1 |
578 |
markom |
'\"
|
2 |
|
|
'\" Copyright (c) 1993 The Regents of the University of California.
|
3 |
|
|
'\" Copyright (c) 1994-1997 Sun Microsystems, Inc.
|
4 |
|
|
'\"
|
5 |
|
|
'\" See the file "license.terms" for information on usage and redistribution
|
6 |
|
|
'\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
|
7 |
|
|
'\"
|
8 |
|
|
'\" RCS: @(#) $Id: expr.n,v 1.1.1.1 2002-01-16 10:25:24 markom Exp $
|
9 |
|
|
'\"
|
10 |
|
|
.so man.macros
|
11 |
|
|
.TH expr n 8.0 Tcl "Tcl Built-In Commands"
|
12 |
|
|
.BS
|
13 |
|
|
'\" Note: do not modify the .SH NAME line immediately below!
|
14 |
|
|
.SH NAME
|
15 |
|
|
expr \- Evaluate an expression
|
16 |
|
|
.SH SYNOPSIS
|
17 |
|
|
\fBexpr \fIarg \fR?\fIarg arg ...\fR?
|
18 |
|
|
.BE
|
19 |
|
|
|
20 |
|
|
.SH DESCRIPTION
|
21 |
|
|
.PP
|
22 |
|
|
Concatenates \fIarg\fR's (adding separator spaces between them),
|
23 |
|
|
evaluates the result as a Tcl expression, and returns the value.
|
24 |
|
|
The operators permitted in Tcl expressions are a subset of
|
25 |
|
|
the operators permitted in C expressions, and they have the
|
26 |
|
|
same meaning and precedence as the corresponding C operators.
|
27 |
|
|
Expressions almost always yield numeric results
|
28 |
|
|
(integer or floating-point values).
|
29 |
|
|
For example, the expression
|
30 |
|
|
.CS
|
31 |
|
|
\fBexpr 8.2 + 6\fR
|
32 |
|
|
.CE
|
33 |
|
|
evaluates to 14.2.
|
34 |
|
|
Tcl expressions differ from C expressions in the way that
|
35 |
|
|
operands are specified. Also, Tcl expressions support
|
36 |
|
|
non-numeric operands and string comparisons.
|
37 |
|
|
.SH OPERANDS
|
38 |
|
|
.PP
|
39 |
|
|
A Tcl expression consists of a combination of operands, operators,
|
40 |
|
|
and parentheses.
|
41 |
|
|
White space may be used between the operands and operators and
|
42 |
|
|
parentheses; it is ignored by the expression's instructions.
|
43 |
|
|
Where possible, operands are interpreted as integer values.
|
44 |
|
|
Integer values may be specified in decimal (the normal case), in octal (if the
|
45 |
|
|
first character of the operand is \fB0\fR), or in hexadecimal (if the first
|
46 |
|
|
two characters of the operand are \fB0x\fR).
|
47 |
|
|
If an operand does not have one of the integer formats given
|
48 |
|
|
above, then it is treated as a floating-point number if that is
|
49 |
|
|
possible. Floating-point numbers may be specified in any of the
|
50 |
|
|
ways accepted by an ANSI-compliant C compiler (except that the
|
51 |
|
|
\fBf\fR, \fBF\fR, \fBl\fR, and \fBL\fR suffixes will not be permitted in
|
52 |
|
|
most installations). For example, all of the
|
53 |
|
|
following are valid floating-point numbers: 2.1, 3., 6e4, 7.91e+16.
|
54 |
|
|
If no numeric interpretation is possible, then an operand is left
|
55 |
|
|
as a string (and only a limited set of operators may be applied to
|
56 |
|
|
it).
|
57 |
|
|
.PP
|
58 |
|
|
Operands may be specified in any of the following ways:
|
59 |
|
|
.IP [1]
|
60 |
|
|
As an numeric value, either integer or floating-point.
|
61 |
|
|
.IP [2]
|
62 |
|
|
As a Tcl variable, using standard \fB$\fR notation.
|
63 |
|
|
The variable's value will be used as the operand.
