This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Accidental interpolation of $@ in Pod::Html [PATCH]
[perl5.git] / pod / perlop.pod
CommitLineData
a0d0e21e
LW
1=head1 NAME
2
3perlop - Perl operators and precedence
4
d042e63d
MS
5=head1 DESCRIPTION
6
7=head2 Operator Precedence and Associativity
8
9Operator precedence and associativity work in Perl more or less like
10they do in mathematics.
11
12I<Operator precedence> means some operators are evaluated before
13others. For example, in C<2 + 4 * 5>, the multiplication has higher
14precedence so C<4 * 5> is evaluated first yielding C<2 + 20 ==
1522> and not C<6 * 5 == 30>.
16
17I<Operator associativity> defines what happens if a sequence of the
18same operators is used one after another: whether the evaluator will
19evaluate the left operations first or the right. For example, in C<8
20- 4 - 2>, subtraction is left associative so Perl evaluates the
21expression left to right. C<8 - 4> is evaluated first making the
22expression C<4 - 2 == 2> and not C<8 - 2 == 6>.
a0d0e21e
LW
23
24Perl operators have the following associativity and precedence,
19799a22
GS
25listed from highest precedence to lowest. Operators borrowed from
26C keep the same precedence relationship with each other, even where
27C's precedence is slightly screwy. (This makes learning Perl easier
28for C folks.) With very few exceptions, these all operate on scalar
29values only, not array values.
a0d0e21e
LW
30
31 left terms and list operators (leftward)
32 left ->
33 nonassoc ++ --
34 right **
35 right ! ~ \ and unary + and -
54310121 36 left =~ !~
a0d0e21e
LW
37 left * / % x
38 left + - .
39 left << >>
40 nonassoc named unary operators
41 nonassoc < > <= >= lt gt le ge
42 nonassoc == != <=> eq ne cmp
43 left &
44 left | ^
45 left &&
c963b151 46 left || //
137443ea 47 nonassoc .. ...
a0d0e21e
LW
48 right ?:
49 right = += -= *= etc.
50 left , =>
51 nonassoc list operators (rightward)
a5f75d66 52 right not
a0d0e21e 53 left and
c963b151 54 left or xor err
a0d0e21e
LW
55
56In the following sections, these operators are covered in precedence order.
57
5a964f20
TC
58Many operators can be overloaded for objects. See L<overload>.
59
a0d0e21e
LW
60=head2 Terms and List Operators (Leftward)
61
62c18ce2 62A TERM has the highest precedence in Perl. They include variables,
5f05dabc 63quote and quote-like operators, any expression in parentheses,
a0d0e21e
LW
64and any function whose arguments are parenthesized. Actually, there
65aren't really functions in this sense, just list operators and unary
66operators behaving as functions because you put parentheses around
67the arguments. These are all documented in L<perlfunc>.
68
69If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
70is followed by a left parenthesis as the next token, the operator and
71arguments within parentheses are taken to be of highest precedence,
72just like a normal function call.
73
74In the absence of parentheses, the precedence of list operators such as
75C<print>, C<sort>, or C<chmod> is either very high or very low depending on
54310121 76whether you are looking at the left side or the right side of the operator.
a0d0e21e
LW
77For example, in
78
79 @ary = (1, 3, sort 4, 2);
80 print @ary; # prints 1324
81
19799a22
GS
82the commas on the right of the sort are evaluated before the sort,
83but the commas on the left are evaluated after. In other words,
84list operators tend to gobble up all arguments that follow, and
a0d0e21e 85then act like a simple TERM with regard to the preceding expression.
19799a22 86Be careful with parentheses:
a0d0e21e
LW
87
88 # These evaluate exit before doing the print:
89 print($foo, exit); # Obviously not what you want.
90 print $foo, exit; # Nor is this.
91
92 # These do the print before evaluating exit:
93 (print $foo), exit; # This is what you want.
94 print($foo), exit; # Or this.
95 print ($foo), exit; # Or even this.
96
97Also note that
98
99 print ($foo & 255) + 1, "\n";
100
d042e63d
MS
101probably doesn't do what you expect at first glance. The parentheses
102enclose the argument list for C<print> which is evaluated (printing
103the result of C<$foo & 255>). Then one is added to the return value
104of C<print> (usually 1). The result is something like this:
105
106 1 + 1, "\n"; # Obviously not what you meant.
107
108To do what you meant properly, you must write:
109
110 print(($foo & 255) + 1, "\n");
111
112See L<Named Unary Operators> for more discussion of this.
a0d0e21e
LW
113
114Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
54310121 115well as subroutine and method calls, and the anonymous
a0d0e21e
LW
116constructors C<[]> and C<{}>.
117
2ae324a7 118See also L<Quote and Quote-like Operators> toward the end of this section,
c07a80fd 119as well as L<"I/O Operators">.
a0d0e21e
LW
120
121=head2 The Arrow Operator
122
35f2feb0 123"C<< -> >>" is an infix dereference operator, just as it is in C
19799a22
GS
124and C++. If the right side is either a C<[...]>, C<{...}>, or a
125C<(...)> subscript, then the left side must be either a hard or
126symbolic reference to an array, a hash, or a subroutine respectively.
127(Or technically speaking, a location capable of holding a hard
128reference, if it's an array or hash reference being used for
129assignment.) See L<perlreftut> and L<perlref>.
a0d0e21e 130
19799a22
GS
131Otherwise, the right side is a method name or a simple scalar
132variable containing either the method name or a subroutine reference,
133and the left side must be either an object (a blessed reference)
134or a class name (that is, a package name). See L<perlobj>.
a0d0e21e 135
5f05dabc 136=head2 Auto-increment and Auto-decrement
a0d0e21e 137
d042e63d
MS
138"++" and "--" work as in C. That is, if placed before a variable,
139they increment or decrement the variable by one before returning the
140value, and if placed after, increment or decrement after returning the
141value.
142
143 $i = 0; $j = 0;
144 print $i++; # prints 0
145 print ++$j; # prints 1
a0d0e21e 146
b033823e
A
147Note that just as in C, Perl doesn't define B<when> the variable is
148incremented or decremented. You just know it will be done sometime
149before or after the value is returned. This also means that modifying
150a variable twice in the same statement will lead to undefined behaviour.
151Avoid statements like:
152
153 $i = $i ++;
154 print ++ $i + $i ++;
155
156Perl will not guarantee what the result of the above statements is.
157
54310121 158The auto-increment operator has a little extra builtin magic to it. If
a0d0e21e
LW
159you increment a variable that is numeric, or that has ever been used in
160a numeric context, you get a normal increment. If, however, the
5f05dabc 161variable has been used in only string contexts since it was set, and
5a964f20 162has a value that is not the empty string and matches the pattern
9c0670e1 163C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
a0d0e21e
LW
164character within its range, with carry:
165
166 print ++($foo = '99'); # prints '100'
167 print ++($foo = 'a0'); # prints 'a1'
168 print ++($foo = 'Az'); # prints 'Ba'
169 print ++($foo = 'zz'); # prints 'aaa'
170
6a61d433
HS
171C<undef> is always treated as numeric, and in particular is changed
172to C<0> before incrementing (so that a post-increment of an undef value
173will return C<0> rather than C<undef>).
174
5f05dabc 175The auto-decrement operator is not magical.
a0d0e21e
LW
176
177=head2 Exponentiation
178
19799a22 179Binary "**" is the exponentiation operator. It binds even more
cb1a09d0
AD
180tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
181implemented using C's pow(3) function, which actually works on doubles
182internally.)
a0d0e21e
LW
183
184=head2 Symbolic Unary Operators
185
5f05dabc 186Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
a0d0e21e
LW
187precedence version of this.
188
189Unary "-" performs arithmetic negation if the operand is numeric. If
190the operand is an identifier, a string consisting of a minus sign
191concatenated with the identifier is returned. Otherwise, if the string
192starts with a plus or minus, a string starting with the opposite sign
bff5667c 193is returned. One effect of these rules is that -bareword is equivalent
d2a1eec3 194to the string "-bareword".
a0d0e21e 195
972b05a9
JH
196Unary "~" performs bitwise negation, i.e., 1's complement. For
197example, C<0666 & ~027> is 0640. (See also L<Integer Arithmetic> and
198L<Bitwise String Operators>.) Note that the width of the result is
199platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
200bits wide on a 64-bit platform, so if you are expecting a certain bit
d042e63d 201width, remember to use the & operator to mask off the excess bits.
a0d0e21e
LW
202
203Unary "+" has no effect whatsoever, even on strings. It is useful
204syntactically for separating a function name from a parenthesized expression
205that would otherwise be interpreted as the complete list of function
5ba421f6 206arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
a0d0e21e 207
19799a22
GS
208Unary "\" creates a reference to whatever follows it. See L<perlreftut>
209and L<perlref>. Do not confuse this behavior with the behavior of
210backslash within a string, although both forms do convey the notion
211of protecting the next thing from interpolation.
a0d0e21e
LW
212
213=head2 Binding Operators
214
c07a80fd 215Binary "=~" binds a scalar expression to a pattern match. Certain operations
cb1a09d0
AD
216search or modify the string $_ by default. This operator makes that kind
217of operation work on some other string. The right argument is a search
2c268ad5
TP
218pattern, substitution, or transliteration. The left argument is what is
219supposed to be searched, substituted, or transliterated instead of the default
f8bab1e9
GS
220$_. When used in scalar context, the return value generally indicates the
221success of the operation. Behavior in list context depends on the particular
d7782e69
SP
222operator. See L</"Regexp Quote-Like Operators"> for details and
223L<perlretut> for examples using these operators.
f8bab1e9
GS
224
225If the right argument is an expression rather than a search pattern,
2c268ad5 226substitution, or transliteration, it is interpreted as a search pattern at run
573e01ca 227time.
a0d0e21e
LW
228
229Binary "!~" is just like "=~" except the return value is negated in
230the logical sense.
231
232=head2 Multiplicative Operators
233
234Binary "*" multiplies two numbers.
235
236Binary "/" divides two numbers.
237
54310121
PP
238Binary "%" computes the modulus of two numbers. Given integer
239operands C<$a> and C<$b>: If C<$b> is positive, then C<$a % $b> is
240C<$a> minus the largest multiple of C<$b> that is not greater than
241C<$a>. If C<$b> is negative, then C<$a % $b> is C<$a> minus the
242smallest multiple of C<$b> that is not less than C<$a> (i.e. the
6bb4e6d4 243result will be less than or equal to zero).
0412d526 244Note that when C<use integer> is in scope, "%" gives you direct access
55d729e4
GS
245to the modulus operator as implemented by your C compiler. This
246operator is not as well defined for negative operands, but it will
247execute faster.
248
62d10b70
GS
249Binary "x" is the repetition operator. In scalar context or if the left
250operand is not enclosed in parentheses, it returns a string consisting
251of the left operand repeated the number of times specified by the right
252operand. In list context, if the left operand is enclosed in
3585017f
YST
253parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
254If the right operand is zero or negative, it returns an empty string
255or an empty list, depending on the context.
a0d0e21e
LW
256
257 print '-' x 80; # print row of dashes
258
259 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
260
261 @ones = (1) x 80; # a list of 80 1's
262 @ones = (5) x @ones; # set all elements to 5
263
264
265=head2 Additive Operators
266
267Binary "+" returns the sum of two numbers.
268
269Binary "-" returns the difference of two numbers.
270
271Binary "." concatenates two strings.
