<!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>Ghostscript and the PostScript language</title>
<!-- Originally: language.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>Ghostscript and the PostScript language</h1>
<!-- [1.1 end headline] ==================================================== -->
<!-- [1.2 begin table of contents] ========================================= -->
<h2>Table of contents</h2>
<blockquote><ul>
<li><a href="#Capabilities">Ghostscript's capabilities in relation to PostScript</a>
<li><a href="#Implementation_limits">Implementation limits</a>
<ul>
<li><a href="#Architectural_limits">Architectural limits</a>
<li><a href="#Typical_memory_limits">Typical memory limits in LanguageLevel 1</a>
<li><a href="#VM_consumption">Other differences in VM consumption</a>
</ul>
<li><a href="#Additional_operators">Additional operators in Ghostscript</a>
<ul>
<li><a href="#Graphics_and_text">Graphics and text operators</a>
<ul>
<li><a href="#Transparency">Transparency</a>
<ul>
<li><a href="#Transparency_graphics_state_operators">Graphics state operators</a>
<li><a href="#Transparency_rendering_stack_operators">Rendering stack operators</a>
<li><a href="#Transparency_ImageType">New ImageType</a>
</ul>
<li><a href="#Graphics_state">Other graphics state operators</a>
<li><a href="#Path">Path operators</a>
<li><a href="#Painting">Painting operators</a>
<li><a href="#Character">Character operators</a>
</ul>
<li><a href="#Other">Other operators</a>
<ul>
<li><a href="#Mathematical">Mathematical operators</a>
<li><a href="#Dictionary">Dictionary operators</a>
<li><a href="#String">String and name operators</a>
<li><a href="#Relational">Relational operators</a>
<li><a href="#File">File operators</a>
<li><a href="#Virtual_memory">Virtual memory operators</a>
<li><a href="#Miscellaneous">Miscellaneous operators</a>
<li><a href="#Device">Device operators</a>
</ul>
</ul>
<li><a href="#Filters">Filters</a>
<ul>
<li><a href="#Standard_filters">Standard filters</a>
<li><a href="#Non_standard_filters">Non-standard filters</a>
<li><a href="#Unstable_filters">Unstable filters</a>
</ul>
<li><a href="#Device_parameters">Device parameters</a>
<li><a href="#Banding_parameters">Banding parameters</a>
<li><a href="#User_parameters">User parameters</a>
<li><a href="#Miscellaneous_additions">Miscellaneous additions</a>
<ul>
<li><a href="#Extended_semantics_of_run">Extended semantics of 'run'</a>
<li><a href="#DecodingResources">Decoding resources</a>
<li><a href="#CIDDecodingResources">CIDDecoding resources</a>
<li><a href="#GlyphNames2Unicode">GlyphNames2Unicode</a>
<li><a href="#MultipleResourceDirectories">Multiple Resource directories</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>.
<!-- [1.3 end hint] ======================================================== -->
<hr>
<!-- [1.0 end visible header] ============================================== -->
<!-- [2.0 begin contents] ================================================== -->
<h2><a name="Capabilities"></a>Ghostscript's capabilities in relation to PostScript</h2>
<p>
The Ghostscript interpreter, except as noted below, is intended to execute
properly any source program written in the (LanguageLevel 3)
<b>PostScript</b> language as defined in the <cite>PostScript
Language Reference, Third Edition</cite> (ISBN 0-201-37922-8) published by
Addison-Wesley in mid-1999. However, the interpreter is configurable in
ways that can restrict it to various subsets of this language.
Specifically, the base interpreter accepts the Level 1 subset of the
PostScript language, as defined in the first edition of the <cite>PostScript
Language Reference Manual</cite> (ISBN 0-201-10174-2) Addison-Wesley 1985,
plus the file system, version 25.0 language, and miscellaneous additions
listed in sections A.1.6, A.1.7, and A.1.8 of the Second Edition
respectively, including allowing a string operand for the
"<code>status</code>" operator. The base interpreter may be configured
(see the <a href="Make.htm">documentation on building Ghostscript</a> for
how to configure it) by adding any combination of the following:
<ul>
<li>The ability to process PostScript Type 1 fonts. This facility is
normally included in the interpreter.
<li>The CMYK color extensions listed in section A.1.4 of the Second Edition
(including <code>colorimage</code>). These facilities are available
only if the <code>color</code>, <code>dps</code>, or
<code>level2</code> feature was selected when Ghostscript was built.
<li>The Display PostScript extensions listed in section A.1.3 of the Second
Edition, but excluding the operators listed in section A.1.2. These
facilities are available only if the <code>dps</code> feature or the
<code>level2</code> feature was selected when Ghostscript was built.
<li>The composite font extensions listed in section A.1.5 of the Second
Edition, and the ability to handle Type 0 fonts. These facilities are
available only if the <code>compfont</code> feature or the
<code>level2</code> feature was selected when Ghostscript was built.
<li>The ability to load TrueType fonts and to handle PostScript Type 42
(encapsulated TrueType) fonts. These facilities are available only if the
<code>ttfont</code> feature was selected when Ghostscript was built.
<li>The PostScript Level 2 "filter" facilities except the
<code>DCTEncode</code> and <code>DCTDecode</code> filters. These
facilities are available only if the <code>filter</code>,
<code>dps</code>, or <code>level2</code> feature was selected when
Ghostscript was built.
<li>The PostScript Level 2 <code>DCTEncode</code> and
<code>DCTDecode</code> filters. These facilities are available only if
the <code>dct</code> or <code>level2</code> feature was selected when
Ghostscript was built.
<li>All the other PostScript Level 2 operators and facilities listed in
section A.1.1 of the Second Edition and not listed in any of the other
A.1.n sections. These facilities are available only if the
<code>level2</code> feature was selected when Ghostscript was built.
<li>All PostScript LanguageLevel 3 operators and facilities listed in the
Third Edition, except as noted below. These facilities are available only
if the <code>psl3</code> feature was selected when Ghostscript was built.
<li>The ability to recognize DOS EPSF files and process only the PostScript
part, ignoring bitmap previews or other information. This facility is
available only if the <code>epsf</code> feature was selected when
Ghostscript was built.
</ul>
<p>
Ghostscript currently does not implement the following PostScript
LanguageLevel 3 facilities:
<ul>
<li>Settable <code>ProcessColorModel</code> for page devices, except for
a very few special devices.
<li><code>IODevice</code>s other than <code>%stdin</code>,
<code>%stdout</code>, <code>%stderr</code>, <code>%lineedit</code>,
<code>%statementedit</code>, <code>%os%</code>, and (if configured)
<code>%pipe%</code> and <code>%disk0%</code> through <code>%disk0%</code>.
</ul>
<p>
Ghostscript can also interpret files in the Portable Document Format (PDF)
1.7 format defined in the
<a href="http://www.adobe.com/devnet/pdf/pdf_reference.html"><em>PDF
Reference</em> Version 1.7</a>,
distributed by <a href="http://www.adobe.com/">Adobe Systems
Incorporated</a>, except as noted below. This facility can be
disabled by deselecting the <code>pdf</code> feature
when Ghostscript is built.
<p>
Ghostscript currently implements the majority of non-interactive
features defined in the PDF reference.
<p>
Ghostscript also includes a number of
<a href="#Additional_operators">additional operators</a> defined below that
are not in the PostScript language defined by Adobe.
<hr>
<h2><a name="Implementation_limits"></a>Implementation limits</h2>
<p>
The implementation limits show here correspond to those in Tables B.1 and
B.2 of the Second and Third Editions, which describe the quantities fully.
Where Ghostscript's limits are different from those of Adobe's
implementations (as shown in the Third Edition), Adobe's limits are also
shown.
<h3><a name="Architectural_limits"></a>Architectural limits</h3>
<blockquote><table cellpadding=0 cellspacing=0>
<tr><th colspan=7 bgcolor="#CCCC00"><hr><font size="+1">Architectural limits (corresponds to Adobe table B.1)</font><hr>
<tr valign=bottom>
<th align=left>Quantity
<td>
<th align=left>Limit
<td>
<th align=left>Type
<td>
<th align=left>Adobe
<tr> <td colspan=7><hr>
<tr valign=top> <td>integer
<td>
<td>32-bit
<td>
<td>twos complement integer
<td>
<td>
<tr valign=top> <td>real
<td>
<td>single-precision
<td>
<td>IEEE float
<td>
<td>
<tr valign=top> <td>array
<td>
<td>16777216
<td>
<td>elements
<td>
<td>65535
<tr valign=top> <td>dictionary
<td>
<td>16777215
<td>
<td>elements
<td>
<td>65535
<tr valign=top> <td>string
<td>
<td>16777216
<td>
<td>characters
<td>
<td>65535
<tr valign=top> <td>name
<td>
<td>16383
<td>
<td>characters
<td>
<td>127
<tr valign=top> <td>filename
<td>
<td>128*
<td>
<td>characters
<td>
<td>
<tr valign=top> <td><code>save</code> level
<td>
<td>none
<td>
<td>(capacity of memory)
<td>
<td>15
<tr valign=top> <td><code>gsave</code> level
<td>
<td>none
<td>
<td>(capacity of memory)
<td>
<td>13
</table></blockquote>
<p>
* The limit on the length of a file name is 128 characters if the name
starts with a %...% IODevice designation, or 124 characters if it does not.
<h3><a name="Typical_memory_limits"></a>Typical memory limits in LanguageLevel 1</h3>
<blockquote><table cellpadding=0 cellspacing=0>
<tr><th colspan=7 bgcolor="#CCCC00"><hr><font size="+1">Memory limits (corresponds to Adobe table B.2)</font><hr>
<tr valign=bottom>
<th align=left>Quantity
<td>
<th align=left>Limit
<td>
<th align=left>Type
<td>
<th align=left>Adobe
<tr> <td colspan=7><hr>
<tr valign=top> <td><code>userdict</code>
<td>
<td>200
<td>
<td>
<td>
<td>
<tr valign=top> <td><code>FontDirectory</code>
<td>
<td>100
<td>
<td>
<td>
<td>
<tr valign=top> <td>operand stack
<td>
<td>800
<td>
<td>
<td>
<td>500
<tr valign=top> <td>dictionary stack
<td>
<td>20
<td>
<td>
<tr valign=top> <td>execution stack
<td>
<td>250
<td>
<td>
<tr valign=top> <td>interpreter level
<td>
<td>none
<td>
<td>(capacity of memory)
<td>
<td>10
<tr valign=top> <td>path
<td>
<td>none
<td>
<td>(capacity of memory)
<td>
<td>1500
<tr valign=top> <td>dash
<td>
<td>11
<td>
<td>
<tr valign=top> <td>VM
<td>
<td>none
<td>
<td>(capacity of memory)
<td>
<td>240000
<tr valign=top> <td>file
<td>
<td>none
<td>
<td>(determined by operating system)
<td>
<td>6
<tr valign=top> <td>image
<td>
<td>65535
<td>
<td>values (samples × components)<br>for 1-, 2-, 4-, or 8-bit samples
<td>
<td>3300
<tr valign=top> <td>
<td>
<td>32767
<td>
<td>values for 12-bit samples
<td>
<td>3300
</table></blockquote>
<h3><a name="VM_consumption"></a>Other differences in VM consumption</h3>
<p>
In 32-bit builds packed array elements occupy either 2 bytes or 12 bytes.