|
64 |
|
|
.IP [3]
|
65 |
|
|
As a string enclosed in double-quotes.
|
66 |
|
|
The expression parser will perform backslash, variable, and
|
67 |
|
|
command substitutions on the information between the quotes,
|
68 |
|
|
and use the resulting value as the operand
|
69 |
|
|
.IP [4]
|
70 |
|
|
As a string enclosed in braces.
|
71 |
|
|
The characters between the open brace and matching close brace
|
72 |
|
|
will be used as the operand without any substitutions.
|
73 |
|
|
.IP [5]
|
74 |
|
|
As a Tcl command enclosed in brackets.
|
75 |
|
|
The command will be executed and its result will be used as
|
76 |
|
|
the operand.
|
77 |
|
|
.IP [6]
|
78 |
|
|
As a mathematical function whose arguments have any of the above
|
79 |
|
|
forms for operands, such as \fBsin($x)\fR. See below for a list of defined
|
80 |
|
|
functions.
|
81 |
|
|
.LP
|
82 |
|
|
Where substitutions occur above (e.g. inside quoted strings), they
|
83 |
|
|
are performed by the expression's instructions.
|
84 |
|
|
However, an additional layer of substitution may already have
|
85 |
|
|
been performed by the command parser before the expression
|
86 |
|
|
processor was called.
|
87 |
|
|
As discussed below, it is usually best to enclose expressions
|
88 |
|
|
in braces to prevent the command parser from performing substitutions
|
89 |
|
|
on the contents.
|
90 |
|
|
.PP
|
91 |
|
|
For some examples of simple expressions, suppose the variable
|
92 |
|
|
\fBa\fR has the value 3 and
|
93 |
|
|
the variable \fBb\fR has the value 6.
|
94 |
|
|
Then the command on the left side of each of the lines below
|
95 |
|
|
will produce the value on the right side of the line:
|
96 |
|
|
.CS
|
97 |
|
|
.ta 6c
|
98 |
|
|
\fBexpr 3.1 + $a 6.1
|
99 |
|
|
expr 2 + "$a.$b" 5.6
|
100 |
|
|
expr 4*[llength "6 2"] 8
|
101 |
|
|
expr {{word one} < "word $a"} 0\fR
|
102 |
|
|
.CE
|
103 |
|
|
.SH OPERATORS
|
104 |
|
|
.PP
|
105 |
|
|
The valid operators are listed below, grouped in decreasing order
|
106 |
|
|
of precedence:
|
107 |
|
|
.TP 20
|
108 |
|
|
\fB\-\0\0+\0\0~\0\0!\fR
|
109 |
|
|
Unary minus, unary plus, bit-wise NOT, logical NOT. None of these operands
|
110 |
|
|
may be applied to string operands, and bit-wise NOT may be
|
111 |
|
|
applied only to integers.
|
112 |
|
|
.TP 20
|
113 |
|
|
\fB*\0\0/\0\0%\fR
|
114 |
|
|
Multiply, divide, remainder. None of these operands may be
|
115 |
|
|
applied to string operands, and remainder may be applied only
|
116 |
|
|
to integers.
|
117 |
|
|
The remainder will always have the same sign as the divisor and
|
118 |
|
|
an absolute value smaller than the divisor.
|
119 |
|
|
.TP 20
|
120 |
|
|
\fB+\0\0\-\fR
|
121 |
|
|
Add and subtract. Valid for any numeric operands.
|
122 |
|
|
.TP 20
|
123 |
|
|
\fB<<\0\0>>\fR
|
124 |
|
|
Left and right shift. Valid for integer operands only.
|
125 |
|
|
A right shift always propagates the sign bit.
|
126 |
|
|
.TP 20
|
127 |
|
|
\fB<\0\0>\0\0<=\0\0>=\fR
|
128 |
|
|
Boolean less, greater, less than or equal, and greater than or equal.
|
129 |
|
|
Each operator produces 1 if the condition is true, 0 otherwise.
|
130 |
|
|
These operators may be applied to strings as well as numeric operands,
|
131 |
|
|
in which case string comparison is used.