272
273=head2 Shift Operators
274
55497cff
PP
275Binary "<<" returns the value of its left argument shifted left by the
276number of bits specified by the right argument. Arguments should be
982ce180 277integers. (See also L<Integer Arithmetic>.)
a0d0e21e 278
55497cff
PP
279Binary ">>" returns the value of its left argument shifted right by
280the number of bits specified by the right argument. Arguments should
982ce180 281be integers. (See also L<Integer Arithmetic>.)
a0d0e21e 282
b16cf6df
JH
283Note that both "<<" and ">>" in Perl are implemented directly using
284"<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
285in force then signed C integers are used, else unsigned C integers are
286used. Either way, the implementation isn't going to generate results
287larger than the size of the integer type Perl was built with (32 bits
288or 64 bits).
289
290The result of overflowing the range of the integers is undefined
291because it is undefined also in C. In other words, using 32-bit
292integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
293of bits is also undefined.
294
a0d0e21e
LW
295=head2 Named Unary Operators
296
297The various named unary operators are treated as functions with one
568e6d8b 298argument, with optional parentheses.
a0d0e21e
LW
299
300If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
301is followed by a left parenthesis as the next token, the operator and
302arguments within parentheses are taken to be of highest precedence,
3981b0eb 303just like a normal function call. For example,
304because named unary operators are higher precedence than ||:
a0d0e21e
LW
305
306 chdir $foo || die; # (chdir $foo) || die
307 chdir($foo) || die; # (chdir $foo) || die
308 chdir ($foo) || die; # (chdir $foo) || die
309 chdir +($foo) || die; # (chdir $foo) || die
310
3981b0eb 311but, because * is higher precedence than named operators:
a0d0e21e
LW
312
313 chdir $foo * 20; # chdir ($foo * 20)
314 chdir($foo) * 20; # (chdir $foo) * 20
315 chdir ($foo) * 20; # (chdir $foo) * 20
316 chdir +($foo) * 20; # chdir ($foo * 20)
317
318 rand 10 * 20; # rand (10 * 20)
319 rand(10) * 20; # (rand 10) * 20
320 rand (10) * 20; # (rand 10) * 20
321 rand +(10) * 20; # rand (10 * 20)
322
568e6d8b
RGS
323Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
324treated like named unary operators, but they don't follow this functional
325parenthesis rule. That means, for example, that C<-f($file).".bak"> is
326equivalent to C<-f "$file.bak">.
327
5ba421f6 328See also L<"Terms and List Operators (Leftward)">.
a0d0e21e
LW
329
330=head2 Relational Operators
331
35f2feb0 332Binary "<" returns true if the left argument is numerically less than
a0d0e21e
LW
333the right argument.
334
35f2feb0 335Binary ">" returns true if the left argument is numerically greater
a0d0e21e
LW
336than the right argument.
337
35f2feb0 338Binary "<=" returns true if the left argument is numerically less than
a0d0e21e
LW
339or equal to the right argument.
340
35f2feb0 341Binary ">=" returns true if the left argument is numerically greater
a0d0e21e
LW
342than or equal to the right argument.
343
344Binary "lt" returns true if the left argument is stringwise less than
345the right argument.
346
347Binary "gt" returns true if the left argument is stringwise greater
348than the right argument.
349
350Binary "le" returns true if the left argument is stringwise less than
351or equal to the right argument.
352
353Binary "ge" returns true if the left argument is stringwise greater
354than or equal to the right argument.
355
356=head2 Equality Operators
357
358Binary "==" returns true if the left argument is numerically equal to
359the right argument.
360
361Binary "!=" returns true if the left argument is numerically not equal
362to the right argument.
363
35f2feb0 364Binary "<=>" returns -1, 0, or 1 depending on whether the left
6ee5d4e7 365argument is numerically less than, equal to, or greater than the right
d4ad863d 366argument. If your platform supports NaNs (not-a-numbers) as numeric
7d3a9d88
NC
367values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
368"<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
369returns true, as does NaN != anything else. If your platform doesn't
370support NaNs then NaN is just a string with numeric value 0.
371
2b54f59f
YST
372 perl -le '$a = "NaN"; print "No NaN support here" if $a == $a'
373 perl -le '$a = "NaN"; print "NaN support here" if $a != $a'
a0d0e21e
LW
374
375Binary "eq" returns true if the left argument is stringwise equal to
376the right argument.
377
378Binary "ne" returns true if the left argument is stringwise not equal
379to the right argument.
380
d4ad863d
JH
381Binary "cmp" returns -1, 0, or 1 depending on whether the left
382argument is stringwise less than, equal to, or greater than the right
383argument.
a0d0e21e 384
a034a98d
DD
385"lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
386by the current locale if C<use locale> is in effect. See L<perllocale>.
387
a0d0e21e
LW
388=head2 Bitwise And
389
2cdc098b 390Binary "&" returns its operands ANDed together bit by bit.
2c268ad5 391(See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
a0d0e21e 392
2cdc098b
MG
393Note that "&" has lower priority than relational operators, so for example
394the brackets are essential in a test like
395
396 print "Even\n" if ($x & 1) == 0;
397
a0d0e21e
LW
398=head2 Bitwise Or and Exclusive Or
399
2cdc098b 400Binary "|" returns its operands ORed together bit by bit.
2c268ad5 401(See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
a0d0e21e 402
2cdc098b 403Binary "^" returns its operands XORed together bit by bit.
2c268ad5 404(See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
a0d0e21e 405
2cdc098b
MG
406Note that "|" and "^" have lower priority than relational operators, so
407for example the brackets are essential in a test like
408
409 print "false\n" if (8 | 2) != 10;
410
a0d0e21e
LW
411=head2 C-style Logical And
412
413Binary "&&" performs a short-circuit logical AND operation. That is,
414if the left operand is false, the right operand is not even evaluated.
415Scalar or list context propagates down to the right operand if it
416is evaluated.
417
418=head2 C-style Logical Or
419
420Binary "||" performs a short-circuit logical OR operation. That is,
421if the left operand is true, the right operand is not even evaluated.
422Scalar or list context propagates down to the right operand if it
423is evaluated.
424
c963b151
BD
425=head2 C-style Logical Defined-Or
426
427Although it has no direct equivalent in C, Perl's C<//> operator is related
428to its C-style or. In fact, it's exactly the same as C<||>, except that it
429tests the left hand side's definedness instead of its truth. Thus, C<$a // $b>
430is similar to C<defined($a) || $b> (except that it returns the value of C<$a>
431rather than the value of C<defined($a)>) and is exactly equivalent to
432C<defined($a) ? $a : $b>. This is very useful for providing default values
d042e63d
MS
433for variables. If you actually want to test if at least one of C<$a> and
434C<$b> is defined, use C<defined($a // $b)>.
c963b151 435
d042e63d
MS
436The C<||>, C<//> and C<&&> operators return the last value evaluated
437(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
438portable way to find out the home directory might be:
a0d0e21e 439
c963b151
BD
440 $home = $ENV{'HOME'} // $ENV{'LOGDIR'} //
441 (getpwuid($<))[7] // die "You're homeless!\n";
a0d0e21e 442
5a964f20
TC
443In particular, this means that you shouldn't use this
444for selecting between two aggregates for assignment:
445
446 @a = @b || @c; # this is wrong
447 @a = scalar(@b) || @c; # really meant this
448 @a = @b ? @b : @c; # this works fine, though
449
c963b151
BD
450As more readable alternatives to C<&&>, C<//> and C<||> when used for
451control flow, Perl provides C<and>, C<err> and C<or> operators (see below).
452The short-circuit behavior is identical. The precedence of "and", "err"
453and "or" is much lower, however, so that you can safely use them after a
5a964f20 454list operator without the need for parentheses:
a0d0e21e
LW
455
456 unlink "alpha", "beta", "gamma"
457 or gripe(), next LINE;
458
459With the C-style operators that would have been written like this:
460
461 unlink("alpha", "beta", "gamma")
462 || (gripe(), next LINE);
463
eeb6a2c9 464Using "or" for assignment is unlikely to do what you want; see below.
5a964f20
TC
465
466=head2 Range Operators
a0d0e21e
LW
467
468Binary ".." is the range operator, which is really two different
fb53bbb2 469operators depending on the context. In list context, it returns a
54ae734e 470list of values counting (up by ones) from the left value to the right
2cdbc966 471value. If the left value is greater than the right value then it
fb53bbb2 472returns the empty list. The range operator is useful for writing
54ae734e 473C<foreach (1..10)> loops and for doing slice operations on arrays. In
2cdbc966
JD
474the current implementation, no temporary array is created when the
475range operator is used as the expression in C<foreach> loops, but older
476versions of Perl might burn a lot of memory when you write something
477like this:
a0d0e21e
LW
478
479 for (1 .. 1_000_000) {
480 # code
54310121 481 }
a0d0e21e 482
54ae734e
MG
483The range operator also works on strings, using the magical auto-increment,
484see below.
485
5a964f20 486In scalar context, ".." returns a boolean value. The operator is
a0d0e21e
LW
487bistable, like a flip-flop, and emulates the line-range (comma) operator
488of B<sed>, B<awk>, and various editors. Each ".." operator maintains its
489own boolean state. It is false as long as its left operand is false.
490Once the left operand is true, the range operator stays true until the
491right operand is true, I<AFTER> which the range operator becomes false
19799a22 492again. It doesn't become false till the next time the range operator is
a0d0e21e
LW
493evaluated. It can test the right operand and become false on the same
494evaluation it became true (as in B<awk>), but it still returns true once.
19799a22
GS
495If you don't want it to test the right operand till the next
496evaluation, as in B<sed>, just use three dots ("...") instead of
497two. In all other regards, "..." behaves just like ".." does.
498
499The right operand is not evaluated while the operator is in the
500"false" state, and the left operand is not evaluated while the
501operator is in the "true" state. The precedence is a little lower
502than || and &&. The value returned is either the empty string for
503false, or a sequence number (beginning with 1) for true. The
504sequence number is reset for each range encountered. The final
505sequence number in a range has the string "E0" appended to it, which
506doesn't affect its numeric value, but gives you something to search
507for if you want to exclude the endpoint. You can exclude the
508beginning point by waiting for the sequence number to be greater
df5f8116
CW
509than 1.
510
511If either operand of scalar ".." is a constant expression,
512that operand is considered true if it is equal (C<==>) to the current
513input line number (the C<$.> variable).
514
515To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
516but that is only an issue if you use a floating point expression; when
517implicitly using C<$.> as described in the previous paragraph, the
518comparison is C<int(EXPR) == int($.)> which is only an issue when C<$.>
519is set to a floating point value and you are not reading from a file.
520Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
521you want in scalar context because each of the operands are evaluated
522using their integer representation.
523
524Examples:
a0d0e21e
LW
525
526As a scalar operator:
527
df5f8116
CW
528 if (101 .. 200) { print; } # print 2nd hundred lines, short for
529 # if ($. == 101 .. $. == 200) ...
530 next line if (1 .. /^$/); # skip header lines, short for
531 # ... if ($. == 1 .. /^$/);
a0d0e21e
LW
532 s/^/> / if (/^$/ .. eof()); # quote body
533
5a964f20
TC
534 # parse mail messages
535 while (<>) {
536 $in_header = 1 .. /^$/;
df5f8116
CW
537 $in_body = /^$/ .. eof;
538 if ($in_header) {
539 # ...
540 } else { # in body
541 # ...
542 }
5a964f20 543 } continue {
df5f8116 544 close ARGV if eof; # reset $. each file
5a964f20
TC
545 }
546
acf31ca5
SF
547Here's a simple example to illustrate the difference between
548the two range operators:
549
550 @lines = (" - Foo",
551 "01 - Bar",
552 "1 - Baz",
553 " - Quux");
554
555 foreach(@lines)
556 {
557 if (/0/ .. /1/)
558 {
559 print "$_\n";
560 }
561 }
562
563This program will print only the line containing "Bar". If
564the range operator is changed to C<...>, it will also print the
565"Baz" line.
