This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
CommitLineData
d74e8afc 2X<operator>
a0d0e21e
LW
3
4perlop - Perl operators and precedence
5
d042e63d
MS
7
d74e8afc
ITB
9X<operator, precedence> X<precedence> X<associativity>
d042e63d
MS
10
11Operator precedence and associativity work in Perl more or less like
12they do in mathematics.
13
14I<Operator precedence> means some operators are evaluated before
15others. For example, in C<2 + 4 * 5>, the multiplication has higher
16precedence so C<4 * 5> is evaluated first yielding C<2 + 20 ==
1722> and not C<6 * 5 == 30>.
18
19I<Operator associativity> defines what happens if a sequence of the
20same operators is used one after another: whether the evaluator will
21evaluate the left operations first or the right. For example, in C<8
22- 4 - 2>, subtraction is left associative so Perl evaluates the
23expression left to right. C<8 - 4> is evaluated first making the
24expression C<4 - 2 == 2> and not C<8 - 2 == 6>.
a0d0e21e
LW
25
26Perl operators have the following associativity and precedence,
19799a22
GS
27listed from highest precedence to lowest. Operators borrowed from
28C keep the same precedence relationship with each other, even where
29C's precedence is slightly screwy. (This makes learning Perl easier
30for C folks.) With very few exceptions, these all operate on scalar
31values only, not array values.
a0d0e21e
LW
32
33 left terms and list operators (leftward)
34 left ->
35 nonassoc ++ --
36 right **
37 right ! ~ \ and unary + and -
54310121 38 left =~ !~
a0d0e21e
LW
39 left * / % x
40 left + - .
41 left << >>
42 nonassoc named unary operators
43 nonassoc < > <= >= lt gt le ge
0d863452 44 nonassoc == != <=> eq ne cmp ~~
a0d0e21e
LW
45 left &
46 left | ^
47 left &&
c963b151 48 left || //
137443ea 49 nonassoc .. ...
a0d0e21e
LW
50 right ?:
51 right = += -= *= etc.
52 left , =>
53 nonassoc list operators (rightward)
a5f75d66 54 right not
a0d0e21e 55 left and
c963b151 56 left or xor err
a0d0e21e
LW
57
58In the following sections, these operators are covered in precedence order.
59
5a964f20
TC
61
a0d0e21e 62=head2 Terms and List Operators (Leftward)
d74e8afc 63X<list operator> X<operator, list> X<term>
a0d0e21e 64
62c18ce2 65A TERM has the highest precedence in Perl. They include variables,
5f05dabc 66quote and quote-like operators, any expression in parentheses,
a0d0e21e
LW
67and any function whose arguments are parenthesized. Actually, there
68aren't really functions in this sense, just list operators and unary
69operators behaving as functions because you put parentheses around
70the arguments. These are all documented in L<perlfunc>.
71
72If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
73is followed by a left parenthesis as the next token, the operator and
74arguments within parentheses are taken to be of highest precedence,
75just like a normal function call.
76
77In the absence of parentheses, the precedence of list operators such as
78C<print>, C<sort>, or C<chmod> is either very high or very low depending on
54310121 79whether you are looking at the left side or the right side of the operator.
a0d0e21e
LW
80For example, in
81
82 @ary = (1, 3, sort 4, 2);
83 print @ary; # prints 1324
84
19799a22
GS
85the commas on the right of the sort are evaluated before the sort,
86but the commas on the left are evaluated after. In other words,
87list operators tend to gobble up all arguments that follow, and
a0d0e21e 88then act like a simple TERM with regard to the preceding expression.
19799a22 89Be careful with parentheses:
a0d0e21e
LW
90
91 # These evaluate exit before doing the print:
92 print(\$foo, exit); # Obviously not what you want.
93 print \$foo, exit; # Nor is this.
94
95 # These do the print before evaluating exit:
96 (print \$foo), exit; # This is what you want.
97 print(\$foo), exit; # Or this.
98 print (\$foo), exit; # Or even this.
99
100Also note that
101
102 print (\$foo & 255) + 1, "\n";
103
d042e63d
MS
104probably doesn't do what you expect at first glance. The parentheses
105enclose the argument list for C<print> which is evaluated (printing
106the result of C<\$foo & 255>). Then one is added to the return value
107of C<print> (usually 1). The result is something like this:
108
109 1 + 1, "\n"; # Obviously not what you meant.
110
111To do what you meant properly, you must write:
112
113 print((\$foo & 255) + 1, "\n");
114
115See L<Named Unary Operators> for more discussion of this.
a0d0e21e
LW
116
117Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
54310121 118well as subroutine and method calls, and the anonymous
a0d0e21e
LW
119constructors C<[]> and C<{}>.
120
2ae324a7 121See also L<Quote and Quote-like Operators> toward the end of this section,
da87341d 122as well as L</"I/O Operators">.
a0d0e21e
LW
123
d74e8afc 125X<arrow> X<dereference> X<< -> >>
a0d0e21e 126
35f2feb0 127"C<< -> >>" is an infix dereference operator, just as it is in C
19799a22
GS
128and C++. If the right side is either a C<[...]>, C<{...}>, or a
129C<(...)> subscript, then the left side must be either a hard or
130symbolic reference to an array, a hash, or a subroutine respectively.
131(Or technically speaking, a location capable of holding a hard
132reference, if it's an array or hash reference being used for
133assignment.) See L<perlreftut> and L<perlref>.
a0d0e21e 134
19799a22
GS
135Otherwise, the right side is a method name or a simple scalar
136variable containing either the method name or a subroutine reference,
137and the left side must be either an object (a blessed reference)
138or a class name (that is, a package name). See L<perlobj>.
a0d0e21e 139
d74e8afc 141X<increment> X<auto-increment> X<++> X<decrement> X<auto-decrement> X<-->
a0d0e21e 142
d042e63d
MS
143"++" and "--" work as in C. That is, if placed before a variable,
144they increment or decrement the variable by one before returning the
145value, and if placed after, increment or decrement after returning the
146value.
147
148 \$i = 0; \$j = 0;
149 print \$i++; # prints 0
150 print ++\$j; # prints 1
a0d0e21e 151
b033823e
A
152Note that just as in C, Perl doesn't define B<when> the variable is
153incremented or decremented. You just know it will be done sometime
154before or after the value is returned. This also means that modifying
155a variable twice in the same statement will lead to undefined behaviour.