|
132 |
|
|
.TP 20
|
133 |
|
|
\fB==\0\0!=\fR
|
134 |
|
|
Boolean equal and not equal. Each operator produces a zero/one result.
|
135 |
|
|
Valid for all operand types.
|
136 |
|
|
.TP 20
|
137 |
|
|
\fB&\fR
|
138 |
|
|
Bit-wise AND. Valid for integer operands only.
|
139 |
|
|
.TP 20
|
140 |
|
|
\fB^\fR
|
141 |
|
|
Bit-wise exclusive OR. Valid for integer operands only.
|
142 |
|
|
.TP 20
|
143 |
|
|
\fB|\fR
|
144 |
|
|
Bit-wise OR. Valid for integer operands only.
|
145 |
|
|
.TP 20
|
146 |
|
|
\fB&&\fR
|
147 |
|
|
Logical AND. Produces a 1 result if both operands are non-zero,
|
148 |
|
|
|
149 |
|
|
Valid for boolean and numeric (integers or floating-point) operands only.
|
150 |
|
|
.TP 20
|
151 |
|
|
\fB||\fR
|
152 |
|
|
Logical OR. Produces a 0 result if both operands are zero, 1 otherwise.
|
153 |
|
|
Valid for boolean and numeric (integers or floating-point) operands only.
|
154 |
|
|
.TP 20
|
155 |
|
|
\fIx\fB?\fIy\fB:\fIz\fR
|
156 |
|
|
If-then-else, as in C. If \fIx\fR
|
157 |
|
|
evaluates to non-zero, then the result is the value of \fIy\fR.
|
158 |
|
|
Otherwise the result is the value of \fIz\fR.
|
159 |
|
|
The \fIx\fR operand must have a numeric value.
|
160 |
|
|
.LP
|
161 |
|
|
See the C manual for more details on the results
|
162 |
|
|
produced by each operator.
|
163 |
|
|
All of the binary operators group left-to-right within the same
|
164 |
|
|
precedence level. For example, the command
|
165 |
|
|
.CS
|
166 |
|
|
\fBexpr 4*2 < 7\fR
|
167 |
|
|
.CE
|
168 |
|
|
returns 0.
|
169 |
|
|
.PP
|
170 |
|
|
The \fB&&\fR, \fB||\fR, and \fB?:\fR operators have ``lazy
|
171 |
|
|
evaluation'', just as in C,
|
172 |
|
|
which means that operands are not evaluated if they are
|
173 |
|
|
not needed to determine the outcome. For example, in the command
|
174 |
|
|
.CS
|
175 |
|
|
\fBexpr {$v ? [a] : [b]}\fR
|
176 |
|
|
.CE
|
177 |
|
|
only one of \fB[a]\fR or \fB[b]\fR will actually be evaluated,
|
178 |
|
|
depending on the value of \fB$v\fR. Note, however, that this is
|
179 |
|
|
only true if the entire expression is enclosed in braces; otherwise
|
180 |
|
|
the Tcl parser will evaluate both \fB[a]\fR and \fB[b]\fR before
|
181 |
|
|
invoking the \fBexpr\fR command.
|
182 |
|
|
.SH "MATH FUNCTIONS"
|
183 |
|
|
.PP
|
184 |
|
|
Tcl supports the following mathematical functions in expressions:
|
185 |
|
|
.DS
|
186 |
|
|
.ta 3c 6c 9c
|
187 |
|
|
\fBacos\fR \fBcos\fR \fBhypot\fR \fBsinh\fR
|
188 |
|
|
\fBasin\fR \fBcosh\fR \fBlog\fR \fBsqrt\fR
|
189 |
|
|
\fBatan\fR \fBexp\fR \fBlog10\fR \fBtan\fR
|
190 |
|
|
\fBatan2\fR \fBfloor\fR \fBpow\fR \fBtanh\fR
|
191 |
|
|
\fBceil\fR \fBfmod\fR \fBsin\fR
|
192 |
|
|
.DE
|
193 |
|
|
Each of these functions invokes the math library function of the same
|
194 |
|
|
name; see the manual entries for the library functions for details
|
195 |
|
|
on what they do. Tcl also implements the following functions for
|
196 |
|
|
conversion between integers and floating-point numbers and the
|
197 |
|
|
generation of random numbers:
|
198 |
|
|
.TP
|
199 |
|
|
\fBabs(\fIarg\fB)\fR
|
200 |
|
|
Returns the absolute value of \fIarg\fR. \fIArg\fR may be either
|
201 |
|
|
integer or floating-point, and the result is returned in the same form.