566
567And now some examples as a list operator:
a0d0e21e
LW
568
569 for (101 .. 200) { print; } # print $_ 100 times
3e3baf6d 570 @foo = @foo[0 .. $#foo]; # an expensive no-op
a0d0e21e
LW
571 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
572
5a964f20 573The range operator (in list context) makes use of the magical
5f05dabc 574auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
575can say
576
577 @alphabet = ('A' .. 'Z');
578
54ae734e 579to get all normal letters of the English alphabet, or
a0d0e21e
LW
580
581 $hexdigit = (0 .. 9, 'a' .. 'f')[$num & 15];
582
583to get a hexadecimal digit, or
584
585 @z2 = ('01' .. '31'); print $z2[$mday];
586
587to get dates with leading zeros. If the final value specified is not
588in the sequence that the magical increment would produce, the sequence
589goes until the next value would be longer than the final value
590specified.
591
df5f8116
CW
592Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
593return two elements in list context.
594
595 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
596
a0d0e21e
LW
597=head2 Conditional Operator
598
599Ternary "?:" is the conditional operator, just as in C. It works much
600like an if-then-else. If the argument before the ? is true, the
601argument before the : is returned, otherwise the argument after the :
cb1a09d0
AD
602is returned. For example:
603
54310121 604 printf "I have %d dog%s.\n", $n,
cb1a09d0
AD
605 ($n == 1) ? '' : "s";
606
607Scalar or list context propagates downward into the 2nd
54310121 608or 3rd argument, whichever is selected.
cb1a09d0
AD
609
610 $a = $ok ? $b : $c; # get a scalar
611 @a = $ok ? @b : @c; # get an array
612 $a = $ok ? @b : @c; # oops, that's just a count!
613
614The operator may be assigned to if both the 2nd and 3rd arguments are
615legal lvalues (meaning that you can assign to them):
a0d0e21e
LW
616
617 ($a_or_b ? $a : $b) = $c;
618
5a964f20
TC
619Because this operator produces an assignable result, using assignments
620without parentheses will get you in trouble. For example, this:
621
622 $a % 2 ? $a += 10 : $a += 2
623
624Really means this:
625
626 (($a % 2) ? ($a += 10) : $a) += 2
627
628Rather than this:
629
630 ($a % 2) ? ($a += 10) : ($a += 2)
631
19799a22
GS
632That should probably be written more simply as:
633
634 $a += ($a % 2) ? 10 : 2;
635
4633a7c4 636=head2 Assignment Operators
a0d0e21e
LW
637
638"=" is the ordinary assignment operator.
639
640Assignment operators work as in C. That is,
641
642 $a += 2;
643
644is equivalent to
645
646 $a = $a + 2;
647
648although without duplicating any side effects that dereferencing the lvalue
54310121
PP
649might trigger, such as from tie(). Other assignment operators work similarly.
650The following are recognized:
a0d0e21e
LW
651
652 **= += *= &= <<= &&=
653 -= /= |= >>= ||=
654 .= %= ^=
655 x=
656
19799a22 657Although these are grouped by family, they all have the precedence
a0d0e21e
LW
658of assignment.
659
b350dd2f
GS
660Unlike in C, the scalar assignment operator produces a valid lvalue.
661Modifying an assignment is equivalent to doing the assignment and
662then modifying the variable that was assigned to. This is useful
663for modifying a copy of something, like this:
a0d0e21e
LW
664
665 ($tmp = $global) =~ tr [A-Z] [a-z];
666
667Likewise,
668
669 ($a += 2) *= 3;
670
671is equivalent to
672
673 $a += 2;
674 $a *= 3;
675
b350dd2f
GS
676Similarly, a list assignment in list context produces the list of
677lvalues assigned to, and a list assignment in scalar context returns
678the number of elements produced by the expression on the right hand
679side of the assignment.
680
748a9306 681=head2 Comma Operator
a0d0e21e 682
5a964f20 683Binary "," is the comma operator. In scalar context it evaluates
a0d0e21e
LW
684its left argument, throws that value away, then evaluates its right
685argument and returns that value. This is just like C's comma operator.
686
5a964f20 687In list context, it's just the list argument separator, and inserts
a0d0e21e
LW
688both its arguments into the list.
689
d042e63d 690The C<< => >> operator is a synonym for the comma, but forces any word
719b43e8 691(consisting entirely of word characters) to its left to be interpreted
a44e5664
MS
692as a string (as of 5.001). This includes words that might otherwise be
693considered a constant or function call.
694
695 use constant FOO => "something";
696
697 my %h = ( FOO => 23 );
698
699is equivalent to:
700
701 my %h = ("FOO", 23);
702
703It is I<NOT>:
704
705 my %h = ("something", 23);
706
707If the argument on the left is not a word, it is first interpreted as
708an expression, and then the string value of that is used.
719b43e8
RGS
709
710The C<< => >> operator is helpful in documenting the correspondence
711between keys and values in hashes, and other paired elements in lists.
748a9306 712
a44e5664
MS
713 %hash = ( $key => $value );
714 login( $username => $password );
715
a0d0e21e
LW
716=head2 List Operators (Rightward)
717
718On the right side of a list operator, it has very low precedence,
719such that it controls all comma-separated expressions found there.
720The only operators with lower precedence are the logical operators
721"and", "or", and "not", which may be used to evaluate calls to list
722operators without the need for extra parentheses:
723
724 open HANDLE, "filename"
725 or die "Can't open: $!\n";
726
5ba421f6 727See also discussion of list operators in L<Terms and List Operators (Leftward)>.
a0d0e21e
LW
728
729=head2 Logical Not
730
731Unary "not" returns the logical negation of the expression to its right.
732It's the equivalent of "!" except for the very low precedence.
733
734=head2 Logical And
735
736Binary "and" returns the logical conjunction of the two surrounding
737expressions. It's equivalent to && except for the very low
5f05dabc 738precedence. This means that it short-circuits: i.e., the right
a0d0e21e
LW
739expression is evaluated only if the left expression is true.
740
c963b151 741=head2 Logical or, Defined or, and Exclusive Or
a0d0e21e
LW
742
743Binary "or" returns the logical disjunction of the two surrounding
5a964f20
TC
744expressions. It's equivalent to || except for the very low precedence.
745This makes it useful for control flow
746
747 print FH $data or die "Can't write to FH: $!";
748
749This means that it short-circuits: i.e., the right expression is evaluated
750only if the left expression is false. Due to its precedence, you should
751probably avoid using this for assignment, only for control flow.
752
753 $a = $b or $c; # bug: this is wrong
754 ($a = $b) or $c; # really means this
755 $a = $b || $c; # better written this way
756
19799a22 757However, when it's a list-context assignment and you're trying to use
5a964f20
TC
758"||" for control flow, you probably need "or" so that the assignment
759takes higher precedence.
760
761 @info = stat($file) || die; # oops, scalar sense of stat!
762 @info = stat($file) or die; # better, now @info gets its due
763
c963b151
BD
764Then again, you could always use parentheses.
765
766Binary "err" is equivalent to C<//>--it's just like binary "or", except it tests
767its left argument's definedness instead of its truth. There are two ways to
768remember "err": either because many functions return C<undef> on an B<err>or,
769or as a sort of correction: C<$a=($b err 'default')>
a0d0e21e
LW
770
771Binary "xor" returns the exclusive-OR of the two surrounding expressions.
772It cannot short circuit, of course.
773
774=head2 C Operators Missing From Perl
775
776Here is what C has that Perl doesn't:
777
778=over 8
779
780=item unary &
781
782Address-of operator. (But see the "\" operator for taking a reference.)
783
784=item unary *
785
54310121 786Dereference-address operator. (Perl's prefix dereferencing
a0d0e21e
LW
787operators are typed: $, @, %, and &.)
788
789=item (TYPE)
790
19799a22 791Type-casting operator.
a0d0e21e
LW
792
793=back
794
5f05dabc 795=head2 Quote and Quote-like Operators
a0d0e21e
LW
796
797While we usually think of quotes as literal values, in Perl they
798function as operators, providing various kinds of interpolating and
799pattern matching capabilities. Perl provides customary quote characters
800for these behaviors, but also provides a way for you to choose your
801quote character for any of them. In the following table, a C<{}> represents
87275199 802any pair of delimiters you choose.
a0d0e21e 803
2c268ad5
TP
804 Customary Generic Meaning Interpolates
805 '' q{} Literal no
806 "" qq{} Literal yes
af9219ee 807 `` qx{} Command yes*
2c268ad5 808 qw{} Word list no
af9219ee
MG
809 // m{} Pattern match yes*
810 qr{} Pattern yes*
811 s{}{} Substitution yes*
2c268ad5 812 tr{}{} Transliteration no (but see below)
7e3b091d 813 <<EOF here-doc yes*
a0d0e21e 814
af9219ee
MG
815 * unless the delimiter is ''.
816
87275199
GS
817Non-bracketing delimiters use the same character fore and aft, but the four
818sorts of brackets (round, angle, square, curly) will all nest, which means
819that
820
821 q{foo{bar}baz}
35f2feb0 822
87275199
GS
823is the same as
824
825 'foo{bar}baz'
826
827Note, however, that this does not always work for quoting Perl code:
828
829 $s = q{ if($a eq "}") ... }; # WRONG
830
83df6a1d
JH
831is a syntax error. The C<Text::Balanced> module (from CPAN, and
832starting from Perl 5.8 part of the standard distribution) is able
833to do this properly.
87275199 834
19799a22 835There can be whitespace between the operator and the quoting
fb73857a 836characters, except when C<#> is being used as the quoting character.
19799a22
GS
837C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
838operator C<q> followed by a comment. Its argument will be taken
839from the next line. This allows you to write:
fb73857a
PP
840
841 s {foo} # Replace foo
842 {bar} # with bar.
843
904501ec
MG
844The following escape sequences are available in constructs that interpolate
845and in transliterations.
a0d0e21e 846
6ee5d4e7 847 \t tab (HT, TAB)
5a964f20 848 \n newline (NL)
6ee5d4e7
PP
849 \r return (CR)
850 \f form feed (FF)
851 \b backspace (BS)
852 \a alarm (bell) (BEL)
853 \e escape (ESC)
a0ed51b3
LW
854 \033 octal char (ESC)
855 \x1b hex char (ESC)
856 \x{263a} wide hex char (SMILEY)
19799a22 857 \c[ control char (ESC)
95cc3e0c 858 \N{name} named Unicode character
2c268ad5 859
4c77eaa2
AE
860B<NOTE>: Unlike C and other languages, Perl has no \v escape sequence for
861the vertical tab (VT - ASCII 11).
862
904501ec
MG
863The following escape sequences are available in constructs that interpolate
864but not in transliterations.
865
a0d0e21e
LW
866 \l lowercase next char
867 \u uppercase next char
868 \L lowercase till \E
869 \U uppercase till \E
870 \E end case modification
1d2dff63 871 \Q quote non-word characters till \E
a0d0e21e 872
95cc3e0c
JH
873If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
874C<\u> and C<\U> is taken from the current locale. See L<perllocale>.
