1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
|
% Copyright 2006 by Till Tantau
%
% This file may be distributed and/or modified
%
% 1. under the LaTeX Project Public License and/or
% 2. under the GNU Free Documentation License.
%
% See the file doc/generic/pgf/licenses/LICENSE for more details.
\section{The Soft Path Subsystem}
\label{section-soft-paths}
\makeatletter
This section describes a set of commands for creating \emph{soft paths} as
opposed to the commands of the previous section, which created \emph{hard
paths}. A soft path is a path that can still be ``changed'' or ``molded''. Once
you (or the \pgfname\ system) is satisfied with a soft path, it is turned into
a hard path, which can be inserted into the resulting |.pdf| or |.ps| file.
Note that the commands described in this section are ``high-level'' in the
sense that they are not implemented in driver files, but rather directly by the
\pgfname-system layer. For this reason, the commands for creating soft paths do
not start with |\pgfsys@|, but rather with |\pgfsyssoftpath@|. On the other
hand, as a user you will never use these commands directly, they are described
as part of the low-level interface.
\subsection{Path Creation Process}
When the user writes a command like |\draw (0bp,0bp) -- (10bp,0bp);| quite a
lot happens behind the scenes:
%
\begin{enumerate}
\item The frontend command is translated by \tikzname\ into commands of the
basic layer. In essence, the command is translated to something like
%
\begin{codeexample}[code only]
\pgfpathmoveto{\pgfpoint{0bp}{0bp}}
\pgfpathlineto{\pgfpoint{10bp}{0bp}}
\pgfusepath{stroke}
\end{codeexample}
%
\item The |\pgfpathxxxx| commands do \emph{not} directly call ``hard''
commands like |\pgfsys@xxxx|. Instead, the command |\pgfpathmoveto|
invokes a special command called |\pgfsyssoftpath@moveto| and
|\pgfpathlineto| invokes |\pgfsyssoftpath@lineto|.
The |\pgfsyssoftpath@xxxx| commands, which are described below,
construct a soft path. Each time such a command is used, special tokens
are added to the end of an internal macro that stores the soft path
currently being constructed.
\item When the |\pgfusepath| is encountered, the soft path stored in the
internal macro is ``invoked''. Only now does a special macro iterate
over the soft path. For each line-to or move-to operation on this path
it calls an appropriate |\pgfsys@moveto| or |\pgfsys@lineto| in order
to, finally, create the desired hard path, namely, the string of
literals in the |.pdf| or |.ps| file.
\item After the path has been invoked, |\pgfsys@stroke| is called to insert
the literal for stroking the path.
\end{enumerate}
Why such a complicated process? Why not have |\pgfpathlineto| directly call
|\pgfsys@lineto| and be done with it? There are two reasons:
%
\begin{enumerate}
\item The \pdf\ specification requires that a path is not interrupted by
any non-path-construction commands. Thus, the following code will
result in a corrupted |.pdf|:
%
\begin{codeexample}[code only]
\pgfsys@moveto{0}{0}
\pgfsys@setlinewidth{1}
\pgfsys@lineto{10}{0}
\pgfsys@stroke
\end{codeexample}
%
Such corrupt code is \emph{tolerated} by most viewers, but not always.
It is much better to create only (reasonably) legal code.
\item A soft path can still be changed, while a hard path is fixed. For
example, one can still change the starting and end points of a soft
path or do optimizations on it. Such transformations are not possible
on hard paths.
\end{enumerate}
\subsection{Starting and Ending a Soft Path}
No special action must be taken in order to start the creation of a soft path.
Rather, each time a command like |\pgfsyssoftpath@lineto| is called, a special
token is added to the (global) current soft path being constructed.
However, you can access and change the current soft path. In this way, it is
possible to store a soft path, to manipulate it, or to invoke it.
\begin{command}{\pgfsyssoftpath@getcurrentpath\marg{macro name}}
This command will store the current soft path in \meta{macro name}.
\end{command}
\begin{command}{\pgfsyssoftpath@setcurrentpath\marg{macro name}}
This command will set the current soft path to be the path stored in
\meta{macro name}. This macro should store a path that has previously been
extracted using the |\pgfsyssoftpath@getcurrentpath| command and has
possibly been modified subsequently.
\end{command}
\begin{command}{\pgfsyssoftpath@invokecurrentpath}
This command will turn the current soft path in a ``hard'' path. To do so,
it iterates over the soft path and calls an appropriate |\pgfsys@xxxx|
command for each element of the path. Note that the current soft path is
\emph{not changed} by this command. Thus, in order to start a new soft path
after the old one has been invoked and is no longer needed, you need to set
the current soft path to be empty. This may seem strange, but it is often
useful to immediately use the last soft path again.
