*get_xfont_device)(gx_device&nbsp;*dev)</code> <b><em>[OPTIONAL]</em></b>
<dd>Return the device that implements <code>get_xfont_procs</code> in a
non-default way for this device, if any.  Except for certain special
internal devices, this is always the device argument.
</dl>

<h3><a name="Page_devices"></a>Page devices</h3>

<dl>
<dt><code>gx_device&nbsp;*(*get_page_device)(gx_device&nbsp;*dev)</code>
<b><em>[OPTIONAL]</em></b>
<dd>According to the Adobe specifications, some devices are "page devices"
and some are not.  This procedure returns <code>NULL</code> if the
device is not a page device, or the device itself if it is a page device.
In the case of forwarding devices, <code>get_page_device</code> returns
the underlying page device (or <code>NULL</code> if the underlying
device is not a page device).
</dl>

<h3><a name="Miscellaneous"></a>Miscellaneous</h3>

<dl>
<dt><code>int (*get_band)(gx_device&nbsp;*dev, int&nbsp;y,
int&nbsp;*band_start)</code> <b><em>[OPTIONAL]</em></b>
<dd>If the device is a band device, this procedure stores in
<code>*band_start</code> the scan line (device Y coordinate) of the band
that includes the given Y coordinate, and returns the number of scan lines
in the band.  If the device is not a band device, this procedure returns 0.
The latter is the default implementation.
</dl>

<dl>
<dt><code>void (*get_clipping_box)(gx_device&nbsp;*dev,
gs_fixed_rect&nbsp;*pbox)</code> <b><em>[OPTIONAL]</em></b>
<dd>Stores in <code>*pbox</code> a rectangle that defines the device's
clipping region.  For all but a few specialized devices, this is
<em>((0,0),(width,height))</em>.
</dl>

<h3><a name="DevSpecOp"></a>Device Specific Operations</h3>

<p>In order to enable the provision of operations that make sense only
to a small range of devices/callers, we provide an extensible function. The
operation to perform is specified by an integer, taken from an enumeration
in <a href="../base/gxdevsop.h">gxdevsop.h</a>.

<p>A typical user of this function might make a call to detect whether
a device works in a particular way (such as whether it has a particular
color mapping) to enable an optimisation elsewhere. Sometimes it may be used
to detect a particular piece of functionality (such as whether
<code>copy_plane</code> is supported); in other cases it may be used both
to detect the presence of other functionality and to perform functions as
well (such as with the pdf specific pattern management calls - moved
here from their own dedicated device function).</p>

<p>This function is designed to be easy to chain through multiple levels of
device without each intermediate device needing to know about the full
range of operations it may be asked to perform.</p>

<dl>
<dt><code>int (*dev_spec_op)(gx_device&nbsp;*dev, int&nbsp;dso,
void&nbsp;*data, int&nbsp;size)</code> <b><em>[OPTIONAL]</em></b>
<dd>Perform device specific operation <code>dso</code>. Returns
<code>gs_error_undefined</code> for an unknown (or unsupported operation),
other negative values for errors, and (<code>dso</code> specific)
non-negative values to indicate success. For details of the meanings of
<code>dso</code>, <code>data</code> and <code>size</code>, see
<a href="../base/gxdevsop.h">gxdevsop.h</a>.
</dl>

<hr>

<h2><a name="Tray"></a>Tray selection</h2>

<!-- Note for documentation maintainers: tray selection overlaps -->
<!-- significantly across the device interface and the PostScript -->
<!-- language implementation of setpagedevice, while the rest of -->
<!-- Drivers.htm focusses on lanugage-independent interfaces. Likely -->
<!-- the documentation should be refactored a bit so that this section -->
<!-- has a comfortable home. -->

<p>The logic for selecting input trays, and modifying other parameters
based on tray selection, can be complex and subtle, largely thanks to
the requirement to be compatible with the PostScript language
setpagedevice mechanism. This section will describe recipes for
several common scenarios for tray selection, with special attention to
the how the overall task factors into configuration options, generic
logic provided by the PostScript language (or not, if the device is
used with other PDL's), and implementation of the put_param /
get_param device functions within the device.