875If Unicode (for example, C<\N{}> or wide hex characters of 0x100 or
876beyond) is being used, the case map used by C<\l>, C<\L>, C<\u> and
877C<\U> is as defined by Unicode. For documentation of C<\N{name}>,
878see L<charnames>.
a034a98d 879
5a964f20
TC
880All systems use the virtual C<"\n"> to represent a line terminator,
881called a "newline". There is no such thing as an unvarying, physical
19799a22 882newline character. It is only an illusion that the operating system,
5a964f20
TC
883device drivers, C libraries, and Perl all conspire to preserve. Not all
884systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
885on a Mac, these are reversed, and on systems without line terminator,
886printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
887you mean a "newline" for your system, but use the literal ASCII when you
888need an exact character. For example, most networking protocols expect
2a380090 889and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
890and although they often accept just C<"\012">, they seldom tolerate just
891C<"\015">. If you get in the habit of using C<"\n"> for networking,
892you may be burned some day.
893
904501ec
MG
894For constructs that do interpolate, variables beginning with "C<$>"
895or "C<@>" are interpolated. Subscripted variables such as C<$a[3]> or
ad0f383a
A
896C<< $href->{key}[0] >> are also interpolated, as are array and hash slices.
897But method calls such as C<< $obj->meth >> are not.
af9219ee
MG
898
899Interpolating an array or slice interpolates the elements in order,
900separated by the value of C<$">, so is equivalent to interpolating
904501ec
MG
901C<join $", @array>. "Punctuation" arrays such as C<@+> are only
902interpolated if the name is enclosed in braces C<@{+}>.
af9219ee 903
1d2dff63
GS
904You cannot include a literal C<$> or C<@> within a C<\Q> sequence.
905An unescaped C<$> or C<@> interpolates the corresponding variable,
906while escaping will cause the literal string C<\$> to be inserted.
907You'll need to write something like C<m/\Quser\E\@\Qhost/>.
908
a0d0e21e
LW
909Patterns are subject to an additional level of interpretation as a
910regular expression. This is done as a second pass, after variables are
911interpolated, so that regular expressions may be incorporated into the
912pattern from the variables. If this is not what you want, use C<\Q> to
913interpolate a variable literally.
914
19799a22
GS
915Apart from the behavior described above, Perl does not expand
916multiple levels of interpolation. In particular, contrary to the
917expectations of shell programmers, back-quotes do I<NOT> interpolate
918within double quotes, nor do single quotes impede evaluation of
919variables when used within double quotes.
a0d0e21e 920
5f05dabc 921=head2 Regexp Quote-Like Operators
cb1a09d0 922
5f05dabc 923Here are the quote-like operators that apply to pattern
cb1a09d0
AD
924matching and related activities.
925
a0d0e21e
LW
926=over 8
927
928=item ?PATTERN?
929
930This is just like the C</pattern/> search, except that it matches only
931once between calls to the reset() operator. This is a useful
5f05dabc 932optimization when you want to see only the first occurrence of
a0d0e21e
LW
933something in each file of a set of files, for instance. Only C<??>
934patterns local to the current package are reset.
935
5a964f20
TC
936 while (<>) {
937 if (?^$?) {
938 # blank line between header and body
939 }
940 } continue {
941 reset if eof; # clear ?? status for next file
942 }
943
483b4840 944This usage is vaguely deprecated, which means it just might possibly
19799a22
GS
945be removed in some distant future version of Perl, perhaps somewhere
946around the year 2168.
a0d0e21e 947
fb73857a 948=item m/PATTERN/cgimosx
a0d0e21e 949
fb73857a 950=item /PATTERN/cgimosx
a0d0e21e 951
5a964f20 952Searches a string for a pattern match, and in scalar context returns
19799a22
GS
953true if it succeeds, false if it fails. If no string is specified
954via the C<=~> or C<!~> operator, the $_ string is searched. (The
955string specified with C<=~> need not be an lvalue--it may be the
956result of an expression evaluation, but remember the C<=~> binds
957rather tightly.) See also L<perlre>. See L<perllocale> for
958discussion of additional considerations that apply when C<use locale>
959is in effect.
a0d0e21e
LW
960
961Options are:
962
fb73857a 963 c Do not reset search position on a failed match when /g is in effect.
5f05dabc 964 g Match globally, i.e., find all occurrences.
a0d0e21e
LW
965 i Do case-insensitive pattern matching.
966 m Treat string as multiple lines.
5f05dabc 967 o Compile pattern only once.
a0d0e21e
LW
968 s Treat string as single line.
969 x Use extended regular expressions.
970
971If "/" is the delimiter then the initial C<m> is optional. With the C<m>
01ae956f 972you can use any pair of non-alphanumeric, non-whitespace characters
19799a22
GS
973as delimiters. This is particularly useful for matching path names
974that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
7bac28a0 975the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
19799a22 976If "'" is the delimiter, no interpolation is performed on the PATTERN.
a0d0e21e
LW
977
978PATTERN may contain variables, which will be interpolated (and the
f70b4f9c 979pattern recompiled) every time the pattern search is evaluated, except
1f247705
GS
980for when the delimiter is a single quote. (Note that C<$(>, C<$)>, and
981C<$|> are not interpolated because they look like end-of-string tests.)
f70b4f9c
AB
982If you want such a pattern to be compiled only once, add a C</o> after
983the trailing delimiter. This avoids expensive run-time recompilations,
984and is useful when the value you are interpolating won't change over
985the life of the script. However, mentioning C</o> constitutes a promise
986that you won't change the variables in the pattern. If you change them,
13a2d996 987Perl won't even notice. See also L<"qr/STRING/imosx">.
a0d0e21e 988
5a964f20 989If the PATTERN evaluates to the empty string, the last
d65afb4b
HS
990I<successfully> matched regular expression is used instead. In this
991case, only the C<g> and C<c> flags on the empty pattern is honoured -
992the other flags are taken from the original pattern. If no match has
993previously succeeded, this will (silently) act instead as a genuine
994empty pattern (which will always match).
a0d0e21e 995
c963b151
BD
996Note that it's possible to confuse Perl into thinking C<//> (the empty
997regex) is really C<//> (the defined-or operator). Perl is usually pretty
998good about this, but some pathological cases might trigger this, such as
999C<$a///> (is that C<($a) / (//)> or C<$a // />?) and C<print $fh //>
1000(C<print $fh(//> or C<print($fh //>?). In all of these examples, Perl
1001will assume you meant defined-or. If you meant the empty regex, just
1002use parentheses or spaces to disambiguate, or even prefix the empty
1003regex with an C<m> (so C<//> becomes C<m//>).
1004
19799a22 1005If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1006list consisting of the subexpressions matched by the parentheses in the
f7e33566
GS
1007pattern, i.e., (C<$1>, C<$2>, C<$3>...). (Note that here C<$1> etc. are
1008also set, and that this differs from Perl 4's behavior.) When there are
1009no parentheses in the pattern, the return value is the list C<(1)> for
1010success. With or without parentheses, an empty list is returned upon
1011failure.
a0d0e21e
LW
1012
1013Examples:
1014
1015 open(TTY, '/dev/tty');
1016 <TTY> =~ /^y/i && foo(); # do foo if desired
1017
1018 if (/Version: *([0-9.]*)/) { $version = $1; }
1019
1020 next if m#^/usr/spool/uucp#;
1021
1022 # poor man's grep
1023 $arg = shift;
1024 while (<>) {
1025 print if /$arg/o; # compile only once
1026 }
1027
1028 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
1029
1030This last example splits $foo into the first two words and the
5f05dabc
PP
1031remainder of the line, and assigns those three fields to $F1, $F2, and
1032$Etc. The conditional is true if any variables were assigned, i.e., if
a0d0e21e
LW
1033the pattern matched.
1034
19799a22
GS
1035The C</g> modifier specifies global pattern matching--that is,
1036matching as many times as possible within the string. How it behaves
1037depends on the context. In list context, it returns a list of the
1038substrings matched by any capturing parentheses in the regular
1039expression. If there are no parentheses, it returns a list of all
1040the matched strings, as if there were parentheses around the whole
1041pattern.
a0d0e21e 1042
7e86de3e 1043In scalar context, each execution of C<m//g> finds the next match,
19799a22 1044returning true if it matches, and false if there is no further match.
7e86de3e
G
1045The position after the last match can be read or set using the pos()
1046function; see L<perlfunc/pos>. A failed match normally resets the
1047search position to the beginning of the string, but you can avoid that
1048by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
1049string also resets the search position.
c90c0ff4
PP
1050
1051You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
1052zero-width assertion that matches the exact position where the previous
5d43e42d
DC
1053C<m//g>, if any, left off. Without the C</g> modifier, the C<\G> assertion
1054still anchors at pos(), but the match is of course only attempted once.
1055Using C<\G> without C</g> on a target string that has not previously had a
1056C</g> match applied to it is the same as using the C<\A> assertion to match
fe4b3f22
RGS
1057the beginning of the string. Note also that, currently, C<\G> is only
1058properly supported when anchored at the very beginning of the pattern.
c90c0ff4
PP
1059
1060Examples:
a0d0e21e
LW
1061
1062 # list context
1063 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1064
1065 # scalar context
5d43e42d 1066 $/ = "";
19799a22
GS
1067 while (defined($paragraph = <>)) {
1068 while ($paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
1069 $sentences++;
a0d0e21e
LW
1070 }
1071 }
1072 print "$sentences\n";
1073
c90c0ff4 1074 # using m//gc with \G
137443ea 1075 $_ = "ppooqppqq";
44a8e56a
PP
1076 while ($i++ < 2) {
1077 print "1: '";
c90c0ff4 1078 print $1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1079 print "2: '";
c90c0ff4 1080 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1081 print "3: '";
c90c0ff4 1082 print $1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1083 }
5d43e42d 1084 print "Final: '$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a
PP
1085
1086The last example should print:
1087
1088 1: 'oo', pos=4
137443ea 1089 2: 'q', pos=5
44a8e56a
PP
1090 3: 'pp', pos=7
1091 1: '', pos=7
137443ea
PP
1092 2: 'q', pos=8
1093 3: '', pos=8
5d43e42d
DC
1094 Final: 'q', pos=8
1095
1096Notice that the final match matched C<q> instead of C<p>, which a match
1097without the C<\G> anchor would have done. Also note that the final match
1098did not update C<pos> -- C<pos> is only updated on a C</g> match. If the
1099final match did indeed match C<p>, it's a good bet that you're running an
1100older (pre-5.6.0) Perl.
44a8e56a 1101
c90c0ff4 1102A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1103combine several regexps like this to process a string part-by-part,
c90c0ff4
PP
1104doing different actions depending on which regexp matched. Each
1105regexp tries to match where the previous one leaves off.
e7ea3e70 1106
3fe9a6f1 1107 $_ = <<'EOL';
e7ea3e70 1108 $url = new URI::URL "http://www/"; die if $url eq "xXx";
3fe9a6f1
PP
1109 EOL
1110 LOOP:
e7ea3e70 1111 {
c90c0ff4
PP
1112 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
1113 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
1114 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
1115 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
1116 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
1117 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
1118 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
e7ea3e70
IZ
1119 print ". That's all!\n";
1120 }
1121
1122Here is the output (split into several lines):
1123
1124 line-noise lowercase line-noise lowercase UPPERCASE line-noise
1125 UPPERCASE line-noise lowercase line-noise lowercase line-noise
1126 lowercase lowercase line-noise lowercase lowercase line-noise
1127 MiXeD line-noise. That's all!
44a8e56a 1128
a0d0e21e
LW
1129=item q/STRING/
1130
1131=item C<'STRING'>
1132
19799a22 1133A single-quoted, literal string. A backslash represents a backslash
68dc0745
PP
1134unless followed by the delimiter or another backslash, in which case
1135the delimiter or backslash is interpolated.
a0d0e21e
LW
1136
1137 $foo = q!I said, "You said, 'She said it.'"!;
1138 $bar = q('This is it.');
68dc0745 1139 $baz = '\n'; # a two-character string
a0d0e21e
LW
1140
1141=item qq/STRING/
1142
1143=item "STRING"
1144
1145A double-quoted, interpolated string.