The average element size is probably about 7 bytes. Names occupy 16 bytes plus the
space for the string.
<p>
In 64-bit builds packed array elements occupy either 2 bytes or 16 bytes.
The average element size is probably about 9 bytes.
Names occupy 24 bytes plus the space for the string.
<p>
The garbage collector doesn't reclaim portions of arrays obtained with
<tt>getinterval</tt>, rather it collects entire arrays.
<hr>
<h2><a name="Additional_operators"></a>Additional operators in Ghostscript</h2>
<h3><a name="Graphics_and_text"></a>Graphics and text operators</h3>
<h4><a name="Transparency"></a>Transparency</h4>
<p>
Ghostscript provides a set of operators for implementing the transparency
and compositing facilities of PDF 1.4. These are defined only if the
<code>transpar</code> option was selected when Ghostscript was built. We
do not attempt to explain the underlying graphics model here: for details,
see <a
href="http://partners.adobe.com/asn/developer/technotes.html#acrobat-pdf"
class="offsite">Adobe
Technical Note</a> #5407, "<a
href="http://partners.adobe.com/asn/developer/acrosdk/DOCS/PDF_Transparency.pdf"
class="offsite">Transparency
in PDF</a>". Note, however, that
Ghostscript's model generalizes that of PDF 1.4 in that Ghostscript
maintains separate alpha and mask values for opacity and shape, rather than
a single value with a Boolean that says whether it represents opacity or
shape.
<h5><a name="Transparency_graphics_state_operators"></a>Graphics state
operators</h5>
<dl>
<dt><code><modename> .setblendmode -</code>
<dd>Sets the blending mode in the graphics state. If the mode name is not
recognized, causes a <code>rangecheck</code> error. The initial value of
the blending mode is <code>/Compatible</code>.
</dl>
<dl>
<dt><code>- .currentblendmode <modename></code>
<dd>Returns the current blending mode.
</dl>
<dl>
<dt><code><0..1> .setopacityalpha -</code>
<dd>Sets the opacity alpha value in the graphics state.
The initial opacity alpha value is 1.
</dl>
<dl>
<dt><code>- .currentopacityalpha <0..1></code>
<dd>Returns the current opacity alpha value.
</dl>
<dl>
<dt><code><0..1> .setshapealpha -</code>
<dd>Sets the shape alpha value in the graphics state.
The initial shape alpha value is 1.
</dl>
<dl>
<dt><code>- .currentshapealpha <0..1></code>
<dd>Returns the current shape alpha value.
</dl>
<dl>
<dt><code><bool> .settextknockout -</code>
<dd>Sets the text knockout flag in the graphics state.
The initial value of the text knockout flag is <code>true</code>.
</dl>
<dl>
<dt><code>- .currenttextknockout <bool></code>
<dd>Returns the current text knockout flag.
</dl>
<h5><a name="Transparency_rendering_stack_operators"></a>Rendering stack
operators</h5>
<p>
The interpreter state is extended to include a (per-context) rendering stack
for handling transparency groups and masks (generically, "layers"). Groups
accumulate a full value for each pixel (paint plus transparency); masks
accumulate only a coverage value. Layers must be properly nested, i.e., the
'end' or 'discard' operator must match the corresponding 'begin' operator.
<p>
Beginning and ending layers must nest properly with respect to
<code>save</code> and <code>restore</code>: <code>save</code> and
<code>restore</code> do not save and restore the layer stack. Currently,
layers are not required to nest with respect to <code>gsave</code> and
<code>grestore</code>, except that the device that is current in the
graphics state when ending a layer must be the same as the device that was
current when beginning the layer. THIS AREA IS SUBJECT TO CHANGE.
<dl>
<dt><code><paramdict> <llx> <lly> <urx> <ury>
.begintransparencygroup -</code>
<dd>Begins a new transparency group. The <code>ll/ur</code> coordinates
are the bounding box of the group in the current user coordinate system.
<code>paramdict</code> has the following keys:
<dl>
<dt><code>/Isolated</code>
<dd>(optional) Boolean; default value = <code>false</code>.
<dt><code>/Knockout</code>
<dd>(optional) Boolean; default value = <code>false</code>.
</dl>
</dl>
<dl>
<dt><code>- .discardtransparencygroup -</code>
<dd>Ends and discards the current transparency group.
</dl>
<dl>
<dt><code>- .endtransparencygroup -</code>
<dd>Ends the current transparency group, compositing the group being ended
onto the group that now becomes current.
</dl>
<dl>
<dt><code><paramdict> <llx> <lly> <urx> <ury>
.begintransparencymaskgroup -</code>
<dd>Begins a new transparency mask, which is represented as a group.
The <code>ll/ur</code> coordinates
are the bounding box of the mask in the current user coordinate system.
<code>paramdict</code> has the following keys:
<dl>
<dt><code>/Subtype</code>
<dd>(required) Name, either <code>/Alpha</code> or
<code>/Luminosity</code>.
<dt><code>/Background</code>
<dd>(optional) Array of number.
<dt><code>/TransferFunction</code>
<dd>(optional) Function object (produced by applying
<code>.buildfunction</code> to a Function dictionary).
</dl>
</dl>
<dl>
<dt><code>- .begintransparencymaskimage -</code>
<dd>Begins a new transparency mask, which is represented as a single image.
</dl>
<dl>
<dt><code>- .discardtransparencymask -</code>
<dd>Ends and discards the current transparency mask.
</dl>
<dl>
<dt><code><masknum> .endtransparencymask -</code>
<dd>Ends the current transparency mask, installing it as the current opacity
(<code>masknum</code> = 0) or shape (<code>masknum</code> = 1) mask in
the graphics state.
</dl>
<dl>
<dt><code><masknum> .inittransparencymask -</code>
<dd>Resets the current opacity (<code>masknum</code> = 0) or shape
(<code>masknum</code> = 1) mask to an infinite mask with alpha = 1
everywhere.
</dl>
<h5><a name="Transparency_ImageType"></a>New ImageType</h5>
<p>
The transparency extension defines a new ImageType 103, similar to ImageType
3 with the following differences:
<ul>
<li>The required <code>MaskDict</code> is replaced by two optional
dictionaries, <code>OpacityMaskDict</code> and
<code>ShapeMaskDict</code>. If present, these dictionaries must have a
<code>BitsPerComponent</code> entry, whose value may be greater than 1.
Note that in contrast to ImageType 3, where any non-zero chunky mask value
is equivalent to 1, ImageType 103 simply takes the low-order bits of chunky
mask values.
<li>A <code>Matte</code> entry may be present in one or both mask
dictionaries, indicating premultiplication of the data values. If both
<code>MaskDict</code>s have a <code>Matte</code> entry and the values
of the two <code>Matte</code> entries are different, a
<code>rangecheck</code> error occurs.
<li><code>InterleaveType</code> appears in the <code>MaskDict</code>s,
not the <code>DataDict</code>, because each mask has its own
<code>InterleaveType</code>. <code>InterleaveType</code> 2
(interlaced scan lines) is not supported.
</ul>
<h4><a name="Graphics_state"></a>Other graphics state operators</h4>
<dl>
<dt><code><bool> .setaccuratecurves -</code>
<dd>Sets a graphics state flag that determines whether curves and arcs,
when flattened, always start and end with a line that is a segment of the
tangent; this also causes butt and square caps to be properly perpendicular
to the tangent. <code>initgraphics</code> sets this flag to false, to
match other PostScript implementations.
</dl>
<dl>
<dt><code>- .currentaccuratecurves <bool></code>
<dd>Returns the current value of the accurate curves flag.
</dl>
<dl>
<dt><code><int> .setcurvejoin -</code>
<dd>Obsolete, left for backward compatibility.
<dd>Sets a graphics state parameter that determines how to treat the joins
between the line segments produced when a curve is flattened. The parameter
value may be either -1 or a value acceptable to <code>setlinejoin</code>.
If the parameter value is -1, the join used for flattened curve line
segments is given by the current line join parameter in the graphics state
(except that if the line join value is "none", a bevel join is used), which
matches the Adobe Red Book, but not some old Adobe implementations; if the curve
join parameter value is a line join value, that type of join is used for
flattened curve line segments, regardless of the value of the graphics state
line join parameter. The initial (and default) value of the curve join
parameter is -1, causing the compatibility to Red Book and to modern Adobe
implementations. <code>initgraphics</code> sets the parameter to its
default value.
</dl>
<dl>
<dt><code>- .currentcurvejoin <int></code>
<dd>Obsolete, left for backward compatibility.
<dd>Returns the current value of the curve join parameter.
</dl>
<dl>
<dt><code><bool> .setdashadapt -</code>
<dd>Sets a graphics state flag that determines whether dash patterns do
(true) or do not (false) automatically scale themselves so that each line
segment consists of an integral number of pattern repetitions.
<code>initgraphics</code> sets this flag to false.
</dl>
<dl>
<dt><code>- .currentdashadapt <bool></code>
<dd>Returns the current value of the dash adaptation flag.
</dl>
<dl>
<dt><code><matrix> .setdefaultmatrix -</code>
<dd>Sets the default matrix that is returned by
<code>defaultmatrix</code> and installed by <code>initmatrix</code>.
Ordinary programs should not use this operator.
</dl>
<dl>
<dt><code><num> <bool> .setdotlength -</code>
<dd>Sets a graphics state parameter that determines the handling of
zero-length lines (dots). If the dot length is zero, dots are painted as
circles if round line caps are in effect, otherwise they are not painted at
all. If the dot length is non-zero, dots are treated exactly like lines of
the given length: the length is specified in user coordinates (like line
width) if <code>bool</code> is false, or in default user coordinates of
points (units of 1/72in; see the <a href="Devices.htm#Measurements">notes
on measurements</a> in the documentation on devices) if
<code>bool</code> is true. Dots occurring as part of dash patterns will
be oriented correctly; isolated dots will be oriented as though they were
part of a vertical line. <code>initgraphics</code> sets the dot length
to zero.
</dl>
<dl>
<dt><code>- .currentdotlength <num> <bool></code>
<dd>Returns the current dot length and dot length mode.
</dl>
<dl>
<dt><code><dx> <dy> .setfilladjust2 -</code>
<dd>Sets graphics state parameters that cause all filled and stroked
regions to be "fattened" by the given amount relative to an algorithm that
only paints pixels whose centers fall within the region to be painted.
<code>dx</code> and <code>dy</code> are numbers between 0 and 0.5,
measured in device space. The only two values that are likely to be useful
are 0, which gives a pure center-of-pixel rule, and 0.5, which gives
Adobe's any-part-of-pixel rule. (0.5 is treated slightly specially in
order to create half-open pixels per Adobe's specification.)