<p>In general, tray selection is determined primarily through the
setpagedevice operator, which is part of the PostScript runtime.
Ghostscript attempts to be as compatible as is reasonable with the
PostScript standard, so for more details, see the description in the
<a
href="http://partners.adobe.com/public/developer/ps/index_specs.html">PostScript
language specifications</a>, including the "supplements", which tend
to have more detail about setpagedevice behavior than the PLRM book itself.

<p>The first step is to set up an /InputAttributes dictionary matching
the trays and so on available in the device. The standard Ghostscript
initialization files set up a large InputAttributes dictionary with
many "known" page sizes (the full list is in
<code>gs_statd.ps</code>, under .setpagesize). It's possible to
edit this list in the Ghostscript source, of course, but most of the
time it is better to execute a snippet of PostScript code after the
default initialization but before sending any actual jobs.

<p>Simply setting a new /InputAttributes dictionary with setpagedevice
will not work, because the the language specification for
setpagedevice demands a "merging" behavior - paper tray keys present
in the old dictionary will be preserved even if the key is not present
in the new /InputAttributes dictionary. Here is a sample invocation
that clears out all existing keys, and installs three new ones: a US letter
page size for trays 0 and 1, and 11x17 for tray 1. Note that you must add at
least one valid entry into the /InputAttributes dictionary; if all are
<code>null</code>, then the setpagedevice will fail with a
/configurationerror.

<blockquote><code>
&lt;&lt; /InputAttributes<br>
&nbsp;&nbsp;currentpagedevice /InputAttributes get<br>
&nbsp;&nbsp;dup { pop 1 index exch null put } forall<br>
<br>
&nbsp;&nbsp;dup 0 &lt;&lt; /PageSize [612 792] &gt;&gt; put<br>
&nbsp;&nbsp;dup 1 &lt;&lt; /PageSize [612 792] &gt;&gt; put<br>
&nbsp;&nbsp;dup 2 &lt;&lt; /PageSize [792 1224] &gt;&gt; put<br>
&gt;&gt; setpagedevice<br>
</code></blockquote>

<p>After this code runs, then requesting a letter page size (612x792
points) from setpagedevice will select tray 0, and requesting an 11x17
size will select tray 2. To explicitly request tray 1, run:

<blockquote><code>
&lt;&lt; /PageSize [612 792] /MediaPosition 1 &gt;&gt; setpagedevice
</code></blockquote>

<p>At this point, the chosen tray is sent to the device as the
(nonstandard) %MediaSource device parameter. Devices with switchable
trays should implement this device parameter in the
<code>put_params</code> procedure. Unlike the usual protocol for
device parameters, it is not necessary for devices to also implement
<code>get_params</code> querying of this paramter; it is
effectively a write-only communication from the language to the
device. Currently, among the devices that ship with Ghostscript, only
PCL (gdevdjet.c) and PCL/XL (gdevpx.c) implement this parameter, but
that list may well grow over time.

If the device has dynamic configuration of trays, etc., then the
easiest way to get that information into the tray selection logic is
to send a setpagedevice request (if using the standard API, then using
gsapi_run_string_continue) to update the /InputAttributes dictionary
immediately before beginning a job.

<h3><a name="LeadingEdge"></a>Tray rotation and the LeadingEdge parameter</h3>

<p>Large, sophisticated printers often have multiple trays supporting
both short-edge and long-edge feed. For example, if the paper path is
11 inches wide, then 11x17 pages must always print short-edge, but
letter size pages print with higher throughput if fed from long-edge
trays. Generally, the device will expect the rasterized bitmap image
to be rotated with respect to the page, so that it's always the same
orientation with respect to the paper feed direction.