1146
1147 $_ .= qq
1148 (*** The previous line contains the naughty word "$1".\n)
19799a22 1149 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 1150 $baz = "\n"; # a one-character string
a0d0e21e 1151
eec2d3df
GS
1152=item qr/STRING/imosx
1153
322edccd 1154This operator quotes (and possibly compiles) its I<STRING> as a regular
19799a22
GS
1155expression. I<STRING> is interpolated the same way as I<PATTERN>
1156in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1157is done. Returns a Perl value which may be used instead of the
1158corresponding C</STRING/imosx> expression.
4b6a7270
IZ
1159
1160For example,
1161
1162 $rex = qr/my.STRING/is;
1163 s/$rex/foo/;
1164
1165is equivalent to
1166
1167 s/my.STRING/foo/is;
1168
1169The result may be used as a subpattern in a match:
eec2d3df
GS
1170
1171 $re = qr/$pattern/;
0a92e3a8
GS
1172 $string =~ /foo${re}bar/; # can be interpolated in other patterns
1173 $string =~ $re; # or used standalone
4b6a7270
IZ
1174 $string =~ /$re/; # or this way
1175
1176Since Perl may compile the pattern at the moment of execution of qr()
19799a22 1177operator, using qr() may have speed advantages in some situations,
4b6a7270
IZ
1178notably if the result of qr() is used standalone:
1179
1180 sub match {
1181 my $patterns = shift;
1182 my @compiled = map qr/$_/i, @$patterns;
1183 grep {
1184 my $success = 0;
a7665c5e 1185 foreach my $pat (@compiled) {
4b6a7270
IZ
1186 $success = 1, last if /$pat/;
1187 }
1188 $success;
1189 } @_;
1190 }
1191
19799a22
GS
1192Precompilation of the pattern into an internal representation at
1193the moment of qr() avoids a need to recompile the pattern every
1194time a match C</$pat/> is attempted. (Perl has many other internal
1195optimizations, but none would be triggered in the above example if
1196we did not use qr() operator.)
eec2d3df
GS
1197
1198Options are:
1199
1200 i Do case-insensitive pattern matching.
1201 m Treat string as multiple lines.
1202 o Compile pattern only once.
1203 s Treat string as single line.
1204 x Use extended regular expressions.
1205
0a92e3a8
GS
1206See L<perlre> for additional information on valid syntax for STRING, and
1207for a detailed look at the semantics of regular expressions.
1208
a0d0e21e
LW
1209=item qx/STRING/
1210
1211=item `STRING`
1212
43dd4d21
JH
1213A string which is (possibly) interpolated and then executed as a
1214system command with C</bin/sh> or its equivalent. Shell wildcards,
1215pipes, and redirections will be honored. The collected standard
1216output of the command is returned; standard error is unaffected. In
1217scalar context, it comes back as a single (potentially multi-line)
1218string, or undef if the command failed. In list context, returns a
1219list of lines (however you've defined lines with $/ or
1220$INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
5a964f20
TC
1221
1222Because backticks do not affect standard error, use shell file descriptor
1223syntax (assuming the shell supports this) if you care to address this.
1224To capture a command's STDERR and STDOUT together:
a0d0e21e 1225
5a964f20
TC
1226 $output = `cmd 2>&1`;
1227
1228To capture a command's STDOUT but discard its STDERR:
1229
1230 $output = `cmd 2>/dev/null`;
1231
1232To capture a command's STDERR but discard its STDOUT (ordering is
1233important here):
1234
1235 $output = `cmd 2>&1 1>/dev/null`;
1236
1237To exchange a command's STDOUT and STDERR in order to capture the STDERR
1238but leave its STDOUT to come out the old STDERR:
1239
1240 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1241
1242To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
1243to redirect them separately to files, and then read from those files
1244when the program is done:
5a964f20 1245
2359510d 1246 system("program args 1>program.stdout 2>program.stderr");
5a964f20
TC
1247
1248Using single-quote as a delimiter protects the command from Perl's
1249double-quote interpolation, passing it on to the shell instead:
1250
1251 $perl_info = qx(ps $$); # that's Perl's $$
1252 $shell_info = qx'ps $$'; # that's the new shell's $$
1253
19799a22 1254How that string gets evaluated is entirely subject to the command
5a964f20
TC
1255interpreter on your system. On most platforms, you will have to protect
1256shell metacharacters if you want them treated literally. This is in
1257practice difficult to do, as it's unclear how to escape which characters.
1258See L<perlsec> for a clean and safe example of a manual fork() and exec()
1259to emulate backticks safely.
a0d0e21e 1260
bb32b41a
GS
1261On some platforms (notably DOS-like ones), the shell may not be
1262capable of dealing with multiline commands, so putting newlines in
1263the string may not get you what you want. You may be able to evaluate
1264multiple commands in a single line by separating them with the command
1265separator character, if your shell supports that (e.g. C<;> on many Unix
1266shells; C<&> on the Windows NT C<cmd> shell).
1267
0f897271
GS
1268Beginning with v5.6.0, Perl will attempt to flush all files opened for
1269output before starting the child process, but this may not be supported
1270on some platforms (see L<perlport>). To be safe, you may need to set
1271C<$|> ($AUTOFLUSH in English) or call the C<autoflush()> method of
1272C<IO::Handle> on any open handles.
1273
bb32b41a
GS
1274Beware that some command shells may place restrictions on the length
1275of the command line. You must ensure your strings don't exceed this
1276limit after any necessary interpolations. See the platform-specific
1277release notes for more details about your particular environment.
1278
5a964f20
TC
1279Using this operator can lead to programs that are difficult to port,
1280because the shell commands called vary between systems, and may in
1281fact not be present at all. As one example, the C<type> command under
1282the POSIX shell is very different from the C<type> command under DOS.
1283That doesn't mean you should go out of your way to avoid backticks
1284when they're the right way to get something done. Perl was made to be
1285a glue language, and one of the things it glues together is commands.
1286Just understand what you're getting yourself into.
bb32b41a 1287
dc848c6f 1288See L<"I/O Operators"> for more discussion.
a0d0e21e 1289
945c54fd
JH
1290=item qw/STRING/
1291
1292Evaluates to a list of the words extracted out of STRING, using embedded
1293whitespace as the word delimiters. It can be understood as being roughly
1294equivalent to:
1295
1296 split(' ', q/STRING/);
1297
efb1e162
CW
1298the differences being that it generates a real list at compile time, and
1299in scalar context it returns the last element in the list. So
945c54fd
JH
1300this expression:
1301
1302 qw(foo bar baz)
1303
1304is semantically equivalent to the list:
1305
1306 'foo', 'bar', 'baz'
1307
1308Some frequently seen examples:
1309
1310 use POSIX qw( setlocale localeconv )
1311 @EXPORT = qw( foo bar baz );
1312
1313A common mistake is to try to separate the words with comma or to
1314put comments into a multi-line C<qw>-string. For this reason, the
1315C<use warnings> pragma and the B<-w> switch (that is, the C<$^W> variable)
1316produces warnings if the STRING contains the "," or the "#" character.
1317
a0d0e21e
LW
1318=item s/PATTERN/REPLACEMENT/egimosx
1319
1320Searches a string for a pattern, and if found, replaces that pattern
1321with the replacement text and returns the number of substitutions
e37d713d 1322made. Otherwise it returns false (specifically, the empty string).
a0d0e21e
LW
1323
1324If no string is specified via the C<=~> or C<!~> operator, the C<$_>
1325variable is searched and modified. (The string specified with C<=~> must
5a964f20 1326be scalar variable, an array element, a hash element, or an assignment
5f05dabc 1327to one of those, i.e., an lvalue.)
a0d0e21e 1328
19799a22 1329If the delimiter chosen is a single quote, no interpolation is
a0d0e21e
LW
1330done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1331PATTERN contains a $ that looks like a variable rather than an
1332end-of-string test, the variable will be interpolated into the pattern
5f05dabc 1333at run-time. If you want the pattern compiled only once the first time
a0d0e21e 1334the variable is interpolated, use the C</o> option. If the pattern
5a964f20 1335evaluates to the empty string, the last successfully executed regular
a0d0e21e 1336expression is used instead. See L<perlre> for further explanation on these.
5a964f20 1337See L<perllocale> for discussion of additional considerations that apply
a034a98d 1338when C<use locale> is in effect.
a0d0e21e
LW
1339
1340Options are:
1341
1342 e Evaluate the right side as an expression.
5f05dabc 1343 g Replace globally, i.e., all occurrences.
a0d0e21e
LW
1344 i Do case-insensitive pattern matching.
1345 m Treat string as multiple lines.
5f05dabc 1346 o Compile pattern only once.
a0d0e21e
LW
1347 s Treat string as single line.
1348 x Use extended regular expressions.
1349
1350Any non-alphanumeric, non-whitespace delimiter may replace the
1351slashes. If single quotes are used, no interpretation is done on the
e37d713d 1352replacement string (the C</e> modifier overrides this, however). Unlike
54310121 1353Perl 4, Perl 5 treats backticks as normal delimiters; the replacement
e37d713d 1354text is not evaluated as a command. If the
a0d0e21e 1355PATTERN is delimited by bracketing quotes, the REPLACEMENT has its own
5f05dabc 1356pair of quotes, which may or may not be bracketing quotes, e.g.,
35f2feb0 1357C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
cec88af6
GS
1358replacement portion to be treated as a full-fledged Perl expression
1359and evaluated right then and there. It is, however, syntax checked at
1360compile-time. A second C<e> modifier will cause the replacement portion
1361to be C<eval>ed before being run as a Perl expression.
a0d0e21e
LW
1362
1363Examples:
1364
1365 s/\bgreen\b/mauve/g; # don't change wintergreen
1366
1367 $path =~ s|/usr/bin|/usr/local/bin|;
1368
1369 s/Login: $foo/Login: $bar/; # run-time pattern
1370
5a964f20 1371 ($foo = $bar) =~ s/this/that/; # copy first, then change
a0d0e21e 1372
5a964f20 1373 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-count
a0d0e21e
LW
1374
1375 $_ = 'abc123xyz';
1376 s/\d+/$&*2/e; # yields 'abc246xyz'
1377 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
1378 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1379
1380 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
1381 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
1382 s/^=(\w+)/&pod($1)/ge; # use function call
1383
5a964f20
TC
1384 # expand variables in $_, but dynamics only, using
1385 # symbolic dereferencing
1386 s/\$(\w+)/${$1}/g;
1387
cec88af6
GS
1388 # Add one to the value of any numbers in the string
1389 s/(\d+)/1 + $1/eg;
1390
1391 # This will expand any embedded scalar variable
1392 # (including lexicals) in $_ : First $1 is interpolated
1393 # to the variable name, and then evaluated
a0d0e21e
LW
1394 s/(\$\w+)/$1/eeg;
1395
5a964f20 1396 # Delete (most) C comments.
a0d0e21e 1397 $program =~ s {
4633a7c4
LW
1398 /\* # Match the opening delimiter.
1399 .*? # Match a minimal number of characters.
1400 \*/ # Match the closing delimiter.
a0d0e21e
LW
1401 } []gsx;
1402
6b0ac556 1403 s/^\s*(.*?)\s*$/$1/; # trim whitespace in $_, expensively
5a964f20 1404
6b0ac556 1405 for ($variable) { # trim whitespace in $variable, cheap
5a964f20
TC
1406 s/^\s+//;
1407 s/\s+$//;
1408 }
a0d0e21e
LW
1409
1410 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
1411
54310121 1412Note the use of $ instead of \ in the last example. Unlike
35f2feb0
GS
1413B<sed>, we use the \<I<digit>> form in only the left hand side.