|
202 |
|
|
.TP
|
203 |
|
|
\fBdouble(\fIarg\fB)\fR
|
204 |
|
|
If \fIarg\fR is a floating value, returns \fIarg\fR, otherwise converts
|
205 |
|
|
\fIarg\fR to floating and returns the converted value.
|
206 |
|
|
.TP
|
207 |
|
|
\fBint(\fIarg\fB)\fR
|
208 |
|
|
If \fIarg\fR is an integer value, returns \fIarg\fR, otherwise converts
|
209 |
|
|
\fIarg\fR to integer by truncation and returns the converted value.
|
210 |
|
|
.TP
|
211 |
|
|
\fBrand()\fR
|
212 |
|
|
Returns a floating point number from zero to just less than one or,
|
213 |
|
|
in mathematical terms, the range [0,1). The seed comes from the
|
214 |
|
|
internal clock of the machine or may be set manual with the srand
|
215 |
|
|
function.
|
216 |
|
|
.TP
|
217 |
|
|
\fBround(\fIarg\fB)\fR
|
218 |
|
|
If \fIarg\fR is an integer value, returns \fIarg\fR, otherwise converts
|
219 |
|
|
\fIarg\fR to integer by rounding and returns the converted value.
|
220 |
|
|
.TP
|
221 |
|
|
\fBsrand(\fIarg\fB)\fR
|
222 |
|
|
The \fIarg\fR, which must be an integer, is used to reset the seed for
|
223 |
|
|
the random number generator. Returns the first random number from
|
224 |
|
|
that seed. Each interpreter has it's own seed.
|
225 |
|
|
.PP
|
226 |
|
|
In addition to these predefined functions, applications may
|
227 |
|
|
define additional functions using \fBTcl_CreateMathFunc\fR().
|
228 |
|
|
.SH "TYPES, OVERFLOW, AND PRECISION"
|
229 |
|
|
.PP
|
230 |
|
|
All internal computations involving integers are done with the C type
|
231 |
|
|
\fIlong\fR, and all internal computations involving floating-point are
|
232 |
|
|
done with the C type \fIdouble\fR.
|
233 |
|
|
When converting a string to floating-point, exponent overflow is
|
234 |
|
|
detected and results in a Tcl error.
|
235 |
|
|
For conversion to integer from string, detection of overflow depends
|
236 |
|
|
on the behavior of some routines in the local C library, so it should
|
237 |
|
|
be regarded as unreliable.
|
238 |
|
|
In any case, integer overflow and underflow are generally not detected
|
239 |
|
|
reliably for intermediate results. Floating-point overflow and underflow
|
240 |
|
|
are detected to the degree supported by the hardware, which is generally
|
241 |
|
|
pretty reliable.
|
242 |
|
|
.PP
|
243 |
|
|
Conversion among internal representations for integer, floating-point,
|
244 |
|
|
and string operands is done automatically as needed.
|
245 |
|
|
For arithmetic computations, integers are used until some
|
246 |
|
|
floating-point number is introduced, after which floating-point is used.
|
247 |
|
|
For example,
|
248 |
|
|
.CS
|
249 |
|
|
\fBexpr 5 / 4\fR
|
250 |
|
|
.CE
|
251 |
|
|
returns 1, while
|
252 |
|
|
.CS
|
253 |
|
|
\fBexpr 5 / 4.0\fR
|
254 |
|
|
\fBexpr 5 / ( [string length "abcd"] + 0.0 )\fR
|
255 |
|
|
.CE
|
256 |
|
|
both return 1.25.