156Avoid statements like:
157
158 \$i = \$i ++;
159 print ++ \$i + \$i ++;
160
161Perl will not guarantee what the result of the above statements is.
162
54310121 163The auto-increment operator has a little extra builtin magic to it. If
a0d0e21e
LW
164you increment a variable that is numeric, or that has ever been used in
165a numeric context, you get a normal increment. If, however, the
5f05dabc 166variable has been used in only string contexts since it was set, and
5a964f20 167has a value that is not the empty string and matches the pattern
9c0670e1 168C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
a0d0e21e
LW
169character within its range, with carry:
170
171 print ++(\$foo = '99'); # prints '100'
172 print ++(\$foo = 'a0'); # prints 'a1'
173 print ++(\$foo = 'Az'); # prints 'Ba'
174 print ++(\$foo = 'zz'); # prints 'aaa'
175
6a61d433
HS
176C<undef> is always treated as numeric, and in particular is changed
177to C<0> before incrementing (so that a post-increment of an undef value
178will return C<0> rather than C<undef>).
179
5f05dabc 180The auto-decrement operator is not magical.
a0d0e21e
LW
181
d74e8afc 183X<**> X<exponentiation> X<power>
a0d0e21e 184
19799a22 185Binary "**" is the exponentiation operator. It binds even more
cb1a09d0
186tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
187implemented using C's pow(3) function, which actually works on doubles
188internally.)
a0d0e21e
LW
189
d74e8afc 191X<unary operator> X<operator, unary>
a0d0e21e 192
5f05dabc 193Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
a0d0e21e 194precedence version of this.
d74e8afc 195X<!>
a0d0e21e
LW
196
197Unary "-" performs arithmetic negation if the operand is numeric. If
198the operand is an identifier, a string consisting of a minus sign
199concatenated with the identifier is returned. Otherwise, if the string
200starts with a plus or minus, a string starting with the opposite sign
bff5667c 201is returned. One effect of these rules is that -bareword is equivalent
8705167b 202to the string "-bareword". If, however, the string begins with a
06705523
SP
203non-alphabetic character (exluding "+" or "-"), Perl will attempt to convert
204the string to a numeric and the arithmetic negation is performed. If the
205string cannot be cleanly converted to a numeric, Perl will give the warning
206B<Argument "the string" isn't numeric in negation (-) at ...>.
d74e8afc 207X<-> X<negation, arithmetic>
a0d0e21e 208
972b05a9
JH
209Unary "~" performs bitwise negation, i.e., 1's complement. For
211L<Bitwise String Operators>.) Note that the width of the result is
212platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
213bits wide on a 64-bit platform, so if you are expecting a certain bit
d042e63d 214width, remember to use the & operator to mask off the excess bits.
d74e8afc 215X<~> X<negation, binary>
a0d0e21e
LW
216
217Unary "+" has no effect whatsoever, even on strings. It is useful
218syntactically for separating a function name from a parenthesized expression
219that would otherwise be interpreted as the complete list of function
5ba421f6 220arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
d74e8afc 221X<+>
a0d0e21e 222
19799a22
GS
223Unary "\" creates a reference to whatever follows it. See L<perlreftut>
224and L<perlref>. Do not confuse this behavior with the behavior of
225backslash within a string, although both forms do convey the notion
226of protecting the next thing from interpolation.
d74e8afc 227X<\> X<reference> X<backslash>
a0d0e21e
LW
228
d74e8afc 230X<binding> X<operator, binding> X<=~> X<!~>
a0d0e21e 231
c07a80fd 232Binary "=~" binds a scalar expression to a pattern match. Certain operations
cb1a09d0
233search or modify the string \$_ by default. This operator makes that kind
234of operation work on some other string. The right argument is a search
TP
235pattern, substitution, or transliteration. The left argument is what is
236supposed to be searched, substituted, or transliterated instead of the default
f8bab1e9
GS
237\$_. When used in scalar context, the return value generally indicates the
238success of the operation. Behavior in list context depends on the particular
d7782e69
SP
239operator. See L</"Regexp Quote-Like Operators"> for details and
240L<perlretut> for examples using these operators.
f8bab1e9
GS
241
242If the right argument is an expression rather than a search pattern,
2c268ad5 243substitution, or transliteration, it is interpreted as a search pattern at run
573e01ca 244time.
a0d0e21e
LW
245
246Binary "!~" is just like "=~" except the return value is negated in
247the logical sense.
248
d74e8afc 250X<operator, multiplicative>
a0d0e21e
LW
251
252Binary "*" multiplies two numbers.
d74e8afc 253X<*>
a0d0e21e
LW
254
255Binary "/" divides two numbers.
d74e8afc 256X</> X<slash>
a0d0e21e 257
54310121 258Binary "%" computes the modulus of two numbers. Given integer
259operands C<\$a> and C<\$b>: If C<\$b> is positive, then C<\$a % \$b> is
260C<\$a> minus the largest multiple of C<\$b> that is not greater than
261C<\$a>. If C<\$b> is negative, then C<\$a % \$b> is C<\$a> minus the
262smallest multiple of C<\$b> that is not less than C<\$a> (i.e. the
SP
263result will be less than or equal to zero). If the operands
264C<\$a> and C<\$b> are floting point values, only the integer portion
265of C<\$a> and C<\$b> will be used in the operation.