1414Anywhere else it's $<I<digit>>.
a0d0e21e 1415
5f05dabc 1416Occasionally, you can't use just a C</g> to get all the changes
19799a22 1417to occur that you might want. Here are two common cases:
a0d0e21e
LW
1418
1419 # put commas in the right places in an integer
19799a22 1420 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
a0d0e21e
LW
1421
1422 # expand tabs to 8-column spacing
1423 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
1424
6940069f 1425=item tr/SEARCHLIST/REPLACEMENTLIST/cds
a0d0e21e 1426
6940069f 1427=item y/SEARCHLIST/REPLACEMENTLIST/cds
a0d0e21e 1428
2c268ad5 1429Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
1430with the corresponding character in the replacement list. It returns
1431the number of characters replaced or deleted. If no string is
2c268ad5 1432specified via the =~ or !~ operator, the $_ string is transliterated. (The
54310121
PP
1433string specified with =~ must be a scalar variable, an array element, a
1434hash element, or an assignment to one of those, i.e., an lvalue.)
8ada0baa 1435
2c268ad5
TP
1436A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
1437does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121
PP
1438For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1439SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1440its own pair of quotes, which may or may not be bracketing quotes,
2c268ad5 1441e.g., C<tr[A-Z][a-z]> or C<tr(+\-*/)/ABCD/>.
a0d0e21e 1442
cc255d5f
JH
1443Note that C<tr> does B<not> do regular expression character classes
1444such as C<\d> or C<[:lower:]>. The <tr> operator is not equivalent to
1445the tr(1) utility. If you want to map strings between lower/upper
1446cases, see L<perlfunc/lc> and L<perlfunc/uc>, and in general consider
1447using the C<s> operator if you need regular expressions.
1448
8ada0baa
JH
1449Note also that the whole range idea is rather unportable between
1450character sets--and even within character sets they may cause results
1451you probably didn't expect. A sound principle is to use only ranges
1452that begin from and end at either alphabets of equal case (a-e, A-E),
1453or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1454character sets in full.
1455
a0d0e21e
LW
1456Options:
1457
1458 c Complement the SEARCHLIST.
1459 d Delete found but unreplaced characters.
1460 s Squash duplicate replaced characters.
1461
19799a22
GS
1462If the C</c> modifier is specified, the SEARCHLIST character set
1463is complemented. If the C</d> modifier is specified, any characters
1464specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
1465(Note that this is slightly more flexible than the behavior of some
1466B<tr> programs, which delete anything they find in the SEARCHLIST,
1467period.) If the C</s> modifier is specified, sequences of characters
1468that were transliterated to the same character are squashed down
1469to a single instance of the character.
a0d0e21e
LW
1470
1471If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1472exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1473than the SEARCHLIST, the final character is replicated till it is long
5a964f20 1474enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
a0d0e21e
LW
1475This latter is useful for counting characters in a class or for
1476squashing character sequences in a class.
1477
1478Examples:
1479
1480 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1481
1482 $cnt = tr/*/*/; # count the stars in $_
1483
1484 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
1485
1486 $cnt = tr/0-9//; # count the digits in $_
1487
1488 tr/a-zA-Z//s; # bookkeeper -> bokeper
1489
1490 ($HOST = $host) =~ tr/a-z/A-Z/;
1491
1492 tr/a-zA-Z/ /cs; # change non-alphas to single space
1493
1494 tr [\200-\377]
1495 [\000-\177]; # delete 8th bit
1496
19799a22
GS
1497If multiple transliterations are given for a character, only the
1498first one is used:
748a9306
LW
1499
1500 tr/AAA/XYZ/
1501
2c268ad5 1502will transliterate any A to X.
748a9306 1503
19799a22 1504Because the transliteration table is built at compile time, neither
a0d0e21e 1505the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
19799a22
GS
1506interpolation. That means that if you want to use variables, you
1507must use an eval():
a0d0e21e
LW
1508
1509 eval "tr/$oldlist/$newlist/";
1510 die $@ if $@;
1511
1512 eval "tr/$oldlist/$newlist/, 1" or die $@;
1513
7e3b091d
DA
1514=item <<EOF
1515
1516A line-oriented form of quoting is based on the shell "here-document"
1517syntax. Following a C<< << >> you specify a string to terminate
1518the quoted material, and all lines following the current line down to
1519the terminating string are the value of the item. The terminating
1520string may be either an identifier (a word), or some quoted text. If
1521quoted, the type of quotes you use determines the treatment of the
1522text, just as in regular quoting. An unquoted identifier works like
1523double quotes. There must be no space between the C<< << >> and
1524the identifier, unless the identifier is quoted. (If you put a space it
1525will be treated as a null identifier, which is valid, and matches the first
1526empty line.) The terminating string must appear by itself (unquoted and
1527with no surrounding whitespace) on the terminating line.
1528
1529 print <<EOF;
1530 The price is $Price.
1531 EOF
1532
1533 print << "EOF"; # same as above
1534 The price is $Price.
1535 EOF
1536
1537 print << `EOC`; # execute commands
1538 echo hi there
1539 echo lo there
1540 EOC
1541
1542 print <<"foo", <<"bar"; # you can stack them
1543 I said foo.
1544 foo
1545 I said bar.
1546 bar
1547
1548 myfunc(<< "THIS", 23, <<'THAT');
1549 Here's a line
1550 or two.
1551 THIS
1552 and here's another.
1553 THAT
1554
1555Just don't forget that you have to put a semicolon on the end
1556to finish the statement, as Perl doesn't know you're not going to
1557try to do this:
1558
1559 print <<ABC
1560 179231
1561 ABC
1562 + 20;
1563
1564If you want your here-docs to be indented with the
1565rest of the code, you'll need to remove leading whitespace
1566from each line manually:
1567
1568 ($quote = <<'FINIS') =~ s/^\s+//gm;
1569 The Road goes ever on and on,
1570 down from the door where it began.
1571 FINIS
1572
1573If you use a here-doc within a delimited construct, such as in C<s///eg>,
1574the quoted material must come on the lines following the final delimiter.
1575So instead of
1576
1577 s/this/<<E . 'that'
1578 the other
1579 E
1580 . 'more '/eg;
1581
1582you have to write
1583
1584 s/this/<<E . 'that'
1585 . 'more '/eg;
1586 the other
1587 E
1588
1589If the terminating identifier is on the last line of the program, you
1590must be sure there is a newline after it; otherwise, Perl will give the
1591warning B<Can't find string terminator "END" anywhere before EOF...>.
1592
1593Additionally, the quoting rules for the identifier are not related to
1594Perl's quoting rules -- C<q()>, C<qq()>, and the like are not supported
1595in place of C<''> and C<"">, and the only interpolation is for backslashing
1596the quoting character:
1597
1598 print << "abc\"def";
1599 testing...
1600 abc"def
1601
1602Finally, quoted strings cannot span multiple lines. The general rule is
1603that the identifier must be a string literal. Stick with that, and you
1604should be safe.
1605
a0d0e21e
LW
1606=back
1607
75e14d17
IZ
1608=head2 Gory details of parsing quoted constructs
1609
19799a22
GS
1610When presented with something that might have several different
1611interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
1612principle to pick the most probable interpretation. This strategy
1613is so successful that Perl programmers often do not suspect the
1614ambivalence of what they write. But from time to time, Perl's
1615notions differ substantially from what the author honestly meant.
1616
1617This section hopes to clarify how Perl handles quoted constructs.
1618Although the most common reason to learn this is to unravel labyrinthine
1619regular expressions, because the initial steps of parsing are the
1620same for all quoting operators, they are all discussed together.
1621
1622The most important Perl parsing rule is the first one discussed
1623below: when processing a quoted construct, Perl first finds the end
1624of that construct, then interprets its contents. If you understand
1625this rule, you may skip the rest of this section on the first
1626reading. The other rules are likely to contradict the user's
1627expectations much less frequently than this first one.
1628
1629Some passes discussed below are performed concurrently, but because
1630their results are the same, we consider them individually. For different
1631quoting constructs, Perl performs different numbers of passes, from
1632one to five, but these passes are always performed in the same order.
75e14d17 1633
13a2d996 1634=over 4
75e14d17
IZ
1635
1636=item Finding the end
1637
19799a22
GS
1638The first pass is finding the end of the quoted construct, whether
1639it be a multicharacter delimiter C<"\nEOF\n"> in the C<<<EOF>
1640construct, a C</> that terminates a C<qq//> construct, a C<]> which
35f2feb0
GS
1641terminates C<qq[]> construct, or a C<< > >> which terminates a
1642fileglob started with C<< < >>.
75e14d17 1643
19799a22
GS
1644When searching for single-character non-pairing delimiters, such
1645as C</>, combinations of C<\\> and C<\/> are skipped. However,
1646when searching for single-character pairing delimiter like C<[>,
1647combinations of C<\\>, C<\]>, and C<\[> are all skipped, and nested
1648C<[>, C<]> are skipped as well. When searching for multicharacter
1649delimiters, nothing is skipped.
75e14d17 1650
19799a22
GS
1651For constructs with three-part delimiters (C<s///>, C<y///>, and
1652C<tr///>), the search is repeated once more.
75e14d17 1653
19799a22
GS
1654During this search no attention is paid to the semantics of the construct.
1655Thus:
75e14d17
IZ
1656
1657 "$hash{"$foo/$bar"}"
1658
2a94b7ce 1659or:
75e14d17
IZ
1660
1661 m/
2a94b7ce 1662 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
1663 /x
1664
19799a22
GS
1665do not form legal quoted expressions. The quoted part ends on the
1666first C<"> and C</>, and the rest happens to be a syntax error.
1667Because the slash that terminated C<m//> was followed by a C<SPACE>,
1668the example above is not C<m//x>, but rather C<m//> with no C</x>
1669modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 1670
0d594e51
ST
1671Also no attention is paid to C<\c\> during this search.
1672Thus the second C<\> in C<qq/\c\/> is interpreted as a part of C<\/>,
1673and the following C</> is not recognized as a delimiter.
1674Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
1675
75e14d17
IZ
1676=item Removal of backslashes before delimiters
1677
19799a22
GS
1678During the second pass, text between the starting and ending
1679delimiters is copied to a safe location, and the C<\> is removed
1680from combinations consisting of C<\> and delimiter--or delimiters,
1681meaning both starting and ending delimiters will should these differ.
1682This removal does not happen for multi-character delimiters.
1683Note that the combination C<\\> is left intact, just as it was.
75e14d17 1684
19799a22
GS
1685Starting from this step no information about the delimiters is
1686used in parsing.
75e14d17
IZ
1687
1688=item Interpolation
1689
19799a22
GS
1690The next step is interpolation in the text obtained, which is now
1691delimiter-independent. There are four different cases.
75e14d17 1692
13a2d996 1693=over 4
75e14d17
IZ
1694
1695=item C<<<'EOF'>, C<m''>, C<s'''>, C<tr///>, C<y///>
1696
1697No interpolation is performed.
1698
1699=item C<''>, C<q//>
1700
1701The only interpolation is removal of C<\> from pairs C<\\>.
1702
35f2feb0 1703=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>
75e14d17 1704
19799a22
GS
1705C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
1706converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
1707is converted to C<$foo . (quotemeta("baz" . $bar))> internally.
1708The other combinations are replaced with appropriate expansions.