</dl>
<dl>
<dt><code>- .currentfilladjust2 <dx> <dy></code>
<dd>Returns the current fill adjustment values.
</dl>
<dl>
<dt><code><bool> .setlimitclamp -</code>
<dd>Sets a graphics state flag that determines whether attempts to set the
current point outside the internally representable range should clamp the
value to the largest representable value (true) or give a
<code>limitcheck</code> error (false). <code>initgraphics</code> sets
this flag to false, to match other PostScript implementations.
</dl>
<dl>
<dt><code>- .currentlimitclamp <bool></code>
<dd>Returns the current value of the limit clamp flag.
</dl>
<dl>
<dt><code><int> .setoverprintmode -</code>
<dd>Sets the overprint mode in the graphics state. Legal values are 0 or 1.
Per the PDF 1.3 specification, if the overprint mode is 1, then when the
current color space is <code>DeviceCMYK</code>, color components whose
value is 0 do not write into the target, rather than writing a 0 value.
THIS BEHAVIOR IS NOT IMPLEMENTED YET. The initial value of the overprint
mode is 0.
</dl>
<dl>
<dt><code>- .currentoverprintmode <int></code>
<dd>Returns the current overprint mode.
</dl>
<h4><a name="Path"></a>Path operators</h4>
<dl>
<dt><code>- .dashpath -</code>
<dd>If there is no current dash pattern, does nothing. Otherwise, does the
equivalent of <code>flattenpath</code> and then chops up the path as
determined by the dash pattern.
</dl>
<dl>
<dt><code><x> <y> <width> <height> .rectappend -</code>
<dt><code><numarray> .rectappend -</code>
<dt><code><numstring> .rectappend -</code>
<dd>Appends a rectangle or rectangles to the current path, in the same
manner as <code>rectfill</code>, <code>rectclip</code>, etc. Defined
only if the <code>dps</code> or <code>level2</code> option was
selected when Ghostscript was built.
</dl>
<dl>
<dt><code>- .getpath <array></code>
<dd>Export the current path to an array of executable arrays. This
operator is similar to <code>upath</code>, except
<ul>
<li>array of arrays is used to overcome 64K-1 limit on the
array lemgth.
<li>no <code>ucache</code> or <code>setbbox</code> entries are generated.
<li>operators are used instead of executable names.
<li>there's no special handling of an empty path or trailing <code>moveto</code>.
The saved path can be restored by <code>newpath { exec } forall</code>.
</ul>
</dl>
<dl>
<dt><code>- .sethpglpathmode</code>
<dd>If the single numerical parameter is non zero, then future path
creation operations will make an HPGL-style path. HPGL style paths
differ from postscript style paths in that moveto operations only
begin a new subpath if there is not already an open subpath. (i.e.
<code>a b moveto c d moveto e f lineto closepath</code> will draw
back to (a,b), not (c,d). HPGL paths treat these non-subpath starting
<tt>moveto</tt>s as <tt>lineto</tt> segments when filling, and as
unstroked gaps when stroking.
</dl>
<dl>
<dt><code>.currenthpglpathmode -</code>
<dd>Leave the current hpglpathmode on the stack. The initial value
of hpglpathmode is 0.
</dl>
<h4><a name="Painting"></a>Painting operators</h4>
<p>
Ghostscript supports an experimental extension of the PostScript imaging
model to include <code>RasterOp</code> and some related facilities.
This extension is available only if the <code>rasterop</code> option was
selected when building Ghostscript.
<p>
With the <code>RasterOp</code> extension, imaging operations compute a
function <b>D = f(D,S,T)</b> in RGB space, where <b>f</b> is an
arbitrary 3-input Boolean function, <b>D</b> is the destination (frame
buffer or print buffer), <b>S</b> is the source (described below), and
<b>T</b> is the texture (the current PostScript color, which may be a
pattern). The source and texture depend on the PostScript imaging
operation:
<ul>
<li>For <code>fill</code> and <code>stroke</code>, the source is
solid black, covering the region to be painted; the texture is the current
PostScript color.
<li>For <code>show</code> and <code>imagemask</code>, the source is
solid black, covering the pixels to be painted; the texture is the current
PostScript color.
<li>For <code>image</code> and <code>colorimage</code>, the source is
the image data; the texture depends on an optional Boolean parameter,
<code>CombineWithColor</code>, in the image dictionary. If
<code>CombineWithColor</code> is false (the default), the texture is
solid black. If <code>CombineWithColor</code> is true, the texture is
the current color. For the non-dictionary form of the image operator,
<code>CombineWithColor</code> is considered to be false.
</ul>
<p>
The <code>rasterop</code> option adds the following operators:
<dl>
<dt><code><int8> .setrasterop -</code>
<dd>Sets the <code>RasterOp</code> function in the graphics state. The
default function is 252, Source | Texture.
</dl>
<dl>
<dt><code>- .currentrasterop <int8></code>
<dd>Returns the current <code>RasterOp</code> function.
</dl>
<dl>
<dt><code><bool> .setsourcetransparent -</code>
<dd>Sets source transparency in the graphics state. When source
transparency is true, white source pixels prevent storing into the
destination, regardless of what the <code>RasterOp</code> function
returns. The default source transparency is false.
</dl>
<dl>
<dt><code>- .currentsourcetransparent <bool> -</code>
<dd>Returns the current source transparency.
</dl>
<dl>
<dt><code><bool> .settexturetransparent -</code>
<dd>Sets texture transparency in the graphics state. When texture
transparency is true, white texture pixels prevent storing into the
destination, regardless of what the <code>RasterOp</code> function
returns. The default texture transparency is false.
</dl>
<dl>
<dt><code>- .currenttexturetransparent <bool> -</code>
<dd>Returns the current texture transparency.
</dl>
<p>
For more information on RasterOp and transparency, please consult chapter 5
of the "PCL 5 Color Technical Reference Manual",
<a href="http://www.hp.com/cposupport/printers/support_doc/bpl01354.html">Hewlett-Packard
Manual Part No. 5961-0635</a>.
<h4><a name="Character"></a>Character operators</h4>
<dl>
<dt><code><string> <bool> .charboxpath -</code>
<dd>For each character <b>C</b> in the rendering of <string>, let the
bounding box of <b>C</b> <b><em>in device space</em></b> be the four
<b><em>user-space</em></b> points p1x/y, p2x/y, p3x/y, and p4x/y. For each
character in order, <code>.charboxpath</code> appends the following to
the current path:
<ul><li>If <code><bool></code> is true, the equivalent of:
<blockquote>
p1x p1y <code>moveto</code><br>
p2x p2y <code>lineto</code><br>
p3x p3y <code>lineto</code><br>
p4x p4y <code>lineto</code><br>
<code>closepath</code>
</blockquote>
</ul>
<p>
This creates a path whose <code>pathbbox</code> is the
<code>bbox</code> of the string.
<ul><li>If <code><bool></code> is false, the equivalent of:
<blockquote>
p1x p1y <code>moveto</code><br>
p3x p3y <code>lineto</code>
</blockquote>
</ul>
<p>
If the CTM is well-behaved (consists only of reflection, scaling, and
rotation by multiples of 90 degrees), this too creates a (simpler) path
whose <code>pathbbox</code> is the <code>bbox</code> of the string.
</dl>
<dl>
<dt><code><font> <charname|charcode> <charname> <charstring> .type1execchar -</code>
<dd>Does all the work for rendering a Type 1 outline. This operator, like
<code>setcharwidth</code> and <code>setcachedevice</code>, is valid
only in the context of a show operator -- that is, it must only be called
from within a <code>BuildChar</code> or <code>BuildGlyph</code>
procedure.
</dl>
<dl>
<dt><code><font> <charcode> %Type1BuildChar -</code>
<dd>This is not a new operator: rather, it is a name known specially to the
interpreter. Whenever the interpreter needs to render a character (during
a ...<code>show</code>, <code>stringwidth</code>, or
<code>charpath</code>), it looks up the name <code>BuildChar</code>
in the font dictionary to find a procedure to run. If it does not find
this name, and if the <code>FontType</code> is 1, the interpreter
instead uses the value (looked up on the dictionary stack in the usual way)
of the name <code>%Type1BuildChar</code>.
<p>
The standard definition of <code>%Type1BuildChar</code> is in the
initialization file <code>gs_type1.ps</code>. Users should not need to
redefine <code>%Type1BuildChar</code>, except perhaps for tracing or
debugging.
</dl>
<dl>
<dt><code><font> <charname> %Type1BuildGlyph -</code>
<dd>Provides the Type 1 implementation of <code>BuildGlyph</code>.
</dl>
<h3><a name="Other"></a>Other operators</h3>
<h4><a name="Mathematical"></a>Mathematical operators</h4>
<dl>
<dt><code><number> arccos <number></code>
<dd>Computes the arc cosine of a number between -1 and 1.
</dl>
<dl>
<dt><code><number> arcsin <number></code>
<dd>Computes the arc sine of a number between -1 and 1.
</dl>
<h4><a name="Dictionary"></a>Dictionary operators</h4>
<dl>
<dt><code>mark <key1> <value1> <key2> <value2> ... .dicttomark <dict></code>
<dd>Creates and returns a dictionary with the given keys and values. This
is the same as the PostScript Level 2 <code>>></code> operator,
but is available even in Level 1 configurations.
</dl>
<dl>
<dt><code><dict> <key> <value> .forceput - </code>
<dd>Equivalent to <code>put</code>, but works even if
<code>dict</code> is not writable, and (if <code>dict</code> is
<code>systemdict</code> or the current save level is 0) even if
<code>dict</code> is in global VM and <code>key</code> and/or
<code>value</code> is in local VM. <strong>This operator should be used
only initialization code, and only in executeonly procedures: it must not be
accessible after initialization.</strong>
</dl>
<dl>
<dt><code><dict> <key> .forceundef - </code>
<dd>Equivalent to <code>undef</code>, but works even if
<code>dict</code> is not writable. <strong>This operator should be used
only initialization code, and only in executeonly procedures: it must not be
accessible after initialization.</strong>
</dl>
<dl>
<dt><code><dict> <key> .knownget <value> true</code>
<dt><code><dict> <key> .knownget false</code>
<dd>Combines <code>known</code> and <code>get</code> in the
obvious way.
</dl>
<dl>
<dt><code><dict> <integer> .setmaxlength -</code>
<dd>Sets the capacity (<code>maxlength</code>) of a dictionary.
Causes a <code>dictfull</code> error if the dictionary has more
occupied entries than the requested capacity.
</dl>
<h4><a name="String"></a>String and name operators</h4>
<dl>
<dt><code><integer> .bytestring <bytestring></code>
<dd>Allocates and returns a bytestring, a special data type that can be
larger than the maximum size of a string (64K-1 bytes) and can be used in
place of a string with a very few operators.
</dl>
<dl>
<dt><code><name> .namestring <string></code>
<dd>Returns the (read-only) string for a name.