0412d526 266Note that when C<use integer> is in scope, "%" gives you direct access
55d729e4
GS
267to the modulus operator as implemented by your C compiler. This
268operator is not as well defined for negative operands, but it will
269execute faster.
d74e8afc 270X<%> X<remainder> X<modulus> X<mod>
55d729e4 271
62d10b70
GS
272Binary "x" is the repetition operator. In scalar context or if the left
273operand is not enclosed in parentheses, it returns a string consisting
274of the left operand repeated the number of times specified by the right
275operand. In list context, if the left operand is enclosed in
3585017f
YST
276parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
277If the right operand is zero or negative, it returns an empty string
278or an empty list, depending on the context.
d74e8afc 279X<x>
a0d0e21e
LW
280
281 print '-' x 80; # print row of dashes
282
283 print "\t" x (\$tab/8), ' ' x (\$tab%8); # tab over
284
285 @ones = (1) x 80; # a list of 80 1's
286 @ones = (5) x @ones; # set all elements to 5
287
288
a0d0e21e
LW
291
292Binary "+" returns the sum of two numbers.
d74e8afc 293X<+>
a0d0e21e
LW
294
295Binary "-" returns the difference of two numbers.
d74e8afc 296X<->
a0d0e21e
LW
297
298Binary "." concatenates two strings.
d74e8afc
ITB
299X<string, concatenation> X<concatenation>
300X<cat> X<concat> X<concatenate> X<.>
a0d0e21e
LW
301
d74e8afc
ITB
303X<shift operator> X<operator, shift> X<<< << >>>
304X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
305X<shl> X<shr> X<shift, right> X<shift, left>
a0d0e21e 306
55497cff 307Binary "<<" returns the value of its left argument shifted left by the
308number of bits specified by the right argument. Arguments should be
a0d0e21e 310
55497cff 311Binary ">>" returns the value of its left argument shifted right by
312the number of bits specified by the right argument. Arguments should
a0d0e21e 314
b16cf6df
JH
315Note that both "<<" and ">>" in Perl are implemented directly using
316"<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
317in force then signed C integers are used, else unsigned C integers are
318used. Either way, the implementation isn't going to generate results
319larger than the size of the integer type Perl was built with (32 bits
320or 64 bits).
321
322The result of overflowing the range of the integers is undefined
323because it is undefined also in C. In other words, using 32-bit
324integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
325of bits is also undefined.
326
d74e8afc 328X<operator, named unary>
a0d0e21e
LW
329
330The various named unary operators are treated as functions with one
568e6d8b 331argument, with optional parentheses.
a0d0e21e
LW
332
333If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
334is followed by a left parenthesis as the next token, the operator and
335arguments within parentheses are taken to be of highest precedence,
3981b0eb
JA
336just like a normal function call. For example,
337because named unary operators are higher precedence than ||:
a0d0e21e
LW
338
339 chdir \$foo || die; # (chdir \$foo) || die
340 chdir(\$foo) || die; # (chdir \$foo) || die
341 chdir (\$foo) || die; # (chdir \$foo) || die
342 chdir +(\$foo) || die; # (chdir \$foo) || die
343
3981b0eb 344but, because * is higher precedence than named operators:
a0d0e21e
LW
345
346 chdir \$foo * 20; # chdir (\$foo * 20)
347 chdir(\$foo) * 20; # (chdir \$foo) * 20
348 chdir (\$foo) * 20; # (chdir \$foo) * 20
349 chdir +(\$foo) * 20; # chdir (\$foo * 20)
350
351 rand 10 * 20; # rand (10 * 20)
352 rand(10) * 20; # (rand 10) * 20
353 rand (10) * 20; # (rand 10) * 20
354 rand +(10) * 20; # rand (10 * 20)
355
568e6d8b
RGS
356Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
357treated like named unary operators, but they don't follow this functional
358parenthesis rule. That means, for example, that C<-f(\$file).".bak"> is
359equivalent to C<-f "\$file.bak">.
d74e8afc 360X<-X> X<filetest> X<operator, filetest>
568e6d8b 361
a0d0e21e
LW
363
d74e8afc 365X<relational operator> X<operator, relational>
a0d0e21e 366
35f2feb0 367Binary "<" returns true if the left argument is numerically less than
a0d0e21e 368the right argument.
d74e8afc 369X<< < >>
a0d0e21e 370
35f2feb0 371Binary ">" returns true if the left argument is numerically greater
a0d0e21e 372than the right argument.
d74e8afc 373X<< > >>
a0d0e21e 374
35f2feb0 375Binary "<=" returns true if the left argument is numerically less than
a0d0e21e 376or equal to the right argument.
d74e8afc 377X<< <= >>
a0d0e21e 378
35f2feb0 379Binary ">=" returns true if the left argument is numerically greater
a0d0e21e 380than or equal to the right argument.
d74e8afc 381X<< >= >>
a0d0e21e
LW
382
383Binary "lt" returns true if the left argument is stringwise less than
384the right argument.
d74e8afc 385X<< lt >>
a0d0e21e
LW
386
387Binary "gt" returns true if the left argument is stringwise greater
388than the right argument.
d74e8afc 389X<< gt >>
a0d0e21e
LW
390
391Binary "le" returns true if the left argument is stringwise less than
392or equal to the right argument.
d74e8afc 393X<< le >>
a0d0e21e
LW
394
395Binary "ge" returns true if the left argument is stringwise greater
396than or equal to the right argument.
d74e8afc 397X<< ge >>
a0d0e21e
LW
398
d74e8afc 400X<equality> X<equal> X<equals> X<operator, equality>
a0d0e21e
LW
401
402Binary "==" returns true if the left argument is numerically equal to
403the right argument.
d74e8afc 404X<==>
a0d0e21e
LW
405
406Binary "!=" returns true if the left argument is numerically not equal
407to the right argument.
d74e8afc 408X<!=>
a0d0e21e 409
35f2feb0 410Binary "<=>" returns -1, 0, or 1 depending on whether the left
6ee5d4e7 411argument is numerically less than, equal to, or greater than the right
7d3a9d88
NC
413values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
414"<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
415returns true, as does NaN != anything else. If your platform doesn't
416support NaNs then NaN is just a string with numeric value 0.
d74e8afc 417X<< <=> >> X<spaceship>
7d3a9d88 418
2b54f59f
YST
419 perl -le '\$a = "NaN"; print "No NaN support here" if \$a == \$a'
420 perl -le '\$a = "NaN"; print "NaN support here" if \$a != \$a'
a0d0e21e
LW
421
422Binary "eq" returns true if the left argument is stringwise equal to
423the right argument.
d74e8afc 424X<eq>
a0d0e21e
LW
425
426Binary "ne" returns true if the left argument is stringwise not equal
427to the right argument.
d74e8afc 428X<ne>
a0d0e21e 429
JH
430Binary "cmp" returns -1, 0, or 1 depending on whether the left
431argument is stringwise less than, equal to, or greater than the right
432argument.
d74e8afc 433X<cmp>
a0d0e21e 434
0d863452
RH
435Binary "~~" does a smart match between its arguments. Smart matching
436is described in L<perlsyn/"Smart Matching in Detail">.