2a94b7ce 1709
19799a22
GS
1710Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
1711is interpolated in the usual way. Something like C<"\Q\\E"> has
1712no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
1713result is the same as for C<"\\\\E">. As a general rule, backslashes
1714between C<\Q> and C<\E> may lead to counterintuitive results. So,
1715C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
1716as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
1717
1718 $str = '\t';
1719 return "\Q$str";
1720
1721may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
1722
19799a22 1723Interpolated scalars and arrays are converted internally to the C<join> and
92d29cee 1724C<.> catenation operations. Thus, C<"$foo XXX '@arr'"> becomes:
75e14d17 1725
19799a22 1726 $foo . " XXX '" . (join $", @arr) . "'";
75e14d17 1727
19799a22 1728All operations above are performed simultaneously, left to right.
75e14d17 1729
19799a22
GS
1730Because the result of C<"\Q STRING \E"> has all metacharacters
1731quoted, there is no way to insert a literal C<$> or C<@> inside a
1732C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to became
1733C<"\\\$">; if not, it is interpreted as the start of an interpolated
1734scalar.
75e14d17 1735
19799a22
GS
1736Note also that the interpolation code needs to make a decision on
1737where the interpolated scalar ends. For instance, whether
35f2feb0 1738C<< "a $b -> {c}" >> really means:
75e14d17
IZ
1739
1740 "a " . $b . " -> {c}";
1741
2a94b7ce 1742or:
75e14d17
IZ
1743
1744 "a " . $b -> {c};
1745
19799a22
GS
1746Most of the time, the longest possible text that does not include
1747spaces between components and which contains matching braces or
1748brackets. because the outcome may be determined by voting based
1749on heuristic estimators, the result is not strictly predictable.
1750Fortunately, it's usually correct for ambiguous cases.
75e14d17
IZ
1751
1752=item C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
1753
19799a22
GS
1754Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
1755happens (almost) as with C<qq//> constructs, but the substitution
1756of C<\> followed by RE-special chars (including C<\>) is not
1757performed. Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
1758a C<#>-comment in a C<//x>-regular expression, no processing is
1759performed whatsoever. This is the first step at which the presence
1760of the C<//x> modifier is relevant.
1761
1762Interpolation has several quirks: C<$|>, C<$(>, and C<$)> are not
1763interpolated, and constructs C<$var[SOMETHING]> are voted (by several
1764different estimators) to be either an array element or C<$var>
1765followed by an RE alternative. This is where the notation
1766C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
1767array element C<-9>, not as a regular expression from the variable
1768C<$arr> followed by a digit, which would be the interpretation of
1769C</$arr[0-9]/>. Since voting among different estimators may occur,
1770the result is not predictable.
1771
1772It is at this step that C<\1> is begrudgingly converted to C<$1> in
1773the replacement text of C<s///> to correct the incorrigible
1774I<sed> hackers who haven't picked up the saner idiom yet. A warning
9f1b1f2d
GS
1775is emitted if the C<use warnings> pragma or the B<-w> command-line flag
1776(that is, the C<$^W> variable) was set.
19799a22
GS
1777
1778The lack of processing of C<\\> creates specific restrictions on
1779the post-processed text. If the delimiter is C</>, one cannot get
1780the combination C<\/> into the result of this step. C</> will
1781finish the regular expression, C<\/> will be stripped to C</> on
1782the previous step, and C<\\/> will be left as is. Because C</> is
1783equivalent to C<\/> inside a regular expression, this does not
1784matter unless the delimiter happens to be character special to the
1785RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
1786alphanumeric char, as in:
2a94b7ce
IZ
1787
1788 m m ^ a \s* b mmx;
1789
19799a22 1790In the RE above, which is intentionally obfuscated for illustration, the
2a94b7ce 1791delimiter is C<m>, the modifier is C<mx>, and after backslash-removal the
aa863641 1792RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
19799a22
GS
1793reason you're encouraged to restrict your delimiters to non-alphanumeric,
1794non-whitespace choices.
75e14d17
IZ
1795
1796=back
1797
19799a22 1798This step is the last one for all constructs except regular expressions,
75e14d17
IZ
1799which are processed further.
1800
1801=item Interpolation of regular expressions
1802
19799a22
GS
1803Previous steps were performed during the compilation of Perl code,
1804but this one happens at run time--although it may be optimized to
1805be calculated at compile time if appropriate. After preprocessing
1806described above, and possibly after evaluation if catenation,
1807joining, casing translation, or metaquoting are involved, the
1808resulting I<string> is passed to the RE engine for compilation.
1809
1810Whatever happens in the RE engine might be better discussed in L<perlre>,
1811but for the sake of continuity, we shall do so here.
1812
1813This is another step where the presence of the C<//x> modifier is
1814relevant. The RE engine scans the string from left to right and
1815converts it to a finite automaton.
1816
1817Backslashed characters are either replaced with corresponding
1818literal strings (as with C<\{>), or else they generate special nodes
1819in the finite automaton (as with C<\b>). Characters special to the
1820RE engine (such as C<|>) generate corresponding nodes or groups of
1821nodes. C<(?#...)> comments are ignored. All the rest is either
1822converted to literal strings to match, or else is ignored (as is
1823whitespace and C<#>-style comments if C<//x> is present).
1824
1825Parsing of the bracketed character class construct, C<[...]>, is
1826rather different than the rule used for the rest of the pattern.
1827The terminator of this construct is found using the same rules as
1828for finding the terminator of a C<{}>-delimited construct, the only
1829exception being that C<]> immediately following C<[> is treated as
1830though preceded by a backslash. Similarly, the terminator of
1831C<(?{...})> is found using the same rules as for finding the
1832terminator of a C<{}>-delimited construct.
1833
1834It is possible to inspect both the string given to RE engine and the
1835resulting finite automaton. See the arguments C<debug>/C<debugcolor>
1836in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 1837switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
1838
1839=item Optimization of regular expressions
1840
7522fed5 1841This step is listed for completeness only. Since it does not change
75e14d17 1842semantics, details of this step are not documented and are subject
19799a22
GS
1843to change without notice. This step is performed over the finite
1844automaton that was generated during the previous pass.
2a94b7ce 1845
19799a22
GS
1846It is at this stage that C<split()> silently optimizes C</^/> to
1847mean C</^/m>.
75e14d17
IZ
1848
1849=back
1850
a0d0e21e
LW
1851=head2 I/O Operators
1852
54310121 1853There are several I/O operators you should know about.
fbad3eb5 1854
7b8d334a 1855A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
1856double-quote interpolation. It is then interpreted as an external
1857command, and the output of that command is the value of the
e9c56f9b
JH
1858backtick string, like in a shell. In scalar context, a single string
1859consisting of all output is returned. In list context, a list of
1860values is returned, one per line of output. (You can set C<$/> to use
1861a different line terminator.) The command is executed each time the
1862pseudo-literal is evaluated. The status value of the command is
1863returned in C<$?> (see L<perlvar> for the interpretation of C<$?>).
1864Unlike in B<csh>, no translation is done on the return data--newlines
1865remain newlines. Unlike in any of the shells, single quotes do not
1866hide variable names in the command from interpretation. To pass a
1867literal dollar-sign through to the shell you need to hide it with a
1868backslash. The generalized form of backticks is C<qx//>. (Because
1869backticks always undergo shell expansion as well, see L<perlsec> for
1870security concerns.)
19799a22
GS
1871
1872In scalar context, evaluating a filehandle in angle brackets yields
1873the next line from that file (the newline, if any, included), or
1874C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
1875(sometimes known as file-slurp mode) and the file is empty, it
1876returns C<''> the first time, followed by C<undef> subsequently.
1877
1878Ordinarily you must assign the returned value to a variable, but
1879there is one situation where an automatic assignment happens. If
1880and only if the input symbol is the only thing inside the conditional
1881of a C<while> statement (even if disguised as a C<for(;;)> loop),
1882the value is automatically assigned to the global variable $_,
1883destroying whatever was there previously. (This may seem like an
1884odd thing to you, but you'll use the construct in almost every Perl
17b829fa 1885script you write.) The $_ variable is not implicitly localized.
19799a22
GS
1886You'll have to put a C<local $_;> before the loop if you want that
1887to happen.
1888
1889The following lines are equivalent:
a0d0e21e 1890
748a9306 1891 while (defined($_ = <STDIN>)) { print; }
7b8d334a 1892 while ($_ = <STDIN>) { print; }
a0d0e21e
LW
1893 while (<STDIN>) { print; }
1894 for (;<STDIN>;) { print; }
748a9306 1895 print while defined($_ = <STDIN>);
7b8d334a 1896 print while ($_ = <STDIN>);
a0d0e21e
LW
1897 print while <STDIN>;
1898
19799a22 1899This also behaves similarly, but avoids $_ :
7b8d334a
GS
1900
1901 while (my $line = <STDIN>) { print $line }
1902
19799a22
GS
1903In these loop constructs, the assigned value (whether assignment
1904is automatic or explicit) is then tested to see whether it is
1905defined. The defined test avoids problems where line has a string
1906value that would be treated as false by Perl, for example a "" or
1907a "0" with no trailing newline. If you really mean for such values
1908to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
1909
1910 while (($_ = <STDIN>) ne '0') { ... }
1911 while (<STDIN>) { last unless $_; ... }
1912
35f2feb0 1913In other boolean contexts, C<< <I<filehandle>> >> without an
9f1b1f2d
GS
1914explicit C<defined> test or comparison elicit a warning if the
1915C<use warnings> pragma or the B<-w>
19799a22 1916command-line switch (the C<$^W> variable) is in effect.
7b8d334a 1917
5f05dabc 1918The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
1919filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
1920in packages, where they would be interpreted as local identifiers
1921rather than global.) Additional filehandles may be created with
1922the open() function, amongst others. See L<perlopentut> and
1923L<perlfunc/open> for details on this.
a0d0e21e 1924
35f2feb0 1925If a <FILEHANDLE> is used in a context that is looking for
19799a22
GS
1926a list, a list comprising all input lines is returned, one line per
1927list element. It's easy to grow to a rather large data space this
1928way, so use with care.
a0d0e21e 1929
35f2feb0 1930<FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
19799a22 1931See L<perlfunc/readline>.
fbad3eb5 1932
35f2feb0
GS
1933The null filehandle <> is special: it can be used to emulate the
1934behavior of B<sed> and B<awk>. Input from <> comes either from
a0d0e21e 1935standard input, or from each file listed on the command line. Here's
35f2feb0 1936how it works: the first time <> is evaluated, the @ARGV array is
5a964f20 1937checked, and if it is empty, C<$ARGV[0]> is set to "-", which when opened
a0d0e21e
LW
1938gives you standard input. The @ARGV array is then processed as a list
1939of filenames. The loop
1940
1941 while (<>) {
1942 ... # code for each line
1943 }
1944
1945is equivalent to the following Perl-like pseudo code:
1946
3e3baf6d 1947 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
1948 while ($ARGV = shift) {
1949 open(ARGV, $ARGV);
1950 while (<ARGV>) {
1951 ... # code for each line
1952 }
1953 }
1954
19799a22
GS
1955except that it isn't so cumbersome to say, and will actually work.
1956It really does shift the @ARGV array and put the current filename
1957into the $ARGV variable. It also uses filehandle I<ARGV>
35f2feb0 1958internally--<> is just a synonym for <ARGV>, which
19799a22 1959is magical. (The pseudo code above doesn't work because it treats
35f2feb0 1960<ARGV> as non-magical.)
a0d0e21e 1961
35f2feb0 1962You can modify @ARGV before the first <> as long as the array ends up
a0d0e21e 1963containing the list of filenames you really want. Line numbers (C<$.>)
19799a22
GS
1964continue as though the input were one big happy file. See the example
1965in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20
TC
1966
1967If you want to set @ARGV to your own list of files, go right ahead.