</dl>
<dl>
<dt><code><string> <charstring> .stringbreak <index|null></code>
<dd>Searches for a character in <code>string</code> that appears
somewhere in <code>charstring</code>. If such a character is found,
returns the index of the first such character; if no such character is
found, returns <code>null</code>.
</dl>
<dl>
<dt><code><obj> <pattern> .stringmatch <bool></code>
<dd>Matches <code>obj</code> against a pattern in which '*' matches 0 or
more characters and '?' matches any single character. If
<code>obj</code> is a string or a name, matches its characters against
the pattern; if <code>obj</code> is of any other type, the result is
<code>true</code> if the pattern is the single character "*" and
<code>false</code> otherwise.
</dl>
<dl>
<dt><code><state> <fromString> <toString> .type1encrypt <newState> <toSubstring></code>
<dd>Encrypts <code>fromString</code> according to the algorithm for
Adobe Type 1 fonts, writing the result into <code>toString</code>.
<code>toString</code> must be at least as long as
<code>fromString</code>, or a rangecheck error occurs.
<code>state</code> is the initial state of the encryption algorithm (a
16-bit non-negative integer); <code>newState</code> is the new state of
the algorithm.
</dl>
<dl>
<dt><code><state> <fromString> <toString> .type1decrypt <newState> <toSubstring></code>
<dd>Decrypts <code>fromString</code> according to the algorithm for
Adobe Type 1 fonts, writing the result into <code>toString</code>.
Other specifications are as for <code>type1encrypt</code>.
</dl>
<h4><a name="Relational"></a>Relational operators</h4>
<dl>
<dt><code><number|string> <number|string> max <number|string></code>
<dd>Returns the larger of two numbers or strings.
</dl>
<dl>
<dt><code><number|string> <number|string> min <number|string></code>
<dd>Returns the smaller of two numbers or strings.
</dl>
<h4><a name="File"></a>File operators</h4>
<dl>
<dt><code><file> .filename <string> true</code>
<dt><code><file> .filename false</code>
<dd>If the file was opened by the <code>file</code> or
<code>.tempfile</code> operator, returns the file name and
<code>true</code>; if the file is a filter, returns
<code>false</code>.
</dl>
<dl>
<dt><code><file> .fileposition <integer> true</code>
<dd>Returns the position of <code>file</code>. Unlike the standard
<code>fileposition</code> operator, which causes an error if the file is
not positionable, <code>.fileposition</code> works on all files,
including filters: for non-positionable files, it returns the total number
of bytes read or written since the file was opened.
</dl>
<dl>
<dt><code><string> findlibfile <foundstring> <file> true</code>
<dt><code><string> findlibfile <string> false</code>
<dd>Opens the file of the given name for reading, searching through
directories <a href="Use.htm#Finding_files">as described in the usage
documentation</a>. If the search fails, <code>findlibfile</code> simply
pushes false on the stack and returns, rather than causing an error.
</dl>
<dl>
<dt><code><file> <string> .peekstring <substring> <filled_bool></code>
<dd>Reads bytes from a file like <code>readstring</code>, but also leaves
the bytes in the file buffer so they will be read again by a subsequent read
operation. Currently gives a <code>rangecheck</code> error if
<code>string</code> is larger than the file's buffer.
</dl>
<a name=Tempfile></a>
<dl>
<dt><code><prefix_string|null> <access_string> .tempfile
<string> <file></code>
<dd>Creates and opens a temporary file
like the <code>file</code> operator, also returning the file name. There
are three cases for the <code><prefix_string|null></code> operand:
<ul>
<li><code>null</code>: create the file in the same directory and with the
same name conventions as other temporary files created by the Ghostscript
implementation on this platform. E.g., the temporary file might be named
<code>/tmp/gs_a1234</code>.
<p>
<li>A string that contains only alphanumeric characters, underline,
and dash: create the file in the standard temporary directory, but use
the
<code><prefix_string></code> as the first part of the file name.
E.g., if <code><prefix_string></code> is <code>xx</code>, the
temporary file might be named <code>/tmp/xxa1234</code>.
<p>
<li>A string that is the beginning of an absolute file name: use the
<code><prefix_string></code> as the first part of the file name.
E.g., if <code><prefix_string></code> is
<code>/my/tmpdir/zz</code>, the temporary file might be named
<code>/my/tmpdir/zza1234</code>.
<p>
When running in <code>SAFER</code> mode, the absolute path must
be one of the strings on the list given by the <code>PermitFileWriting</code>
userparameter. Temporary files created with <code>.tempfile</code> can
be deleted when in SAFER mode, and can be renamed to one of the paths
that is on <b>both</b> the PermitFileControl and PermitFileWriting
paths.
</ul>
</dl>
<dl>
<dt><code><file> <integer> .unread -</code>
<dd>Pushes back the last-read character onto the front of the file. If the
file is open only for writing, or if the integer argument is not the same
as the last character read from the file, causes an <code>ioerror</code>
error. May also cause an <code>ioerror</code> if the last operation on
the file was not a reading operation. This operator is now deprecated:
use <code>.peekstring</code> in new code.
</dl>
<p>
Ghostscript also supports the following <code>IODevice</code> in
addition to a subset of those defined in the Adobe documentation:
<ul>
<li>
<code>%pipe%command</code>, which opens a pipe on the given command.
This is supported only on operating systems that provide
<code>popen</code> (primarily Unix systems, and not all of those).
<p>
<li>
<code>%disk#%</code>, which emulates the %disk0
through %disk9 devices on some Adobe PostScript printers. This pseudo
device provides a flat filenaming system with a user definable location
for the files (/Root). These devices will only be present if the
diskn.dev feature is specified during the build.
<p>
This feature is intended to allow compatibility with font downloaders
that expect to store fonts on the %disk device of the printer.
<p>
Use of the %disk#% devices requires that the location of files be given
by the user setting the /Root device parameter. The syntax for setting
the /Root parameter is:<pre>
mark /Root (directory_specification) (%disk#) .putdevparams
</pre>
For example, to store the files of the %disk0 device on the directory
/tmp/disk0, use:<pre>
mark /Root (/tmp/disk0/) (%disk0) .putdevparams
</pre>
The files will be stored in the specified directory with arbitrary names.
A mapping file is used to store the association between the file
names given for the file operations on the %diskn# device and the file
that resides in the /Root directory.
</ul>
<h4><a name="Virtual_memory"></a>Virtual memory operators</h4>
<dl>
<dt><code><save> .forgetsave -</code>
<dd>Cancels the effect of a save, making it as though the save never
happened.
</dl>
<h4><a name="Miscellaneous"></a>Miscellaneous operators</h4>
<dl>
<dt><code><array> bind <array></code>
<dd>Depending on the command line parameters <code>bind</code> is redefined as:
</dl>
<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=bottom>
<th valign=bottom align=left>Flag
<td>
<th valign=bottom align=left>Definition
<tr> <td colspan=3><hr>
<tr valign=top> <td>NOBIND
<td>
<td>/bind {} def ;
no operation, returns the argument
<tr valign=top> <td>DELAYBIND
<td>
<td>returns the argument, stores the argument for later use by <code>.bindnow</code>
</table></blockquote>
<dl>
<dt><code><array> .bind <array></code>
<dd>Performs standard <code>bind</code> operation as defined in PLRM regardless of
NOBIND or DELAYBIND flags.
</dl>
<a name="bindnow"></a>
<dl>
<dt><code>- .bindnow -</code>
<dd>Applies <code>bind</code> operator to all savad procedures after binding has been
deferred through -dDELAYBIND. Note that idiom recognition has no effect for the deferred
binding because the value returned from <code>bind</code> is discarded.
<p>
Since v. 8.12 <code>.bindnow</code> undefines itself and restores standard definition of
<code>bind</code> operator. In earlier versions after calling <code>.bindnow</code>,
the postscript <code>bind</code> operator needs to be rebound to the internal implementation
<code>.bind</code>, as in this fragment from the ps2ascii script:
<blockquote><pre><tt>DELAYBIND {
.bindnow
/bind /.bind load def
} if
</tt></pre></blockquote>
This is necessary for correct behavior with later code that uses the <code>bind</code> operator.
</dl>
<dl>
<dt><code><obj1> <obj2> ... <objn> <n> .execn ...</code>
<dd>This executes <code>obj1</code> through <code>objn</code> in that
order, essentially equivalent to
<blockquote><pre>
<obj1> <obj2> ... <objn> <n> array astore {exec} forall
</pre></blockquote>
<p>
except that it doesn't actually create the array.
</dl>
<dl>
<dt><code><string> getenv <string> true</code>
<dt><code><string> getenv false</code>
<dd>Looks up a name in the shell environment. If the name is found,
returns the corresponding value and true; if the name is not found, returns
false.
</dl>
<dl>
<dt><code>.defaultpapersize <string> true</code>
<dt><code>.defaultpapersize false</code>
<dd>Get the system default paper size, which is usually
<code>a4</code> for countries using the metric system, and
<code>letter</code> for countries using the imperial system.
</dl>
<dl>
<dt><code><name> <array> .makeoperator <operator></code>
<dd>Constructs and returns a new operator that is actually the given
procedure in disguise. The name is only used for printing. The operator
has the executable attribute.
<p>
Operators defined in this way do one other thing besides running the
procedure: if an error occurs during the execution of the procedure, and
there has been no net reduction in operand or dictionary stack depth, the
operand or dictionary stack pointer respectively is reset to its position
at the beginning of the procedure.
</dl>
<dl>
<dt><code><string> <boolean> .setdebug -</code>
<dd>Sets or clears any subset of the debugging flags included in
<code><string></code> based on the value of
<code><boolean></code>. These correspond to the debug
flags set by <code>-Z</code> on the command line and enable
debug and tracing output from various internal modules.
<p>Note that most tracing output is only produced if the Ghostscript
interpreter was built with the <code>DEBUG</code> preprocessor
symbol defined.
<p>The <code>zsetdebug()</code> C function, which implements this
operator, is a useful breakpoint for debuggers.
Inserting '<code>() true .setdebug</code>' in the interpreted code will
trigger a breakpoint at that location without side effects. The
current flag state is available in C as the <code>gs_debug[]</code>
array, indexed by character value. The <code>zsetdebug</code> function will
be entered, and <code>gs_debug[]</code> updated, whether or not Ghostscript
is built with the <code>DEBUG</code> preprocessor symbol defined, so this
is useful even with release builds.
</dl>
<dl>
<dt><code>- .oserrno <errno></code>
<dd>Returns the error code for the most recent operating system error.
</dl>
<dl>
<dt><code>- .oserrorstring <string></code>
<dd>Returns the error string for the most recent operating system error.
</dl>
<a name="Runandhide"></a>
<dl>
<dt><code><array> <procedure> .runandhide ... <array></code>
<dd>Runs the <i><tt><procedure></tt></i> after removing the
<i><tt><array></tt></i> from the stack. As long as <i><tt><array></tt></i>
is not contained in any readable dictionaries or elsewhere on stacks, it
will not be accessible to <i><tt><procedure></tt></i>.