<p>The simplest way to achieve this behavior is to call
<code>gx_device_request_leadingedge</code> to request a LeadingEdge
value
<code>LeadingEdge</code> field in the device structure based on the
%MediaSource tray selection index and knowledge of the device's
trays. The default put_params implementation will then handle this
request (it's done this way to preserve the transactional semantics of
put_params; it needs the new value, but the changes can't actually be
made until all params succeed). For example, if tray 0 is long-edge,
while trays 1 and 2 are short-edge, the following code outline should
select the appropriate rotation:

<blockquote><code>
my_put_params(gx_device *pdev, gs_param_list *plist) {<br>
&nbsp;&nbsp;&nbsp;&nbsp;my_device *dev = (my_device *)pdev;<br>
&nbsp;&nbsp;&nbsp;&nbsp;int MediaSource = dev-&gt;myMediaSource;<br>
<br>
&nbsp;&nbsp;&nbsp;&nbsp;code = param_read_int(plist, "%MediaSource", &amp;MediaSource);<br>
<br>
&nbsp;&nbsp;&nbsp;&nbsp;switch (MediaSource) {<br>
&nbsp;&nbsp;&nbsp;&nbsp;case 0:<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gx_device_req_leadingedge(dev, 1);<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;break;<br>
&nbsp;&nbsp;&nbsp;&nbsp;case 1:<br>
&nbsp;&nbsp;&nbsp;&nbsp;case 2:<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gx_device_req_leadingedge(dev, 0);<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;break;<br>
&nbsp;&nbsp;&nbsp;&nbsp;}<br>
&nbsp;&nbsp;&nbsp;&nbsp;...call default put_params, which makes the change...<br>
<br>
&nbsp;&nbsp;&nbsp;&nbsp;dev-&gt;myMediaSource = MediaSource;<br>
&nbsp;&nbsp;&nbsp;&nbsp;return 0;<br>
}
</code></blockquote>

<p>Ghostscript also supports explicit rotation of the page through
setting the /LeadingEdge parameter with setpagedevice. The above code
snippet will simply override this request. To give manual setting
through setpagedevice priority, don't change the LeadingEdge field in
the device if its LEADINGEDGE_SET_MASK bit is set. In other words,
simply enclose the above <tt>switch</tt> statement inside an <code>if
(!(dev-&gt;LeadingEdge &amp; LEADINGEDGE_SET_MASK) { ... }</code> statement.

<!-- Note for doc maintainers: the following is much more of a -->
<!-- discussion of the PS language than a device interface issue, but -->
<!-- it is essential info for people implementing this stuff. -->

<h3><a name="LeadingPage"></a>Interaction between LeadingEdge and PageSize</h3>

<p>As of LanguageLevel 3, PostScript now has two mechanisms for rotating
the imaging of the page: the LeadingEdge parameter described in detail
above, and the automatic rotation as enabled by the /PageSize page
device parameter (described in detail in Table 6.2 of the PLRM3).
Briefly, the PageSize autorotation handles the case where the page
size requested in setpagedevice matches the <i>swapped</i> size of the
paper source (as set in the InputAttributesDictionary). This mechanism
can be, and has been, used to implement long-edge feed, but has
several disadvantages. Among other things, it's overly tied to the PostScript
language, while the device code above will work with other
languages. Also, it only specifies one direction of rotation (90
degrees counterclockwise). Thus, given the choice, LeadingEdge is to
be preferred.

<p>If PageSize is used, the following things are different:

<ul>
<li>The PageSize array in InputAttributes is swapped, so it is [long
short].
<li>The .MediaSize device parameter is similarly swapped.
<li>The initial matrix established by the device through the
<code>get_initial_matrix</code> procedure is the same as for the
non-rotated case.
<li>The CTM rotation is done in the setpagedevice implementation.
</ul>

<!-- Why oh why does it all have to be so complicated? -->

<!-- [2.0 end contents] ==================================================== -->

<!-- [3.0 begin visible trailer] =========================================== -->
<hr>

<p>
<small>Copyright &copy; 2000-2007 Artifex Software, Inc.  All rights reserved.</small>

<p>
This software is provided AS-IS with no warranty, either express or
implied.

This software is distributed under license and may not be copied, modified
or distributed except as expressly authorized under the terms of that
license.  Refer to licensing information at http://www.artifex.com/
or contact Artifex Software, Inc.,  7 Mt. Lassen Drive - Suite A-134,
San Rafael, CA  94903, U.S.A., +1(415)492-9861, for further information.

<p>
<small>Ghostscript version 9.05, 8 February 2012

<!-- [3.0 end visible trailer] ============================================= -->

</body>
</html>
ghostscript-doc 9.05~dfsg-0ubuntu4.5 / usr / share / doc / ghostscript-doc / Drivers.htm