1968This sets @ARGV to all plain text files if no @ARGV was given:
1969
1970 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 1971
5a964f20
TC
1972You can even set them to pipe commands. For example, this automatically
1973filters compressed arguments through B<gzip>:
1974
1975 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
1976
1977If you want to pass switches into your script, you can use one of the
a0d0e21e
LW
1978Getopts modules or put a loop on the front like this:
1979
1980 while ($_ = $ARGV[0], /^-/) {
1981 shift;
1982 last if /^--$/;
1983 if (/^-D(.*)/) { $debug = $1 }
1984 if (/^-v/) { $verbose++ }
5a964f20 1985 # ... # other switches
a0d0e21e 1986 }
5a964f20 1987
a0d0e21e 1988 while (<>) {
5a964f20 1989 # ... # code for each line
a0d0e21e
LW
1990 }
1991
35f2feb0 1992The <> symbol will return C<undef> for end-of-file only once.
19799a22
GS
1993If you call it again after this, it will assume you are processing another
1994@ARGV list, and if you haven't set @ARGV, will read input from STDIN.
a0d0e21e 1995
b159ebd3 1996If what the angle brackets contain is a simple scalar variable (e.g.,
35f2feb0 1997<$foo>), then that variable contains the name of the
19799a22
GS
1998filehandle to input from, or its typeglob, or a reference to the
1999same. For example:
cb1a09d0
AD
2000
2001 $fh = \*STDIN;
2002 $line = <$fh>;
a0d0e21e 2003
5a964f20
TC
2004If what's within the angle brackets is neither a filehandle nor a simple
2005scalar variable containing a filehandle name, typeglob, or typeglob
2006reference, it is interpreted as a filename pattern to be globbed, and
2007either a list of filenames or the next filename in the list is returned,
19799a22 2008depending on context. This distinction is determined on syntactic
35f2feb0
GS
2009grounds alone. That means C<< <$x> >> is always a readline() from
2010an indirect handle, but C<< <$hash{key}> >> is always a glob().
5a964f20 2011That's because $x is a simple scalar variable, but C<$hash{key}> is
ef191992
YST
2012not--it's a hash element. Even C<< <$x > >> (note the extra space)
2013is treated as C<glob("$x ")>, not C<readline($x)>.
5a964f20
TC
2014
2015One level of double-quote interpretation is done first, but you can't
35f2feb0 2016say C<< <$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
2017in the previous paragraph. (In older versions of Perl, programmers
2018would insert curly brackets to force interpretation as a filename glob:
35f2feb0 2019C<< <${foo}> >>. These days, it's considered cleaner to call the
5a964f20 2020internal function directly as C<glob($foo)>, which is probably the right
19799a22 2021way to have done it in the first place.) For example:
a0d0e21e
LW
2022
2023 while (<*.c>) {
2024 chmod 0644, $_;
2025 }
2026
3a4b19e4 2027is roughly equivalent to:
a0d0e21e
LW
2028
2029 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
2030 while (<FOO>) {
5b3eff12 2031 chomp;
a0d0e21e
LW
2032 chmod 0644, $_;
2033 }
2034
3a4b19e4
GS
2035except that the globbing is actually done internally using the standard
2036C<File::Glob> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
2037
2038 chmod 0644, <*.c>;
2039
19799a22
GS
2040A (file)glob evaluates its (embedded) argument only when it is
2041starting a new list. All values must be read before it will start
2042over. In list context, this isn't important because you automatically
2043get them all anyway. However, in scalar context the operator returns
069e01df 2044the next value each time it's called, or C<undef> when the list has
19799a22
GS
2045run out. As with filehandle reads, an automatic C<defined> is
2046generated when the glob occurs in the test part of a C<while>,
2047because legal glob returns (e.g. a file called F<0>) would otherwise
2048terminate the loop. Again, C<undef> is returned only once. So if
2049you're expecting a single value from a glob, it is much better to
2050say
4633a7c4
LW
2051
2052 ($file) = <blurch*>;
2053
2054than
2055
2056 $file = <blurch*>;
2057
2058because the latter will alternate between returning a filename and
19799a22 2059returning false.
4633a7c4 2060
b159ebd3 2061If you're trying to do variable interpolation, it's definitely better
4633a7c4 2062to use the glob() function, because the older notation can cause people
e37d713d 2063to become confused with the indirect filehandle notation.
4633a7c4
LW
2064
2065 @files = glob("$dir/*.[ch]");
2066 @files = glob($files[$i]);
2067
a0d0e21e
LW
2068=head2 Constant Folding
2069
2070Like C, Perl does a certain amount of expression evaluation at
19799a22 2071compile time whenever it determines that all arguments to an
a0d0e21e
LW
2072operator are static and have no side effects. In particular, string
2073concatenation happens at compile time between literals that don't do
19799a22 2074variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
2075compile time. You can say
2076
2077 'Now is the time for all' . "\n" .
2078 'good men to come to.'
2079
54310121 2080and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
2081you say
2082
2083 foreach $file (@filenames) {
5a964f20 2084 if (-s $file > 5 + 100 * 2**16) { }
54310121 2085 }
a0d0e21e 2086
19799a22
GS
2087the compiler will precompute the number which that expression
2088represents so that the interpreter won't have to.
a0d0e21e 2089
fd1abbef
DN
2090=head2 No-ops
2091
2092Perl doesn't officially have a no-op operator, but the bare constants
2093C<0> and C<1> are special-cased to not produce a warning in a void
2094context, so you can for example safely do
2095
2096 1 while foo();
2097
2c268ad5
TP
2098=head2 Bitwise String Operators
2099
2100Bitstrings of any size may be manipulated by the bitwise operators
2101(C<~ | & ^>).
2102
19799a22
GS
2103If the operands to a binary bitwise op are strings of different
2104sizes, B<|> and B<^> ops act as though the shorter operand had
2105additional zero bits on the right, while the B<&> op acts as though
2106the longer operand were truncated to the length of the shorter.
2107The granularity for such extension or truncation is one or more
2108bytes.
2c268ad5
TP
2109
2110 # ASCII-based examples
2111 print "j p \n" ^ " a h"; # prints "JAPH\n"
2112 print "JA" | " ph\n"; # prints "japh\n"
2113 print "japh\nJunk" & '_____'; # prints "JAPH\n";
2114 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
2115
19799a22 2116If you are intending to manipulate bitstrings, be certain that
2c268ad5 2117you're supplying bitstrings: If an operand is a number, that will imply
19799a22 2118a B<numeric> bitwise operation. You may explicitly show which type of
2c268ad5
TP
2119operation you intend by using C<""> or C<0+>, as in the examples below.
2120
2121 $foo = 150 | 105 ; # yields 255 (0x96 | 0x69 is 0xFF)
2122 $foo = '150' | 105 ; # yields 255
2123 $foo = 150 | '105'; # yields 255
2124 $foo = '150' | '105'; # yields string '155' (under ASCII)
2125
2126 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
2127 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
a0d0e21e 2128
1ae175c8
GS
2129See L<perlfunc/vec> for information on how to manipulate individual bits
2130in a bit vector.
2131
55497cff 2132=head2 Integer Arithmetic
a0d0e21e 2133
19799a22 2134By default, Perl assumes that it must do most of its arithmetic in
a0d0e21e
LW
2135floating point. But by saying
2136
2137 use integer;
2138
2139you may tell the compiler that it's okay to use integer operations
19799a22
GS
2140(if it feels like it) from here to the end of the enclosing BLOCK.
2141An inner BLOCK may countermand this by saying
a0d0e21e
LW
2142
2143 no integer;
2144
19799a22
GS
2145which lasts until the end of that BLOCK. Note that this doesn't
2146mean everything is only an integer, merely that Perl may use integer
2147operations if it is so inclined. For example, even under C<use
2148integer>, if you take the C<sqrt(2)>, you'll still get C<1.4142135623731>
2149or so.
2150
2151Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
13a2d996
SP
2152and ">>") always produce integral results. (But see also
2153L<Bitwise String Operators>.) However, C<use integer> still has meaning for
19799a22
GS
2154them. By default, their results are interpreted as unsigned integers, but
2155if C<use integer> is in effect, their results are interpreted
2156as signed integers. For example, C<~0> usually evaluates to a large
2157integral value. However, C<use integer; ~0> is C<-1> on twos-complement
2158machines.
68dc0745
PP
2159
2160=head2 Floating-point Arithmetic
2161
2162While C<use integer> provides integer-only arithmetic, there is no
19799a22
GS
2163analogous mechanism to provide automatic rounding or truncation to a
2164certain number of decimal places. For rounding to a certain number
2165of digits, sprintf() or printf() is usually the easiest route.
2166See L<perlfaq4>.
68dc0745 2167
5a964f20
TC
2168Floating-point numbers are only approximations to what a mathematician
2169would call real numbers. There are infinitely more reals than floats,
2170so some corners must be cut. For example:
2171
2172 printf "%.20g\n", 123456789123456789;
2173 # produces 123456789123456784
2174
2175Testing for exact equality of floating-point equality or inequality is
2176not a good idea. Here's a (relatively expensive) work-around to compare
2177whether two floating-point numbers are equal to a particular number of
2178decimal places. See Knuth, volume II, for a more robust treatment of
2179this topic.
2180
2181 sub fp_equal {
2182 my ($X, $Y, $POINTS) = @_;
2183 my ($tX, $tY);
2184 $tX = sprintf("%.${POINTS}g", $X);
2185 $tY = sprintf("%.${POINTS}g", $Y);
2186 return $tX eq $tY;
2187 }
2188
68dc0745 2189The POSIX module (part of the standard perl distribution) implements
19799a22
GS
2190ceil(), floor(), and other mathematical and trigonometric functions.
2191The Math::Complex module (part of the standard perl distribution)
2192defines mathematical functions that work on both the reals and the
2193imaginary numbers. Math::Complex not as efficient as POSIX, but
68dc0745
PP
2194POSIX can't work with complex numbers.
2195
2196Rounding in financial applications can have serious implications, and
2197the rounding method used should be specified precisely. In these
2198cases, it probably pays not to trust whichever system rounding is
2199being used by Perl, but to instead implement the rounding function you
2200need yourself.
5a964f20
TC
2201
2202=head2 Bigger Numbers
2203
2204The standard Math::BigInt and Math::BigFloat modules provide
19799a22 2205variable-precision arithmetic and overloaded operators, although
cd5c4fce 2206they're currently pretty slow. At the cost of some space and
19799a22
GS
2207considerable speed, they avoid the normal pitfalls associated with
2208limited-precision representations.
5a964f20
TC
2209
2210 use Math::BigInt;
2211 $x = Math::BigInt->new('123456789123456789');
2212 print $x * $x;
2213
2214 # prints +15241578780673678515622620750190521
19799a22 2215
cd5c4fce
T
2216There are several modules that let you calculate with (bound only by
2217memory and cpu-time) unlimited or fixed precision. There are also
2218some non-standard modules that provide faster implementations via
2219external C libraries.
2220
2221Here is a short, but incomplete summary:
2222
2223 Math::Fraction big, unlimited fractions like 9973 / 12967
2224 Math::String treat string sequences like numbers
2225 Math::FixedPrecision calculate with a fixed precision
2226 Math::Currency for currency calculations
2227 Bit::Vector manipulate bit vectors fast (uses C)
2228 Math::BigIntFast Bit::Vector wrapper for big numbers
2229 Math::Pari provides access to the Pari C library
2230 Math::BigInteger uses an external C library
2231 Math::Cephes uses external Cephes C library (no big numbers)
2232 Math::Cephes::Fraction fractions via the Cephes library
2233 Math::GMP another one using an external C library
2234
2235Choose wisely.
16070b82
GS
2236
2237=cut