<p>
This operator is intended to allow hiding a <i><tt><save></tt></i> object
during execution of procedures or files that run in <b>SAFER</b> mode.
If a <code>save</code> is performed prior to entering <b>SAFER</b> mode
with <code>.setsafe</code>, using the save object as the operand to
<code>restore</code> will return to <b>NOSAFER</b> mode. In order to
prevent the procedures running in <b>SAFER</b> mode from being able to
return to <b>NOSAFER</b> mode, this operator should be used.
Upon return from the file or procedure <code>restore</code> can be used
to return to <b>NOSAFER</b> mode.
<p>
<b>Note:</b> The array operand hidden during the execution of the file or
procedure will be placed at the top of the operand stack which may be on
top of objects that the file or procedure leaves on top of the stack.
Thus removing objects below the array may be needed to prevent an
<code>invalidrestore</code> error.
<p>
For example, in order for a script or job server to execute a file
<tt>somefile.ps</tt> with the <b>SAFER</b> mode restrictions in place, returning
to unrestricted <b>NOSAFER</b> mode when the procedure exits is as follows:
<pre>
Start Ghostscript with <b>-dNOSAFER</b>
... % perform any device set up w/o restrictions
[ save ] % create a save object before SAFER
(somefile.ps) (r) file cvx % open the file to process
.setsafe % enter SAFER mode
.runandhide % run the file hiding the save object
count 1 roll % place array below anything left over
count 1 sub { pop } repeat % pop left over stuff
cleardictstack % prevent invalidrestore from dicts
0 get restore % go back to NOSAFER mode
</pre>
Another refinement on the above would be to execute <code>.runandhide</code>
using <code>stopped</code> in order to report errors but continue processing.
</dl>
<dl>
<dt><code>- .setsafe -</code>
<dd>If Ghostscript is started with <code>-dNOSAFER</code> or
<code>-dDELAYSAFER</code>, this operator can be used to enter <b>SAFER</b>
mode (see <a href="Use.htm#Safer"><b>-dSAFER</b></a>)
<p>
Since <b>SAFER</b> mode is implemented with userparameters and device parameters,
it is possible to use <code>save</code> and <code>restore</code> before
and after <code>.setsafe</code> to return to <b>NOSAFER</b> mode, but care
should be taken to ensure that the <i><tt>save</tt></i> object is not
accessible to any procedures or file run in <b>SAFER</b> mode (see
<a href="#Runandhide"><b>.runandhide</b></a> above).
<p>
<b>Note: This uses setpagedevice to change .LockSafetyParams, so the page
will be erased as a side effect of this operator</b>
</dl>
<dl>
<dt><code>- .locksafe -</code>
<dd>
This operator sets the current device's <code>.LockSafetyParams</code>
and the <code>LockFilePermissions</code> userparameter true as well as
adding the paths on LIBPATH and FONTPATH and the paths given by the
system params /GenericResourceDir and /FontResourceDir to the current
PermitFileReading list of paths.
<p>
If Ghostscript is started with <code>-dNOSAFER</code> or
<code>-dDELAYSAFER</code>, this operator can be used to enter <b>SAFER</b>
mode with the current set of <code>PermitFile...</code> user parameters
in effect. Since <code>.setsafe</code> sets the <code>PermitFile...</code>
user parameters to empty arrays, a script or job server that needs to
enable certain paths for file Reading, Writing and/or Control can use this
operator to perform the locking needed to enter <b>SAFER</b> mode.
<p>
For example, to enable reading everywhere, but disallow writing and file
control (deleting and renaming files), the following can be used:
<pre>
{ << /PermitFileReading [ (*) ]
/PermitFileWriting [ ]
/PermitFileControl [ ]
>> setuserparams
.locksafe
} stopped pop
</pre>
In the above example, use of stopped will allow the use of this sequence on
older versions of Ghostscript where <code>.locksafe</code> was not an operator.
<p>
<b>Note: This uses setpagedevice to change .LockSafetyParams, so the page
will be erased as a side effect of this operator</b>
<p>
See also <a href="#LockSafetyParams">.LockSafetyParams</a> and
<a href="#User_parameters">User Parameters</a>.
<p>
</dl>
<dl>
<dt><a name=".setpdfwrite"></a>
<code>.setpdfwrite</code></dt>
<dd>This operator conditions the environment for the <tt>pdfwrite</tt> output device.
It is a shorthand for setting parameters that have been deemed benificial. While not strictly necessary, it is usually helpful to set call this when using the pdfwrite device.
For example, this is how the ps2pdf script calls Ghostscript:
<blockquote><code>
gs -q -dSAFER -dNOPAUSE -dBATCH -sOutputFile=file.pdf </code><em>[more options]</em><code> \<br>
-sDEVICE=pdfwrite -c .setpdfwrite -f </code><em>source1.ps [more files]</em>
</blockquote>
<p>Currently, the operator just sets a minimum 3 MB vmthreshold to allow for
accumulating shared object data and to reduce the incidence of garbage
collection as a performance improvement. Additional settings may be added in the future.
</dl>
<dl>
<dt><code>.color_test</code> and <code>.color_test_all</code></dt>
<dd>These operators are used for the verification of device encode_color and
decode_color routines. They are for internal use only. Their function
can, and probably will, change as Artifex's requirements change.
<p>
<dd>Currently these operators loop through a set of possible values for the inputs
to the encode_color routine and then veify that the decode_color routines produce
values that match the input set to within a tolerance which is based upon the number
of bits used to encode a pixel. The operators also verify that if the device
is 'separable' then that the values produced by gx_default_encode_color and
gx_default_decode_color (the default encode/decode color handlers for a separable
device) are consistent to within the same tolerance.
</dl>
<h4><a name="Device"></a>Device operators</h4>
<dl>
<dt><code><device> copydevice <device></code>
<dd>Copies a device. The copy is writable and installable. The copy is
created in the current VM (local or global), usually local VM for executing
ordinary PostScript files.
</dl>
<dl>
<dt><code><devicename> finddevice <device></code>
<dd>Creates a default instance of a device specified by name. The instance
is created in global VM. If <code>finddevice</code> is called more than
once with the same device name, it creates the default instance the first
time, and returns the same instance thereafter.
</dl>
<dl>
<dt><code><devicename> findprotodevice <device></code>
<dd>Finds the prototype of a device specified by name. A prototype can be
used with <code>.getdeviceparams</code> or other parameter-reading
operators, but it is read-only and cannot be set with
<code>setdevice</code>: it must be copied first.
</dl>
<dl>
<dt><code><device> <x> <y> <width> <max_height> <alpha?> <std_depth|null> <string> .getbitsrect <height> <substring></code>
<dd>Reads a rectangle of rendered bits back from a device. This is only
guaranteed to be implemented for image devices (see below).
<code>alpha?</code> is 0 for no alpha, -1 for alpha first, 1 for alpha
last. <code>std_depth</code> is null for native pixels, number of bits
per component for a standard color space.
</dl>
<dl>
<dt><code><index> .getdevice <device></code>
<dd>Returns a device from the set of devices known to the system. The
first device is numbered 0. If the <code>index</code> is out of
range, a <code>rangecheck</code> error occurs.
This device is actually a prototype, not a directly usable device,
and is marked read-only; it cannot have its parameters changed or be
installed as the current device.
</dl>
<dl>
<dt><code>- .getdefaultdevice <device></code>
<dd>Returns the default device. This device is actually a prototype,
and is marked read-only; it cannot have its parameters changed or
be installed as the current device.
Use <code>copydevice</code> to obtain a workable device.
</dl>
<dl>
<dt><code><matrix> <width> <height> <palette> makeimagedevice <device></code>
<dd>Makes a new device that accumulates an image in memory. <code>
matrix</code> is the initial transformation matrix: it must be orthogonal
(that is, [a 0 0 b x y] or
[0 a b 0 x y]). <code>palette</code> is a
string of 2^<small><sup><b>N</b></sup></small> or
3 × 2^<small><sup><b>N</b></sup></small> elements,
specifying how the 2^<small><sup><b>N</b></sup></small> possible pixel
values will be interpreted. Each element is interpreted as a gray value,
or as RGB values, multiplied by 255. For example, if you want a monochrome
image for which 0=white and 1=black, the palette should be
<code><ff 00></code>; if you want a 3-bit deep image with
just the primary colors and their complements (ignoring the fact that 3-bit
images are not supported), the palette might be <code><000000 0000ff
00ff00 00ffff ff0000 ff00ff ffff00 ffffff></code>. At present, the
palette must contain exactly 2, 4, 16, or 256 entries, and must contain an
entry for black and an entry for white; if it contains any entries that
aren't black, white, or gray, it must contain at least the six primary
colors (red, green, blue, and their complements cyan, magenta, and yellow);
aside from this, its contents are arbitrary.
<p>
Alternatively, palette can be 16, 24, 32, or null (equivalent to 24).
These are interpreted as:
<blockquote><table cellpadding=0 cellspacing=0>
<tr valign=bottom>
<th valign=bottom align=left>Palette
<td>
<th valign=bottom align=left>Bits allocated per color
<tr> <td colspan=3><hr>
<tr valign=top> <td>16
<td>
<td>5 red, 6 green, 5 blue
<tr valign=top> <td>24
<td>
<td>8 red, 8 green, 8 blue
<tr valign=top> <td>32
<td>
<td>8C, 8M, 8Y, 8K
</table></blockquote>
<p>
Note that one can also make an image device (with the same palette as an
existing image device) by copying a device using the
<code>copydevice</code> operator.
</dl>
<dl>
<dt><code><matrix> <width> <height> <palette> <word?> makewordimagedevice <device></code>
<dd>Makes an image device as described above. <code>word?</code> is a
Boolean value indicating whether the data should be stored in a
word-oriented format internally. No ordinary PostScript programs should
use this operator.
</dl>
<dl>
<dt><code><device> <index> <string> copyscanlines <substring></code>
<dd>Copies one or more scan lines from an image device into a string,
starting at a given scan line in the image. The data is in the same format
as for the <code>image</code> operator. It is an error if the device is
not an image device or if the string is too small to hold at least one
complete scan line. Always copies an integral number of scan lines.
</dl>
<dl>
<dt><code><device> setdevice -</code>
<dd>
<p>
Sets the current device to the specified device. Also resets the
transformation and clipping path to the initial values for the device.
Signals an <code>invalidaccess</code> error if the device is a
prototype or if <a href="Language.htm#LockSafetyParams">.LockSafetyParams</a>
is true for the current device.
<p>
Some device properties may need to be set with <tt>putdeviceprops</tt> before
<code>setdevice</code> is called. For example, the pdfwrite device will try
to open its output file, causing an <tt>undefinedfilename</tt> error if
<code>OutputFile</code> hasn't been set to a valid filename. Another
method in such cases is to use the level 2 operator instead:
<code><< /OutputDevice /pdfwrite /OutputFile
(MyPDF.pdf) >> setpagedevice</code>.