\end{command}
\begin{command}{\pgfsyssoftpath@flushcurrentpath}
This command will invoke the current soft path and then set it to be empty.
\end{command}
\subsection{Soft Path Creation Commands}
\begin{command}{\pgfsyssoftpath@moveto\marg{x}\marg{y}}
This command appends a ``move-to'' segment to the current soft path. The
coordinates \meta{x} and \meta{y} are given as normal \TeX\ dimensions.
\example One way to draw a line:
%
\begin{codeexample}[code only]
\pgfsyssoftpath@moveto{0pt}{0pt}
\pgfsyssoftpath@lineto{10pt}{10pt}
\pgfsyssoftpath@flushcurrentpath
\pgfsys@stroke
\end{codeexample}
%
\end{command}
\begin{command}{\pgfsyssoftpath@lineto\marg{x}\marg{y}}
Appends a ``line-to'' segment to the current soft path.
\end{command}
\begin{command}{\pgfsyssoftpath@curveto\marg{a}\marg{b}\marg{c}\marg{d}\marg{x}\marg{y}}
Appends a ``curve-to'' segment to the current soft path with controls
$(a,b)$ and $(c,d)$.
\end{command}
\begin{command}{\pgfsyssoftpath@rect\marg{lower left x}\marg{lower left y}\marg{width}\marg{height}}
Appends a rectangle segment to the current soft path.
\end{command}
\begin{command}{\pgfsyssoftpath@closepath}
Appends a ``close-path'' segment to the current soft path.
\end{command}
\subsection{The Soft Path Data Structure}
A soft path is stored in a standardized way, which makes it possible to modify
it before it becomes ``hard''. Basically, a soft path is a long sequence of
triples. Each triple starts with a \emph{token} that identifies what is going
on. This token is followed by two dimensions in braces. For example, the
following is a soft path that means ``the path starts at $(0\mathrm{bp},
0\mathrm{bp})$ and then continues in a straight line to $(10\mathrm{bp},
0\mathrm{bp})$''.
%
\begin{codeexample}[code only]
\pgfsyssoftpath@movetotoken{0bp}{0bp}\pgfsyssoftpath@linetotoken{10bp}{0bp}
\end{codeexample}
A curve-to is hard to express in this way since we need six numbers to express
it, not two. For this reasons, a curve-to is expressed using three triples as
follows: The command
%
\begin{codeexample}[code only]
\pgfsyssoftpath@curveto{1bp}{2bp}{3bp}{4bp}{5bp}{6bp}
\end{codeexample}
%
\noindent results in the following three triples:
%
\begin{codeexample}[code only]
\pgfsyssoftpath@curvetosupportatoken{1bp}{2bp}
\pgfsyssoftpath@curvetosupportbtoken{3bp}{4bp}
\pgfsyssoftpath@curvetotoken{5bp}{6bp}
\end{codeexample}
These three triples must always ``remain together''. Thus, a lonely
|supportbtoken| is forbidden.
In details, the following tokens exist:
%
\begin{itemize}
\item \declare{|\pgfsyssoftpath@movetotoken|} indicates a move-to
operation. The two following numbers indicate the position to which the
current point should be moved.
\item \declare{|\pgfsyssoftpath@linetotoken|} indicates a line-to
operation.
\item \declare{|\pgfsyssoftpath@curvetosupportatoken|} indicates the first
control point of a curve-to operation. The triple must be followed by a
|\pgfsyssoftpath@curvetosupportbtoken|.
\item \declare{|\pgfsyssoftpath@curvetosupportbtoken|} indicates the second
control point of a curve-to operation. The triple must be followed by a
|\pgfsyssoftpath@curvetotoken|.
\item \declare{|\pgfsyssoftpath@curvetotoken|} indicates the target of a
curve-to operation.
\item \declare{|\pgfsyssoftpath@rectcornertoken|} indicates the corner of a
rectangle on the soft path. The triple must be followed by a
|\pgfsyssoftpath@rectsizetoken|.
\item \declare{|\pgfsyssoftpath@rectsizetoken|} indicates the size of a
rectangle on the soft path.
\item \declare{|\pgfsyssoftpath@closepath|} indicates that the subpath
begun with the last move-to operation should be closed. The parameter
numbers are currently not important, but if set to anything different
from |{0pt}{0pt}|, they should be set to the coordinate of the original
move-to operation to which the path ``returns'' now.
\end{itemize}
|