437This operator is only available if you enable the "~~" feature:
439X<~~>
440
a034a98d
DD
441"lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
442by the current locale if C<use locale> is in effect. See L<perllocale>.
443
d74e8afc 445X<operator, bitwise, and> X<bitwise and> X<&>
a0d0e21e 446
2cdc098b 447Binary "&" returns its operands ANDed together bit by bit.
a0d0e21e 449
2cdc098b
MG
450Note that "&" has lower priority than relational operators, so for example
451the brackets are essential in a test like
452
453 print "Even\n" if (\$x & 1) == 0;
454
a0d0e21e 455=head2 Bitwise Or and Exclusive Or
d74e8afc
ITB
456X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
457X<bitwise xor> X<^>
a0d0e21e 458
2cdc098b 459Binary "|" returns its operands ORed together bit by bit.
a0d0e21e 461
2cdc098b 462Binary "^" returns its operands XORed together bit by bit.
a0d0e21e 464
2cdc098b
MG
465Note that "|" and "^" have lower priority than relational operators, so
466for example the brackets are essential in a test like
467
468 print "false\n" if (8 | 2) != 10;
469
d74e8afc 471X<&&> X<logical and> X<operator, logical, and>
a0d0e21e
LW
472
473Binary "&&" performs a short-circuit logical AND operation. That is,
474if the left operand is false, the right operand is not even evaluated.
475Scalar or list context propagates down to the right operand if it
476is evaluated.
477
d74e8afc 479X<||> X<operator, logical, or>
a0d0e21e
LW
480
481Binary "||" performs a short-circuit logical OR operation. That is,
482if the left operand is true, the right operand is not even evaluated.
483Scalar or list context propagates down to the right operand if it
484is evaluated.
485
d74e8afc 487X<//> X<operator, logical, defined-or>
c963b151
BD
488
489Although it has no direct equivalent in C, Perl's C<//> operator is related
490to its C-style or. In fact, it's exactly the same as C<||>, except that it
491tests the left hand side's definedness instead of its truth. Thus, C<\$a // \$b>
492is similar to C<defined(\$a) || \$b> (except that it returns the value of C<\$a>
493rather than the value of C<defined(\$a)>) and is exactly equivalent to
494C<defined(\$a) ? \$a : \$b>. This is very useful for providing default values
d042e63d
MS
495for variables. If you actually want to test if at least one of C<\$a> and
496C<\$b> is defined, use C<defined(\$a // \$b)>.
c963b151 497
d042e63d
MS
498The C<||>, C<//> and C<&&> operators return the last value evaluated
499(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
500portable way to find out the home directory might be:
a0d0e21e 501
c963b151
BD
502 \$home = \$ENV{'HOME'} // \$ENV{'LOGDIR'} //
503 (getpwuid(\$<))[7] // die "You're homeless!\n";
a0d0e21e 504
5a964f20
TC
505In particular, this means that you shouldn't use this
506for selecting between two aggregates for assignment:
507
508 @a = @b || @c; # this is wrong
509 @a = scalar(@b) || @c; # really meant this
510 @a = @b ? @b : @c; # this works fine, though
511
c963b151
BD
512As more readable alternatives to C<&&>, C<//> and C<||> when used for
513control flow, Perl provides C<and>, C<err> and C<or> operators (see below).
514The short-circuit behavior is identical. The precedence of "and", "err"
515and "or" is much lower, however, so that you can safely use them after a
5a964f20 516list operator without the need for parentheses:
a0d0e21e
LW
517
519 or gripe(), next LINE;
520
521With the C-style operators that would have been written like this:
522
524 || (gripe(), next LINE);
525
eeb6a2c9 526Using "or" for assignment is unlikely to do what you want; see below.
5a964f20
TC
527
d74e8afc 529X<operator, range> X<range> X<..> X<...>
a0d0e21e
LW
530
531Binary ".." is the range operator, which is really two different
fb53bbb2 532operators depending on the context. In list context, it returns a
54ae734e 533list of values counting (up by ones) from the left value to the right
2cdbc966 534value. If the left value is greater than the right value then it
fb53bbb2 535returns the empty list. The range operator is useful for writing
54ae734e 536C<foreach (1..10)> loops and for doing slice operations on arrays. In
2cdbc966
JD
537the current implementation, no temporary array is created when the
538range operator is used as the expression in C<foreach> loops, but older
539versions of Perl might burn a lot of memory when you write something
540like this:
a0d0e21e
LW
541
542 for (1 .. 1_000_000) {
543 # code
54310121 544 }
a0d0e21e 545
54ae734e
MG
546The range operator also works on strings, using the magical auto-increment,
547see below.
548
5a964f20 549In scalar context, ".." returns a boolean value. The operator is
a0d0e21e
LW
550bistable, like a flip-flop, and emulates the line-range (comma) operator
551of B<sed>, B<awk>, and various editors. Each ".." operator maintains its
552own boolean state. It is false as long as its left operand is false.
553Once the left operand is true, the range operator stays true until the
554right operand is true, I<AFTER> which the range operator becomes false
19799a22 555again. It doesn't become false till the next time the range operator is
a0d0e21e
LW
556evaluated. It can test the right operand and become false on the same
557evaluation it became true (as in B<awk>), but it still returns true once.
19799a22
GS
558If you don't want it to test the right operand till the next
559evaluation, as in B<sed>, just use three dots ("...") instead of
560two. In all other regards, "..." behaves just like ".." does.
561
562The right operand is not evaluated while the operator is in the
563"false" state, and the left operand is not evaluated while the
564operator is in the "true" state. The precedence is a little lower
565than || and &&. The value returned is either the empty string for
566false, or a sequence number (beginning with 1) for true. The
567sequence number is reset for each range encountered. The final
568sequence number in a range has the string "E0" appended to it, which
569doesn't affect its numeric value, but gives you something to search
570for if you want to exclude the endpoint. You can exclude the
571beginning point by waiting for the sequence number to be greater
df5f8116
CW
572than 1.
573
574If either operand of scalar ".." is a constant expression,
575that operand is considered true if it is equal (C<==>) to the current
576input line number (the C<\$.> variable).
577
578To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
579but that is only an issue if you use a floating point expression; when
580implicitly using C<\$.> as described in the previous paragraph, the
581comparison is C<int(EXPR) == int(\$.)> which is only an issue when C<\$.>
582is set to a floating point value and you are not reading from a file.
583Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
584you want in scalar context because each of the operands are evaluated
585using their integer representation.
586
587Examples:
a0d0e21e
LW
588
589As a scalar operator:
590
df5f8116
CW
591 if (101 .. 200) { print; } # print 2nd hundred lines, short for
592 # if (\$. == 101 .. \$. == 200) ...
9f10b797
RGS
593
594 next LINE if (1 .. /^\$/); # skip header lines, short for
df5f8116 595 # ... if (\$. == 1 .. /^\$/);
9f10b797
RGS
596 # (typically in a loop labeled LINE)
597
598 s/^/> / if (/^\$/ .. eof()); # quote body
a0d0e21e 599
5a964f20
TC
600 # parse mail messages
601 while (<>) {
602 \$in_header = 1 .. /^\$/;
df5f8116
CW
603 \$in_body = /^\$/ .. eof;
605 # ...
606 } else { # in body
607 # ...
608 }
5a964f20 609 } continue {
df5f8116 610 close ARGV if eof; # reset \$. each file
5a964f20
TC
611 }
612
acf31ca5
SF
613Here's a simple example to illustrate the difference between
614the two range operators:
615
616 @lines = (" - Foo",
617 "01 - Bar",
618 "1 - Baz",
619 " - Quux");
620
9f10b797
RGS
621 foreach (@lines) {
622 if (/0/ .. /1/) {
acf31ca5
SF
623 print "\$_\n";
624 }
625 }
626
9f10b797
RGS
627This program will print only the line containing "Bar". If
628the range operator is changed to C<...>, it will also print the
acf31ca5
SF
629"Baz" line.
630
631And now some examples as a list operator:
a0d0e21e
LW
632
633 for (101 .. 200) { print; } # print \$_ 100 times
3e3baf6d 634 @foo = @foo[0 .. \$#foo]; # an expensive no-op
a0d0e21e
LW
635 @foo = @foo[\$#foo-4 .. \$#foo]; # slice last 5 items
636
5a964f20 637The range operator (in list context) makes use of the magical
5f05dabc 638auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
639can say
640
641 @alphabet = ('A' .. 'Z');
642
54ae734e 643to get all normal letters of the English alphabet, or
a0d0e21e
LW
644
645 \$hexdigit = (0 .. 9, 'a' .. 'f')[\$num & 15];
646
647to get a hexadecimal digit, or
648
649 @z2 = ('01' .. '31'); print \$z2[\$mday];
650
651to get dates with leading zeros. If the final value specified is not
652in the sequence that the magical increment would produce, the sequence
653goes until the next value would be longer than the final value
654specified.
655
df5f8116
CW
656Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
657return two elements in list context.
658
659 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
660
d74e8afc 662X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
a0d0e21e
LW
663
664Ternary "?:" is the conditional operator, just as in C. It works much
665like an if-then-else. If the argument before the ? is true, the
666argument before the : is returned, otherwise the argument after the :
cb1a09d0
667is returned. For example:
668
54310121 669 printf "I have %d dog%s.\n", \$n,
cb1a09d0
670 (\$n == 1) ? '' : "s";
671
672Scalar or list context propagates downward into the 2nd
54310121 673or 3rd argument, whichever is selected.
cb1a09d0
674
675 \$a = \$ok ? \$b : \$c; # get a scalar
676 @a = \$ok ? @b : @c; # get an array
677 \$a = \$ok ? @b : @c; # oops, that's just a count!
678
679The operator may be assigned to if both the 2nd and 3rd arguments are
680legal lvalues (meaning that you can assign to them):
a0d0e21e
LW
681
682 (\$a_or_b ? \$a : \$b) = \$c;
683
5a964f20
TC
684Because this operator produces an assignable result, using assignments
685without parentheses will get you in trouble. For example, this:
686
687 \$a % 2 ? \$a += 10 : \$a += 2
688
689Really means this:
690
691 ((\$a % 2) ? (\$a += 10) : \$a) += 2
692
693Rather than this:
694
695 (\$a % 2) ? (\$a += 10) : (\$a += 2)
696
19799a22
GS
697That should probably be written more simply as:
698
699 \$a += (\$a % 2) ? 10 : 2;
700
d74e8afc 702X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
5ac3b81c 703X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
d74e8afc 704X<%=> X<^=> X<x=>
a0d0e21e
LW
705
706"=" is the ordinary assignment operator.
707
708Assignment operators work as in C. That is,
709
710 \$a += 2;
711
712is equivalent to
713
714 \$a = \$a + 2;
715
716although without duplicating any side effects that dereferencing the lvalue
54310121 717might trigger, such as from tie(). Other assignment operators work similarly.
718The following are recognized:
a0d0e21e
LW
719
720 **= += *= &= <<= &&=
9f10b797
RGS
721 -= /= |= >>= ||=
722 .= %= ^= //=
723 x=
a0d0e21e 724
19799a22 725Although these are grouped by family, they all have the precedence
a0d0e21e
LW
726of assignment.
727
b350dd2f
GS
728Unlike in C, the scalar assignment operator produces a valid lvalue.
729Modifying an assignment is equivalent to doing the assignment and
730then modifying the variable that was assigned to. This is useful
731for modifying a copy of something, like this:
a0d0e21e
LW
732
733 (\$tmp = \$global) =~ tr [A-Z] [a-z];
734
735Likewise,
736
737 (\$a += 2) *= 3;
738
739is equivalent to
740
741 \$a += 2;
742 \$a *= 3;
743
b350dd2f
GS
744Similarly, a list assignment in list context produces the list of
745lvalues assigned to, and a list assignment in scalar context returns
746the number of elements produced by the expression on the right hand
747side of the assignment.
748
d74e8afc 750X<comma> X<operator, comma> X<,>
a0d0e21e 751
5a964f20 752Binary "," is the comma operator. In scalar context it evaluates
a0d0e21e
LW
753its left argument, throws that value away, then evaluates its right
754argument and returns that value. This is just like C's comma operator.
755
5a964f20 756In list context, it's just the list argument separator, and inserts
ed5c6d31
PJ
757both its arguments into the list. These arguments are also evaluated
758from left to right.
a0d0e21e 759
d042e63d 760The C<< => >> operator is a synonym for the comma, but forces any word
719b43e8 761(consisting entirely of word characters) to its left to be interpreted
a44e5664
MS
762as a string (as of 5.001). This includes words that might otherwise be
763considered a constant or function call.