</dl>
<dl>
<dt><code>- currentdevice <device></code>
<dd>Gets the current device from the graphics state.
</dl>
<dl>
<dt><code><device> getdeviceprops <mark> <name1> <value1> ... <namen> <valuen></code>
<dd>Gets the properties of a device. See the section on
<a href="#Device_parameters">device parameters</a> below for details.
</dl>
<dl>
<dt><code><mark> <name1> <value1> ... <namen> <valuen> <device> putdeviceprops <device></code>
<dd>Sets properties of a device. May cause <code>undefined</code>,
<code>invalidaccess</code>, <code>typecheck</code>, <code>rangecheck</code>, or
<code>limitcheck</code> errors.
</dl>
<dl>
<dt><code>- flushpage -</code>
<dd>On displays, flushes any buffered output, so that it is guaranteed to
show up on the screen; on printers, has no effect.
</dl>
<hr>
<h2><a name="Filters"></a>Filters</h2>
<h3><a name="Standard_filters"></a>Standard filters</h3>
<p>
In its usual configuration, Ghostscript supports all the standard PostScript
LanguageLevel 3 filters, both encoding and decoding, except that it does not
currently support:
<ul>
<li>the <code>EarlyChange</code> key in the <code>LZWEncode</code>
filter.
</ul>
<p>
Ghostscript also supports additional keys in the optional dictionary
operands for some filters. For the <code>LZWDecode</code> filter:
<dl>
<dt><code>InitialCodeLength <integer></code> (default 8)
<dd>An integer between 2 and 11 specifying the initial number of data bits
per code. Note that the actual initial code length is 1 greater than this,
to allow for the reset and end-of-data code values.
</dl>
<dl>
<dt><code>FirstBitLowOrder <boolean></code> (default false)
<dd>If true, codes appear with their low-order bit first.
</dl>
<dl>
<dt><code>BlockData <boolean></code> (default false)
<dd>If true, the data is broken into blocks in the manner specified for the
GIF file format.
</dl>
<p>
For the <code>CCITTFaxEncode</code> and <code>CCITTFaxDecode</code>
filters:
<dl>
<dt><code>DecodedByteAlign <integer></code> (default 1)
<dd>An integer <b>N</b> with the value 1, 2, 4, 8, or 16, specifying that
decoded data scan lines are always a multiple of <b>N</b> bytes. The
encoding filter skips data in each scan line from Columns to the next
multiple of <b>N</b> bytes; the decoding filter pads each scan line to a
multiple of <b>N</b> bytes.
</dl>
<h3><a name="Non_standard_filters"></a>Non-standard filters</h3>
<p>
In addition to the standard PostScript LanguageLevel 3 filters, Ghostscript
supports the following non-standard filters. Many of these filters are used
internally to implement standard filters or facilities; they are almost
certain to remain, in their present form or a backward-compatible one, in
future Ghostscript releases.
<dl>
<dt><code><target> /BCPEncode filter <file></code>
<dt><code><source> /BCPDecode filter <file></code>
<dd>Create filters that implement the Adobe Binary Communications Protocol.
See Adobe documentation for details.
</dl>
<dl>
<dt><code><target> <seed_integer> /eexecEncode filter <file></code>
<dd>Creates a filter for encrypting data into the encrypted format described
in the Adobe Type 1 Font Format documentation. The
<code>seed_integer</code> must be 55665 for the <code>eexec</code>
section of a font, or 4330 for a <code>CharString</code>. Note that for
the <code>eexec</code> section of a font, this filter produces binary
output and does not include the initial 4 (or <code>lenIV</code>) garbage
bytes.
</dl>
<dl>
<dt><code><source> <seed_integer> /eexecDecode filter <file></code>
<dt><code><source> <dict> /eexecDecode filter <file></code>
<dd>Creates a filter for decrypting data encrypted as described in the Adobe
Type 1 Font Format documentation. The <code>seed_integer</code> must be
55665 or 4330 as described just above. PDF interpreters don't skip space characters
after operator <code>eexec</code>. Use <code>keep_spaces = true</code> for
decoding embedded PDF fonts. Recognized dictionary keys are:
<blockquote>
<code>seed <16-bit integer></code> (required)<br>
<code>lenIV <non-negative integer></code> (default=4)<br>
<code>eexec <bool></code> (default=<code>false</code>)<br>
<code>keep_spaces <bool></code> (default=<code>false</code>)
</blockquote>
</dl>
<dl>
<dt><code><target> /MD5Encode filter <file></code>
<dd>Creates a filter that produces the 16-byte MD5 digest of the input.
Note that no output is produced until the filter is closed.
</dl>
<dl>
<dt><code><source> <hex_boolean> /PFBDecode filter <file></code>
<dd>Creates a filter that decodes data in <code>.PFB</code> format, the
usual semi-binary representation for Type 1 font files on IBM PC and
compatible systems. If <code>hex_boolean</code> is true, binary packets
are converted to hex; if false, binary packets are not converted.
</dl>
<dl>
<dt><code><target> <dict> /PixelDifferenceEncode filter <file></code>
<dt><code><source> <dict> /PixelDifferenceDecode filter <file></code>
<dd>Implements the Predictor=2 pixel-differencing option of the LZW
filters. Recognized keys are:
<blockquote>
<code>Colors <integer></code> (1 to 4, default=1)<br>
<code>BitsPerComponent <integer></code> (1, 2, 4, or 8, default=8)<br>
<code>Columns <integer></code> (>= 0, required)
</blockquote>
<p>
See the Adobe <a href="http://partners.adobe.com/public/developer/pdf/index_reference.html"><em>PDF Reference Manual</em></a> for details.
</dl>
<dl>
<dt><code><target> <dict> /PNGPredictorEncode filter <file></code>
<dt><code><source> <dict> /PNGPredictorDecode filter <file></code>
<dd>Implements the "filter" algorithms of the
<a href="http://www.libpng.org/pub/png/">Portable Network Graphics (PNG)
graphics format</a>. Recognized keys are:
<blockquote><table cellpadding=0 cellspacing=0>
<tr><th colspan=5 bgcolor="#CCCC00"><hr><font size="+1">Keys recognized in PNG filter algorithms</font><hr>
<tr valign=bottom>
<th align=left>Key
<td>
<th align=left>Range
<td>
<th align=left>Default
<tr> <td colspan=5><hr>
<tr valign=top> <td><code>Colors <integer></code>
<td>
<td>1 to 16
<td>
<td>16
<tr valign=top> <td><code>BitsPerComponent <integer></code>
<td>
<td>1, 2, 4, 8, or 16
<td>
<td>8
<tr valign=top> <td><code>Columns <integer></code>
<td>
<td>>= 0
<td>
<td>1
<tr valign=top> <td><code>Predictor <integer></code>
<td>
<td>10 to 15
<td>
<td>15
</table></blockquote>
<p>
The <code>Predictor</code> is the PNG algorithm number + 10 for the
<code>Encoding</code> filter; the <code>Decoding</code> filter
ignores <code>Predictor</code>. 15 means the encoder attempts to
optimize the choice of algorithm. For more details see the PNG
specification
<blockquote>
<a href="http://www.w3.org/TR/WD-png-960128.html">http://www.w3.org/TR/WD-png-960128.html</a>
</blockquote>
</dl>
<dl>
<dt><code><target> /TBCPEncode filter <file></code>
<dt><code><source> /TBCPDecode filter <file></code>
<dd>Create filters that implement the Adobe Tagged Binary Communications
Protocol. See Adobe documentation for details.
</dl>
<dl>
<dt><code><target> /zlibEncode filter <file></code>
<dt><code><source> /zlibDecode filter <file></code>
<dd>Creates filters that use the data compression method variously known as
'zlib' (the name of a popular library that implements it), 'Deflate' (as in
<a href="http://www.ietf.org/rfc/rfc1951.txt">RFC 1951</a>, which is a
detailed specification for the method), 'gzip' (the name of a popular
compression application that uses it), or 'Flate' (Adobe's name). Note that
the PostScript <code>Flate</code> filters are actually a combination of
this filter with an optional predictor filter.
</dl>
<h3><a name="Unstable_filters"></a>Unstable filters</h3>
<p>
Some versions of Ghostscript may also support other non-standard filters for
experimental purposes. The current version includes the following such
filters, which are not documented further. No code should assume that these
filters will exist in compatible form, or at all, in future versions.
<dl>
<dt><code><target/source> <string> ByteTranslateEncode/Decode filter <file></code>
<dd><code>string</code> must be a string of exactly 256 bytes. Creates a
filter that converts each input byte <em>b</em> to
<code>string</code>[<em>b</em>]. Note that the <code>Encode</code>
and <code>Decode</code> filters operate identically: the client must
provide a <code>string</code> for the <code>Decode</code> filter that
is the inverse mapping of the <code>string</code> for the
<code>Encode</code> filter.
</dl>
<dl>
<dt><code><target/source> <dict> BoundedHuffmanEncode/Decode filter <file></code>
<dd>These filters encode and decode data using Huffman codes. Since these
filters aren't used anywhere, we don't document them further, except to note
the recognized dictionary keys, which must be set identically for encoding
and decoding:
<blockquote>
<code>FirstBitLowOrder <bool></code> (default=false)<br>
<code>MaxCodeLength <int></code> (default=16)<br>
<code>EndOfData <bool></code> (default=true)<br>
<code>EncodeZeroRuns <int></code> (default=256)<br>
<code>Tables <int_array></code>
</blockquote>
</dl>
<dl>
<dt><code><target/source> <dict> BWBlockSortEncode/Decode filter <file></code>
<dd>This filter implements the Burroughs-Wheeler block sorting compression
method, which we've heard is also used in the popular <code>bzip2</code>
compression application. See <a
href="http://sources.redhat.com/bzip2/">http://sources.redhat.com/bzip2/</a>
for more information. The only recognized dictionary key is:
<blockquote>
<code>BlockSize <integer></code> (default=16384)
</blockquote>
</dl>
<dl>
<dt><code><target/source> MoveToFrontEncode/Decode filter <file></code>
<dd>The <code>Encode</code> filter starts by initializing an internal
256-byte array <code>a</code> to the values 0 .. 255. This array will
always hold a permutation of these values. Then for each input byte
<em>b</em>, the filter outputs the index <em>i</em> such that
<code>a</code>[<em>i</em>] = <em>b</em>, and moves that element to the
front (element 0) of <code>a</code>, moving elements 0 .. <em>i-1</em> to
positions 1 .. <em>i</em>. The <code>Decode</code> filter inverts this
process.
</dl>
<hr>
<h2><a name="Device_parameters"></a>Device parameters</h2>
Ghostscript supports the concept of device parameters for all devices, not
just page devices. (For non-page devices, these are accessible through
<code>getdeviceprops</code> and <code>putdeviceprops</code>, as
indicated above.) Here are the currently defined parameters for all
devices:
<dl>
<dt><a name="LockSafetyParams"></a>
<code>.LockSafetyParams <boolean></code>
<dd>This parameter allows for improved system security by preventing
PostScript programs from being able to change potentially dangerous
device paramters such as OutputFile. This parameter cannot be set false
if it is already true.