|
257 |
|
|
Floating-point values are always returned with a ``\fB.\fR''
|
258 |
|
|
or an \fBe\fR so that they will not look like integer values. For
|
259 |
|
|
example,
|
260 |
|
|
.CS
|
261 |
|
|
\fBexpr 20.0/5.0\fR
|
262 |
|
|
.CE
|
263 |
|
|
returns \fB4.0\fR, not \fB4\fR.
|
264 |
|
|
|
265 |
|
|
.SH "STRING OPERATIONS"
|
266 |
|
|
.PP
|
267 |
|
|
String values may be used as operands of the comparison operators,
|
268 |
|
|
although the expression evaluator tries to do comparisons as integer
|
269 |
|
|
or floating-point when it can.
|
270 |
|
|
If one of the operands of a comparison is a string and the other
|
271 |
|
|
has a numeric value, the numeric operand is converted back to
|
272 |
|
|
a string using the C \fIsprintf\fR format specifier
|
273 |
|
|
\fB%d\fR for integers and \fB%g\fR for floating-point values.
|
274 |
|
|
For example, the commands
|
275 |
|
|
.CS
|
276 |
|
|
\fBexpr {"0x03" > "2"}\fR
|
277 |
|
|
\fBexpr {"0y" < "0x12"}\fR
|
278 |
|
|
.CE
|
279 |
|
|
both return 1. The first comparison is done using integer
|
280 |
|
|
comparison, and the second is done using string comparison after
|
281 |
|
|
the second operand is converted to the string \fB18\fR.
|
282 |
|
|
Because of Tcl's tendency to treat values as numbers whenever
|
283 |
|
|
possible, it isn't generally a good idea to use operators like \fB==\fR
|
284 |
|
|
when you really want string comparison and the values of the
|
285 |
|
|
operands could be arbitrary; it's better in these cases to use the
|
286 |
|
|
\fBstring compare\fR command instead.
|
287 |
|
|
|
288 |
|
|
.SH "PERFORMANCE CONSIDERATIONS"
|
289 |
|
|
.VS
|
290 |
|
|
.PP
|
291 |
|
|
Enclose expressions in braces for the best speed and the smallest
|
292 |
|
|
storage requirements.
|
293 |
|
|
This allows the Tcl bytecode compiler to generate the best code.
|
294 |
|
|
.PP
|
295 |
|
|
As mentioned above, expressions are substituted twice:
|
296 |
|
|
once by the Tcl parser and once by the \fBexpr\fR command.
|
297 |
|
|
For example, the commands
|
298 |
|
|
.CS
|
299 |
|
|
\fBset a 3\fR
|
300 |
|
|
\fBset b {$a + 2}\fR
|
301 |
|
|
\fBexpr $b*4\fR
|
302 |
|
|
.CE
|
303 |
|
|
return 11, not a multiple of 4.
|
304 |
|
|
This is because the Tcl parser will first substitute \fB$a + 2\fR for
|
305 |
|
|
the variable \fBb\fR,
|
306 |
|
|
then the \fBexpr\fR command will evaluate the expression \fB$a + 2*4\fR.
|
307 |
|
|
.PP
|
308 |
|
|
Most expressions do not require a second round of substitutions.
|
309 |
|
|
Either they are enclosed in braces or, if not,
|
310 |
|
|
their variable and command substitutions yield numbers or strings
|
311 |
|
|
that don't themselves require substitutions.
|
312 |
|
|
However, because a few unbraced expressions
|
313 |
|
|
need two rounds of substitutions,
|
314 |
|
|
the bytecode compiler must emit
|
315 |
|
|
additional instructions to handle this situation.
|
316 |
|
|
The most expensive code is required for
|
317 |
|
|
unbraced expressions that contain command substitutions.
|
318 |
|
|
These expressions must be implemented by generating new code
|
319 |
|
|
each time the expression is executed.
|
320 |
|
|
.VE
|
321 |
|
|
|
322 |
|
|
.SH KEYWORDS
|
323 |
|
|
arithmetic, boolean, compare, expression, fuzzy comparison
|