764
765 use constant FOO => "something";
766
767 my %h = ( FOO => 23 );
768
769is equivalent to:
770
771 my %h = ("FOO", 23);
772
773It is I<NOT>:
774
775 my %h = ("something", 23);
776
777If the argument on the left is not a word, it is first interpreted as
778an expression, and then the string value of that is used.
719b43e8
RGS
779
780The C<< => >> operator is helpful in documenting the correspondence
781between keys and values in hashes, and other paired elements in lists.
748a9306 782
a44e5664
MS
783 %hash = ( \$key => \$value );
785
d74e8afc 787X<operator, list, rightward> X<list operator>
a0d0e21e
LW
788
789On the right side of a list operator, it has very low precedence,
790such that it controls all comma-separated expressions found there.
791The only operators with lower precedence are the logical operators
792"and", "or", and "not", which may be used to evaluate calls to list
793operators without the need for extra parentheses:
794
795 open HANDLE, "filename"
796 or die "Can't open: \$!\n";
797
5ba421f6 798See also discussion of list operators in L<Terms and List Operators (Leftward)>.
a0d0e21e
LW
799
d74e8afc 801X<operator, logical, not> X<not>
a0d0e21e
LW
802
803Unary "not" returns the logical negation of the expression to its right.
804It's the equivalent of "!" except for the very low precedence.
805
d74e8afc 807X<operator, logical, and> X<and>
a0d0e21e
LW
808
809Binary "and" returns the logical conjunction of the two surrounding
810expressions. It's equivalent to && except for the very low
5f05dabc 811precedence. This means that it short-circuits: i.e., the right
a0d0e21e
LW
812expression is evaluated only if the left expression is true.
813
c963b151 814=head2 Logical or, Defined or, and Exclusive Or
d74e8afc
ITB
815X<operator, logical, or> X<operator, logical, xor> X<operator, logical, err>
816X<operator, logical, defined or> X<operator, logical, exclusive or>
817X<or> X<xor> X<err>
a0d0e21e
LW
818
819Binary "or" returns the logical disjunction of the two surrounding
5a964f20
TC
820expressions. It's equivalent to || except for the very low precedence.
821This makes it useful for control flow
822
823 print FH \$data or die "Can't write to FH: \$!";
824
825This means that it short-circuits: i.e., the right expression is evaluated
826only if the left expression is false. Due to its precedence, you should
827probably avoid using this for assignment, only for control flow.
828
829 \$a = \$b or \$c; # bug: this is wrong
830 (\$a = \$b) or \$c; # really means this
831 \$a = \$b || \$c; # better written this way
832
19799a22 833However, when it's a list-context assignment and you're trying to use
5a964f20
TC
834"||" for control flow, you probably need "or" so that the assignment
835takes higher precedence.
836
837 @info = stat(\$file) || die; # oops, scalar sense of stat!
838 @info = stat(\$file) or die; # better, now @info gets its due
839
c963b151
BD
840Then again, you could always use parentheses.
841
9f10b797
RGS
842Binary "err" is equivalent to C<//>--it's just like binary "or", except it
843tests its left argument's definedness instead of its truth. There are two
844ways to remember "err": either because many functions return C<undef> on
845an B<err>or, or as a sort of correction: C<\$a = (\$b err 'default')>. This
846keyword is only available when the 'err' feature is enabled: see
a0d0e21e
LW
848
849Binary "xor" returns the exclusive-OR of the two surrounding expressions.
850It cannot short circuit, of course.
851
852=head2 C Operators Missing From Perl
d74e8afc
ITB
853X<operator, missing from perl> X<&> X<*>
854X<typecasting> X<(TYPE)>
a0d0e21e
LW
855
856Here is what C has that Perl doesn't:
857
858=over 8
859
860=item unary &
861
862Address-of operator. (But see the "\" operator for taking a reference.)
863
864=item unary *
865
54310121 866Dereference-address operator. (Perl's prefix dereferencing
a0d0e21e
LW
867operators are typed: \$, @, %, and &.)
868
869=item (TYPE)
870
19799a22 871Type-casting operator.
a0d0e21e
LW
872
873=back
874
5f05dabc 875=head2 Quote and Quote-like Operators
d74e8afc
ITB
876X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
877X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
878X<escape sequence> X<escape>
879
a0d0e21e
LW
880
881While we usually think of quotes as literal values, in Perl they
882function as operators, providing various kinds of interpolating and
883pattern matching capabilities. Perl provides customary quote characters
884for these behaviors, but also provides a way for you to choose your
885quote character for any of them. In the following table, a C<{}> represents
9f10b797 886any pair of delimiters you choose.
a0d0e21e 887
TP
888 Customary Generic Meaning Interpolates
889 '' q{} Literal no
890 "" qq{} Literal yes
af9219ee 891 `` qx{} Command yes*
2c268ad5 892 qw{} Word list no
af9219ee
MG
893 // m{} Pattern match yes*
894 qr{} Pattern yes*
895 s{}{} Substitution yes*
2c268ad5 896 tr{}{} Transliteration no (but see below)
7e3b091d 897 <<EOF here-doc yes*
a0d0e21e 898
af9219ee
MG
899 * unless the delimiter is ''.
900
87275199
GS
901Non-bracketing delimiters use the same character fore and aft, but the four
902sorts of brackets (round, angle, square, curly) will all nest, which means
9f10b797 903that
87275199 904
9f10b797 905 q{foo{bar}baz}
35f2feb0 906
9f10b797 907is the same as
87275199
GS
908
909 'foo{bar}baz'
910
911Note, however, that this does not always work for quoting Perl code:
912
913 \$s = q{ if(\$a eq "}") ... }; # WRONG
914
83df6a1d
JH
915is a syntax error. The C<Text::Balanced> module (from CPAN, and
916starting from Perl 5.8 part of the standard distribution) is able
917to do this properly.
87275199 918
19799a22 919There can be whitespace between the operator and the quoting
fb73857a 920characters, except when C<#> is being used as the quoting character.
19799a22
GS
921C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
922operator C<q> followed by a comment. Its argument will be taken
923from the next line. This allows you to write:
fb73857a 924
925 s {foo} # Replace foo
926 {bar} # with bar.