<p>
If this parameter is true for the current device, attempt to set a new
device that has <code>.LockSafetyParams</code> false will signal an
<code> invalidaccess</code> error.
</dl>
<dl>
<dt><code>BitsPerPixel <integer> (usually read-only)</code>
<dd>Number of bits per pixel.
</dl>
<dl>
<dt><code>.HWMargins [<four floats>]</code>
<dd>Size of non-imageable regions around the edges of the page, in points
(units of 1/72in; see the <a href="Devices.htm#Measurements">notes on
measurements</a> in the documentation on devices).
</dl>
<dl>
<dt><code>HWSize [<integer> <integer>]</code>
<dd>X and Y size in pixels.
</dl>
<dl>
<dt><code>%MediaSource <integer></code>
<dd>The input tray key as determined by setpagedevice. PostScript
language programs don't set this parameter directly; they can
<em>request</em> a particular tray through the MediaPosition
setpagedevice parameter, but the setpagedevice logic need not
necessarily honor the request. Devices which support switchable trays
should implement %MediaSource in their put_params device procedure,
but (unlike most other such parameters) need not implement
corresponding reading logic in get_params.
</dl>
<dl>
<dt><code>%MediaDestination <integer></code>
<dd>The output tray key as determined by setpagedevice. Handling by
devices should be parallel to %MediaSource.
</dl>
<dl>
<dt><code>.IgnoreNumCopies <boolean></code>
<dd>Some page description languages support a NumCopies parameter.
This parameter instructs the device to ignore this, producing only
one copy of the document on output. Note that some devices ignore
NumCopies regardless because of limitation of the output format
or the implementation.
</dl>
<dl>
<dt><code>Name <string> (read-only)</code>
<dd>The device name. Currently the same as <code>OutputDevice</code>.
</dl>
<dl>
<dt><code>Colors, GrayValues, RedValues, GreenValues, BlueValues, ColorValues (usually read-only)</code>
<dd>As for the <code>deviceinfo</code> operator of Display PostScript.
<code>Red</code>, <code>Green</code>, <code>Blue</code>, and
<code>ColorValues</code> are only defined if
<code>Colors</code> > 1.
</dl>
<dl>
<dt><code>TextAlphaBits, GraphicsAlphaBits (usually read-only)</code>
<dd>The number of bits of anti-aliasing information for text or graphics
respectively. Legal values are 1 (no anti-aliasing, the default for most
devices), 2, or 4.
</dl>
<p>
Ghostscript also supports the following read-only parameter that is not a
true device parameter:
<dl>
<dt><code>.EmbedFontObjects <integer></code>
<dd>If non-zero, indicates that the device may embed font objects (as
opposed to bitmaps for individual characters) in the output. The purpose of
this parameter is to disable third-party font renderers for such devices.
(This is zero for almost all devices.)
</dl>
<p>
In addition, the following are defined per Adobe's documentation for the
<code>setpagedevice</code> operator:
<blockquote>
<code>Duplex</code> (if supported)<br>
<code>HWResolution</code><br>
<code>ImagingBBox</code><br>
<code>Margins</code><br>
<code>LeadingEdge</code><br>
<code>MediaPosition</code><br>
<code>NumCopies</code> (for printers only)<br>
<code>Orientation</code> (if supported)<br>
<code>OutputDevice</code><br>
<code>PageOffset</code> (write-only)<br>
<code>PageSize</code><br>
<code>ProcessColorModel</code> (usually read-only)<br>
</blockquote>
<p>
Some devices may only allow certain values for <code>HWResolution</code>
and <code>PageSize</code>. The null device ignores attempts to set
<code>PageSize</code>; its size is always <code>[0 0]</code>.
<p>
It should be noted that calling <tt>setpagedevice</tt> with one of the above keys may reset the effects of any <code>pdfmark</code> commands up to that point. In particular this is true of HWResolution, a behavior that differs from Adobe Distiller.
<a name="Banding_parameters"></a>
<p><b>
For raster printers and image format (jpeg*, tiff*, png* ...) devices these
page device parameters are also defined:</b>
<dl>
<dt><code>MaxBitmap <integer></code>
<dd>Maximum space for a full page raster image (bitmap) in memory.
<p>This value includes the space for padding raster lines and for an array of
pointers for each raster line, thus the <code>MaxBitmap</code> value to allow
a given PageSize of a specific number of bits per pixel to be rendered in a
full page buffer may be somewhat larger than the bitmap size alone.
</dl>
<dl>
<dt><code>BandListStorage <file|memory></code>
<dd>The default is determined by the make file macro <code>BAND_LIST_STORAGE</code>.
Since <code>memory</code> is always included, specifying <code>-sBandListStorage=memory</code>
when the default is <code>file</code> will use memory based storage for the
band list of the page. This is primarily intended for testing, but if the disk I/O is
slow, band list storage in memory may be faster.
</dl>
<dl>
<dt><code>BufferSpace <integer></code>
<dd>Size of the buffer space for band lists, if the full page raster image
(bitmap) is larger than <code>MaxBitmap</code> (see above.)
<p>The buffer space is used to collect display list (clist) commands for the
bands and then to consolidate those commands when writing the clist to the
selected BAND_LIST_STORAGE device (memory or file) set when Ghostscript is compiled.
<p>If <code>MaxBitmap</code> (above) forces banding mode, and if <code>BufferSpace</code>
is large enough, the display list (clist) will consist of a single band.
<p>The <code>BufferSpace</code> will determine the size of the 'consolidation'
buffer (above) even if the <code>MaxBitmap</code> value is low enough to force
banding/clist mode.
</dl>
<dl>
<dt><code>BGPrint <boolean></code>
<dd>With many printer devices, when the display list (clist) banding mode is being used,
the page rendering and output can be performed in a background thread.
The default value, <code>false</code>, causes the rendering and printing to be
done in the same thread as the parser. When <code>-dBGPrint=true</code>, the
page output will be overlapped with parsing and writing the clist for the next page.
<p>If the device does not support background printing, rendering and printing will
be performed as if <code>-dBGPrint=false</code>.
<p>Note that the background printing thread will allocate a band buffer (size determined
by the <code>BufferSpace</code> or <code>BandBufferSpace</code> values) in addition to
the band buffer in the 'main' parsing thread.
<p>If <code>NumRenderingThreads</code> is > 0, then the background printing thread
will use the specified number of rendering threads as children of the background printing
thread. The background printing thread will perform any processing of the raster data
delivered by the rendering threads.
</dl>
<dl>
<dt><code>GrayDetection <boolean></code>
<dd>When <code>true</code>, and when the display list (clist) banding mode is being used,
during writing of the clist, the color processing logic collects information about the
colors used <b>before</b> the device color profile is applied. This allows special devices
that examine <code>pageneutralcolor</code> with the information that all colors on
the page are near <i>neutral</i>, i.e. monochrome, and converting the rendered raster to
gray may be used to reduce the use of color toners/inks.
<p>
Since the determination of whether or not the page uses colors is determined before the
conversion to device colors, this information is independent of the device output profile.
<p>
Among the devices distributed with the source, currently only the <code>pnmcmyk</code>
device supports this parameter and will produce either a <code>P7 PAM</code> CMYK output
or a <code>P5 PGM</code> Gray output depending on the use of color on the page.
<p>
<b>Note:</b> Since <code>-dGrayDetection=true</code> requires extra checking during
writing of the clist, this option should <b>only</b> be used for devices that
support the optimization of pages to monochrome, otherwise performance may be degraded
for no benefit.
</dl>
<dl>
<dt><code>NumRenderingThreads <integer></code>
<dd>When the display list (clist) banding mode is being used, bands can be rendered
in separate threads. The default value, 0, causes the rendering of bands to be
done in the same thread as the parser and device driver. <code>NumRenderingThreads</code>
of 1 or higher results in bands rendering in the specified number of 'background'
threads.
<p>The number of threads should generally be set to the number of available
processor cores for best throughput.
<p>Note that each thread will allocate a band buffer (size determined by the
<code>BufferSpace</code> or <code>BandBufferSpace</code> values) in addition to
the band buffer in the 'main' thread.
</dl>
<dl>
<dt><code>OutputFile <string></code>
<dd>An empty string means "send to printer directly", otherwise specifies
the file name for output; <code>%d</code> is replaced by the page number
for page-oriented output devices;
on Unix systems <code>%pipe%</code><em>command</em> writes to a pipe.
(<code>|</code><em>command</em> also writes to a pipe, but is now
deprecated). Also see the <code>-o</code> parameter.
<p>
Attempts to set this parameter if <code>.LockSafetyParams</code> is true
will signal an <code>invalidaccess</code> error.
</dl>
<dl>
<dt><code>OpenOutputFile <boolean></code>
<dd>If true, open the device's output file when the device is opened,
rather than waiting until the first page is ready to print.
</dl>
<dl>
<dt><code>PageCount <integer> (read-only)</code>
<dd>Counts the number of pages printed on the device.
</dl>
<p>
The following parameters are for use only by very specialized applications
that separate band construction from band rasterization. <b>Improper use may
cause unpredictable errors.</b> In particular, if you only want to allocate
more memory for banding, to increase band size and improve performance, use
the <code>BufferSpace</code> parameter, not <code>BandBufferSpace</code>.
<dl>
<dt><code>BandHeight <integer></code>
<dd>The height of bands when banding. 0 means use the largest band height
that will fit within the <code>BandBufferSpace</code> (or <code>BufferSpace</code>,
if <code>BandBufferSpace</code> is not specified). If <code>BandHeight</code>
is larger than the number of lines that will fit in the buffer, opening the device will fail.
</dl>
<dl>
<dt><code>BandWidth <integer></code>
<dd>The width of bands in the rasterizing pass, in pixels. 0 means use the
actual page width. A BandWidth value smaller than the width of the page
will be ignored, and the actual page width will be used instead.
</dl>
<dl>
<dt><code>BandBufferSpace <integer></code>
<dd>The size of the band buffer in the rasterizing pass, in bytes. 0 means
use the same buffer size as for the interpretation pass.
</dl>
<p>
Ghostscript supports the following parameter for
<code>setpagedevice</code> and <code>currentpagedevice</code> that is
not a device parameter per se:
<dl>
<dt><code>ViewerPreProcess <procedure></code>
<dd>Specifies a procedure to be applied to the page device dictionary
before any other processing is done. The procedure may not alter the
dictionary, but it may return a modified copy. This "hook" is provided for
use by viewing programs such as GSview.
</dl>
<hr>
<h2><a name="User_parameters"></a>User parameters</h2>
Ghostscript supports the following non-standard user parameters:
<dl>
<dt><code>ProcessDSCComment <procedure|null></code>
<dd>If not null, this procedure is called whenever the scanner detects a DSC
comment (comment beginning with <code>%%</code> or <code>%!</code>).