927
904501ec
MG
928The following escape sequences are available in constructs that interpolate
929and in transliterations.
d74e8afc 930X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N>
a0d0e21e 931
6ee5d4e7 932 \t tab (HT, TAB)
5a964f20 933 \n newline (NL)
6ee5d4e7 934 \r return (CR)
935 \f form feed (FF)
936 \b backspace (BS)
937 \a alarm (bell) (BEL)
938 \e escape (ESC)
a0ed51b3
LW
939 \033 octal char (ESC)
940 \x1b hex char (ESC)
941 \x{263a} wide hex char (SMILEY)
19799a22 942 \c[ control char (ESC)
95cc3e0c 943 \N{name} named Unicode character
4c77eaa2
AE
945B<NOTE>: Unlike C and other languages, Perl has no \v escape sequence for
946the vertical tab (VT - ASCII 11).
947
904501ec
MG
948The following escape sequences are available in constructs that interpolate
949but not in transliterations.
d74e8afc 950X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q>
904501ec 951
a0d0e21e
LW
952 \l lowercase next char
953 \u uppercase next char
954 \L lowercase till \E
955 \U uppercase till \E
956 \E end case modification
1d2dff63 957 \Q quote non-word characters till \E
a0d0e21e 958
95cc3e0c
JH
959If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
960C<\u> and C<\U> is taken from the current locale. See L<perllocale>.
961If Unicode (for example, C<\N{}> or wide hex characters of 0x100 or
962beyond) is being used, the case map used by C<\l>, C<\L>, C<\u> and
963C<\U> is as defined by Unicode. For documentation of C<\N{name}>,
964see L<charnames>.
a034a98d 965
5a964f20
TC
966All systems use the virtual C<"\n"> to represent a line terminator,
967called a "newline". There is no such thing as an unvarying, physical
19799a22 968newline character. It is only an illusion that the operating system,
5a964f20
TC
969device drivers, C libraries, and Perl all conspire to preserve. Not all
970systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
971on a Mac, these are reversed, and on systems without line terminator,
972printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
973you mean a "newline" for your system, but use the literal ASCII when you
974need an exact character. For example, most networking protocols expect
2a380090 975and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
976and although they often accept just C<"\012">, they seldom tolerate just
977C<"\015">. If you get in the habit of using C<"\n"> for networking,
978you may be burned some day.
d74e8afc
ITB
979X<newline> X<line terminator> X<eol> X<end of line>
980X<\n> X<\r> X<\r\n>
5a964f20 981
904501ec
MG
982For constructs that do interpolate, variables beginning with "C<\$>"
983or "C<@>" are interpolated. Subscripted variables such as C<\$a[3]> or
A
984C<< \$href->{key}[0] >> are also interpolated, as are array and hash slices.
985But method calls such as C<< \$obj->meth >> are not.
af9219ee
MG
986
987Interpolating an array or slice interpolates the elements in order,
988separated by the value of C<\$">, so is equivalent to interpolating
904501ec
MG
989C<join \$", @array>. "Punctuation" arrays such as C<@+> are only
990interpolated if the name is enclosed in braces C<@{+}>.
af9219ee 991
1d2dff63
GS
992You cannot include a literal C<\$> or C<@> within a C<\Q> sequence.
993An unescaped C<\$> or C<@> interpolates the corresponding variable,
994while escaping will cause the literal string C<\\$> to be inserted.
995You'll need to write something like C<m/\Quser\E\@\Qhost/>.
996
a0d0e21e
LW
997Patterns are subject to an additional level of interpretation as a
998regular expression. This is done as a second pass, after variables are
999interpolated, so that regular expressions may be incorporated into the
1000pattern from the variables. If this is not what you want, use C<\Q> to
1001interpolate a variable literally.
1002
19799a22
GS
1003Apart from the behavior described above, Perl does not expand
1004multiple levels of interpolation. In particular, contrary to the
1005expectations of shell programmers, back-quotes do I<NOT> interpolate
1006within double quotes, nor do single quotes impede evaluation of
1007variables when used within double quotes.
a0d0e21e 1008
d74e8afc 1010X<operator, regexp>
cb1a09d0 1011
5f05dabc 1012Here are the quote-like operators that apply to pattern
cb1a09d0
1013matching and related activities.
1014
a0d0e21e
LW
1015=over 8
1016
1017=item ?PATTERN?
d74e8afc 1018X<?>
a0d0e21e
LW
1019
1020This is just like the C</pattern/> search, except that it matches only
1021once between calls to the reset() operator. This is a useful
5f05dabc 1022optimization when you want to see only the first occurrence of
a0d0e21e
LW
1023something in each file of a set of files, for instance. Only C<??>
1024patterns local to the current package are reset.
1025
5a964f20
TC
1026 while (<>) {
1027 if (?^\$?) {
1028 # blank line between header and body
1029 }
1030 } continue {
1031 reset if eof; # clear ?? status for next file
1032 }
1033
483b4840 1034This usage is vaguely deprecated, which means it just might possibly
19799a22
GS
1035be removed in some distant future version of Perl, perhaps somewhere
1036around the year 2168.
a0d0e21e 1037
fb73857a 1038=item m/PATTERN/cgimosx
d74e8afc
ITB
1039X<m> X<operator, match>
1040X<regexp, options> X<regexp> X<regex, options> X<regex>
1041X</c> X</i> X</m> X</o> X</s> X</x>
a0d0e21e 1042
fb73857a 1043=item /PATTERN/cgimosx
a0d0e21e 1044
5a964f20 1045Searches a string for a pattern match, and in scalar context returns
19799a22
GS
1046true if it succeeds, false if it fails. If no string is specified
1047via the C<=~> or C<!~> operator, the \$_ string is searched. (The
1048string specified with C<=~> need not be an lvalue--it may be the
1049result of an expression evaluation, but remember the C<=~> binds
1051discussion of additional considerations that apply when C<use locale>
1052is in effect.
a0d0e21e
LW
1053
1054Options are:
1055
fb73857a 1056 c Do not reset search position on a failed match when /g is in effect.
5f05dabc 1057 g Match globally, i.e., find all occurrences.
a0d0e21e
LW
1058 i Do case-insensitive pattern matching.
1059 m Treat string as multiple lines.