There are two operands, the file and the comment (minus any terminating
EOL), which the procedure must consume.
</dl>
<dl>
<dt><code>ProcessComment <procedure|null></code>
<dd>If not null, this procedure is called whenever the scanner detects a
comment (or, if <code>ProcessDSCComment</code> is also not null, a
comment other than a DSC comment). The operands are the same as for
<code>ProcessDSCComment</code>.
</dl>
<dl>
<dt><code>LockFilePermissions <boolean></code>
<dd>If <tt>true</tt>, this parameter and the three <tt>PermitFile...</tt>
parameters cannot be changed. Attempts to change any of the values
when LockFilePermissions is <tt>true</tt> will signal <code>invalidaccess</code>.
Also, when this value is <tt>true</tt>, the <code>file</code> operator
will give <code>invalidaccess</code> when attempting to open files
(processes) using the <code>%pipe</code> device.
<p>
Also when <code>LockFilePermissions</code> is <tt>true</tt>, strings
cannot reference the parent directory (platform specific). For example
<code>(../../xyz)</code> is illegal on unix, Windows
and Macintosh, and <code>([.#.#.XYZ])</code> is illegal on VMS.
<p>
This parameter is set <tt>true</tt> by the <code>.setsafe</code> and
<code>.locksafe</code> operators.
</dl>
<dl>
<dt><code>PermitFileReading <array of strings></code>
<dt><code>PermitFileWriting <array of strings></code>
<dt><code>PermitFileControl <array of strings></code>
<dd>These parameters specify paths where file reading, writing and the
'control' operations are permitted, respectively. File control
operations are <code>deletefile</code> and <code>renamefile</code>.
For <code>renamefile</code>, the filename for the current filename
must match one of the paths on the PermitFileControl list, and the
new filename must be on <b>both</b> the PermitFileControl and the
PermitFileWriting lists of paths.
<p>
The strings can contain wildcard characters as for the <code>filenameforall</code>
operator and unless specifying a single file, will end with a <b>*</b>
for directories (folders) to allow access to all files and sub-directories
in that directory.
<p>
<b>Note:</b> The strings are used for stringmatch operations similar
to <code>filenameforall</code>, thus on MS Windows platforms, use the '/'
character to separate directories and filenames or use '\\\\' to
have the string contain '\\' which will match a single '\' in the
target filename (use of '/' is strongly recommended).
<p>
The <a href="Use.htm#Safer"><b>SAFER</b></a> mode and the
<code>.setsafe</code> operator set all three lists to empty arrays,
thus the only files that can be read are the <code>%stdin</code> device and
on LIBPATH or FONTPATH or the Resource paths specified by the /FontResourceDir
or /GenericResourceDir system params. Files cannot be opened for writing
anywhere and cannot be deleted or renamed except for files created with the
<a href="#Tempfile"><b>.tempfile</b></a> operator).
<p>
<b>Note: </b>Limiting file reading as above is <b>NOT</b> compatible with
SAFER mode in release versions before 7.11 and corresponds to the use of
<code>-dPARANOIDSAFER</code> in version 7.04 (up to and not including
version 7.10) and GPL versions 6.53 (up to and not including 6.60).
</dl>
<dl>
<dt><code>AlignToPixels <integer></code>
<dd>Control sub-pixel positioning of character glyphs (where
applicable). A value of 1 specifies alignment of text characters to
pixels boundaries. A value of 0 to subpixels where the division factor
is set by the device parameter <code>TextAlphaBits</code>. If the
latter is 1, the same rendering results regardless of the value of
<code>AlignToPixels</code>. The initial value defaults to 1, but this
may be overridden by the command line argument
<code>-dAlignToPixels</code>.
</dl>
<dl>
<dt><a name="GridFitTT"></a>
<code>GridFitTT <integer></code>
<dd>Control the use of True Type grid fitting.
Ghostscript, by default, uses Freetype for rendering Truetype (and most other) glyphs
(but other scaler/renderer libraries can be used), thus has access to a complete Truetype
bytecode interpreter.
<p>
This parameter controls the hinting of Truetype glyphs.
<ul>
<li>
A value of 0 disables grid fitting for all True Type fonts (not generally recommended).
</li>
<li>
A value of 1 enables the grid fitting using the native Truetype hinting bytecode
program(s). Fonts or glyphs with faulty bytecode program(s) will be rendered unhinted.
</li>
<li>
A value 2 is scaler/renderer dependent (generally, if no alternative hinting engine is available
this will be equivalent to 1). With the Freetype (our default) this enables Freetype's built-in
autohinter.
</li>
<li>
With Freetype, a value of 3 is effectively equivalent to 1.
</li>
</ul>
<p>
This parameter defaults to 1, but this
may be overridden on the command line with
<code>-dGridFitTT=n</code>.
<p>
</dl>
<hr>
<h2><a name="Miscellaneous_additions"></a>Miscellaneous additions</h2>
<h3><a name="Extended_semantics_of_run"></a>Extended semantics of 'run'</h3>
<p>
The operator <code>run</code> can take either a string or a file as its argument. In
the latter case, it just runs the file, closing it at the end, and trapping
errors just as for the string case.
<h3><a name="DecodingResources"></a>Decoding resources</h3>
<p>
<code>Decoding</code> is a Ghostscript-specific resource category. It contains
various resources for emulating PostScript fonts with other font technologies.
Instances of the <tt>Decoding</tt> category are tables which map PostScript glyph
names to character codes used with TrueType, Intellifont, Microtype and other font formats.
<p>
Currently Ghostscript is capable of PostScript font emulation in 2 ways :
<ul>
<li>
1. Through <a href="./Use.htm#FAPI_run">FAPI</a> plugins, and
</li>
<li>
2. With TrueType font files, using the native font renderer, by
specifying TrueType font names or files in <a href="../lib/Fontmap">lib/Fontmap</a>.
</li>
</ul>
<p>
<code>Decoding</code> resources are not currently used by the native font renderer.
<p>
An instance of the <code>Decoding</code> resource category is
a dictionary. The dictionary keys are PostScript glyph names and the
values are either character codes, or arrays of character codes.
Arrays are used when a single name may be mapped to various character codes -
in this case Ghostscript tries all alternatives until a success.
The name of the resource instance should
reflect the character set for which it maps. For example,
<code>/Unicode</code> <code>/Decoding</code> resource maps to
Unicode UTF-16.
<p>
The rules for using <code>Decoding</code> resources in particular
cases are specified in the configuration file
<a href="../lib/xlatmap">lib/xlatmap</a>. See the file itself for more
information.
<p>
The file format for <code>Decoding</code> resource files is
generic PostScript.
Users may want to define custom <code>Decoding</code> resources.
The <code>ParseDecoding</code> procset defined in
<a href="../Resource/Init/gs_ciddc.ps">Resource/Init/gs_ciddc.ps</a> allows representation
of the table in a comfortable form.
<h3><a name="CIDDecodingResources"></a>CIDDecoding resources</h3>
<p>
<code>CIDDecoding</code> resources are similar to <code>Decoding</code>
resources, except they map Character Identifiers (CIDs) rather than glyph names.
Another difference is that the native Ghostscript font renderer uses
<code>CIDDecoding</code> resources while emulate CID fonts with TrueType or OpenType fonts.
<p>
An instance of the <code>CIDDecoding</code> resource category is
a dictionary of arrays. Keys in the dictionary are integers,
which correspond to high order byte of a CID.
Values are 256-element arrays, and their indices correspond to the low order byte of a CID.
Each elemet of an array is either null, or character code (integer), or an array
of character codes (integers). The zero code represents mapping to the default character.
<p>
The dictionary includes the additional key <code>CIDCount</code>.
Its value is the maximal CID defined, plus one.
<p>
The Ghostscript library is capable of generating some <code>CIDDecoding</code>
instances automatically, using the appropriate <code>CMap</code> (character map)
resources. This covers most of practical cases if the neccessary <code>CMap</code>
resources are provided. See the table <code>.CMapChooser</code> in
<a href="../Resource/Init/gs_ciddc.ps">Resource/Init/gs_ciddc.ps</a>
for the names of automatically gerenated resources and associated <code>CMap</code>s.
They allow to mapping CNS1, GB1, Japan1, Japan2 and Korea1 CID sets to TrueType
character sets known as Unicode (exactly UTF-16), Big5,
GB1213, ShiftJIS, Johab and Wansung.
<p>
The file format for <code>CIDDecoding</code> resource file is
generic PostScript.
Users may want to define custom resources to <code>CIDDecoding</code>
resource category.
<h3><a name="GlyphNames2Unicode"></a>GlyphNames2Unicode</h3>
<p>
<code>GlyphNames2Unicode</code> is an undocumented dictionary which Adobe
PostScript printer driver uses to communicate with Adobe Distiller.
In this dictionary the keys are glyph names, the values are Unicode UTF-16 codes for them.
The dictionaly is stored in the <code>FontInfo</code> dictionary under
the key <code>GlyphNames2Unicode</code>. Ghostscript recognises it and uses
to generate <code>ToUnicode</code> CMaps with pdfwrite.
<p>
<h3><a name="MultipleResourceDirectories"></a>Multiple Resource directories</h3>
<p>
Since 8.10 release Ghostscript maintains multiple resource directories.
<p>
Ghostscript does not distinguish <code>lib</code> and
<code>Resource</code> directories.
There is no file name conflicts because <code>lib</code> does not
contain subdirectories, but <code>Resource</code>
always store files in subdirectories.
<p>
The search method with multiple resource directories
appears not fully conforming to PLRM. We cannot unconditionally call
<code>ResourceFileName</code> while executing <code>findresource</code>
or <code>resourcestatus</code>, <code>resourceforall</code>, because
per PLRM it always returns a single path. Therefore Ghostscript
implements an extended search method in <code>findresource</code>,
<code>resourcestatus</code> and <code>resourceforall</code>, which
first calls <code>ResourceFileName</code> and checks whether the
returned path points to an existing file. If yes, the file is used,
othervise Ghostscript searches all directories specified in
<code>LIB_PATH</code>. With a single resource directory
it appears conforming to PLRM and equivalent to Adobe implementations.
<p>
<code>ResourceFileName</code> may be used for obtaining a path
where a resource file to be installed. In this case
Ghostscript to be invoked with <code>-sGenericResourceDir=path</code>,
specifying an absolute path. The default value for
<code>GenericResourceDir</code> is a relative path. Therefore
a default invocation with a PostScript installer
will install resource files into <code>/gs/Resource</code>.
<p>
<!-- [2.0 end contents] ==================================================== -->
<!-- [3.0 begin visible trailer] =========================================== -->
<hr>
<p>
<small>Copyright © 2000-2006 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 9.14, 26 March 2014
<!-- [3.0 end visible trailer] ============================================= -->
</body>
</html>
distrib
>
Mageia
>
5
>
x86_64
>
media
>
core-updates
>
by-pkgid
>
0764ff2a42cce6d799de9aa7a1f29305
>
files
>
32