5f05dabc 1060 o Compile pattern only once.
a0d0e21e
LW
1061 s Treat string as single line.
1062 x Use extended regular expressions.
1063
1064If "/" is the delimiter then the initial C<m> is optional. With the C<m>
01ae956f 1065you can use any pair of non-alphanumeric, non-whitespace characters
19799a22
GS
1066as delimiters. This is particularly useful for matching path names
1067that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
7bac28a0 1068the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
19799a22 1069If "'" is the delimiter, no interpolation is performed on the PATTERN.
a0d0e21e
LW
1070
1071PATTERN may contain variables, which will be interpolated (and the
f70b4f9c 1072pattern recompiled) every time the pattern search is evaluated, except
1f247705
GS
1073for when the delimiter is a single quote. (Note that C<\$(>, C<\$)>, and
1074C<\$|> are not interpolated because they look like end-of-string tests.)
f70b4f9c
AB
1075If you want such a pattern to be compiled only once, add a C</o> after
1076the trailing delimiter. This avoids expensive run-time recompilations,
1077and is useful when the value you are interpolating won't change over
1078the life of the script. However, mentioning C</o> constitutes a promise
1079that you won't change the variables in the pattern. If you change them,
a0d0e21e 1081
5a964f20 1082If the PATTERN evaluates to the empty string, the last
d65afb4b
HS
1083I<successfully> matched regular expression is used instead. In this
1084case, only the C<g> and C<c> flags on the empty pattern is honoured -
1085the other flags are taken from the original pattern. If no match has
1086previously succeeded, this will (silently) act instead as a genuine
1087empty pattern (which will always match).
a0d0e21e 1088
c963b151
BD
1089Note that it's possible to confuse Perl into thinking C<//> (the empty
1090regex) is really C<//> (the defined-or operator). Perl is usually pretty
1092C<\$a///> (is that C<(\$a) / (//)> or C<\$a // />?) and C<print \$fh //>
1093(C<print \$fh(//> or C<print(\$fh //>?). In all of these examples, Perl
1094will assume you meant defined-or. If you meant the empty regex, just
1095use parentheses or spaces to disambiguate, or even prefix the empty
1096regex with an C<m> (so C<//> becomes C<m//>).
1097
19799a22 1098If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1099list consisting of the subexpressions matched by the parentheses in the
f7e33566
GS
1100pattern, i.e., (C<\$1>, C<\$2>, C<\$3>...). (Note that here C<\$1> etc. are
1101also set, and that this differs from Perl 4's behavior.) When there are
1102no parentheses in the pattern, the return value is the list C<(1)> for
1103success. With or without parentheses, an empty list is returned upon
1104failure.
a0d0e21e
LW
1105
1106Examples:
1107
1108 open(TTY, '/dev/tty');
1109 <TTY> =~ /^y/i && foo(); # do foo if desired
1110
1111 if (/Version: *([0-9.]*)/) { \$version = \$1; }
1112
1113 next if m#^/usr/spool/uucp#;
1114
1115 # poor man's grep
1116 \$arg = shift;
1117 while (<>) {
1118 print if /\$arg/o; # compile only once
1119 }
1120
1121 if ((\$F1, \$F2, \$Etc) = (\$foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
1122
1123This last example splits \$foo into the first two words and the
5f05dabc 1124remainder of the line, and assigns those three fields to \$F1, \$F2, and
1125\$Etc. The conditional is true if any variables were assigned, i.e., if
a0d0e21e
LW
1126the pattern matched.
1127
19799a22
GS
1128The C</g> modifier specifies global pattern matching--that is,
1129matching as many times as possible within the string. How it behaves
1130depends on the context. In list context, it returns a list of the
1131substrings matched by any capturing parentheses in the regular
1132expression. If there are no parentheses, it returns a list of all
1133the matched strings, as if there were parentheses around the whole
1134pattern.
a0d0e21e 1135
7e86de3e 1136In scalar context, each execution of C<m//g> finds the next match,
19799a22 1137returning true if it matches, and false if there is no further match.
7e86de3e
MG
1138The position after the last match can be read or set using the pos()
1139function; see L<perlfunc/pos>. A failed match normally resets the
1140search position to the beginning of the string, but you can avoid that
1141by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
1142string also resets the search position.
c90c0ff4 1143
1144You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
1145zero-width assertion that matches the exact position where the previous
5d43e42d
DC
1146C<m//g>, if any, left off. Without the C</g> modifier, the C<\G> assertion
1147still anchors at pos(), but the match is of course only attempted once.
1148Using C<\G> without C</g> on a target string that has not previously had a
1149C</g> match applied to it is the same as using the C<\A> assertion to match
fe4b3f22
RGS
1150the beginning of the string. Note also that, currently, C<\G> is only
1151properly supported when anchored at the very beginning of the pattern.
c90c0ff4 1152
1153Examples:
a0d0e21e
LW
1154
1155 # list context
1156 (\$one,\$five,\$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1157
1158 # scalar context
5d43e42d 1159 \$/ = "";
19799a22
GS
1160 while (defined(\$paragraph = <>)) {
1161 while (\$paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
1162 \$sentences++;
a0d0e21e
LW
1163 }
1164 }
1165 print "\$sentences\n";
1166
c90c0ff4 1167 # using m//gc with \G
137443ea 1168 \$_ = "ppooqppqq";
44a8e56a 1169 while (\$i++ < 2) {
1170 print "1: '";
c90c0ff4 1171 print \$1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1172 print "2: '";
c90c0ff4 1173 print \$1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1174 print "3: '";
c90c0ff4 1175 print \$1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1176 }
5d43e42d 1177 print "Final: '\$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a 1178
1179The last example should print:
1180
1181 1: 'oo', pos=4
137443ea 1182 2: 'q', pos=5
44a8e56a 1183 3: 'pp', pos=7
1184 1: '', pos=7
137443ea 1185 2: 'q', pos=8
1186 3: '', pos=8
5d43e42d
DC
1187 Final: 'q', pos=8
1188
1189Notice that the final match matched C<q> instead of C<p>, which a match
1190without the C<\G> anchor would have done. Also note that the final match
1191did not update C<pos> -- C<pos> is only updated on a C</g> match. If the
1192final match did indeed match C<p>, it's a good bet that you're running an
1193older (pre-5.6.0) Perl.
44a8e56a 1194
c90c0ff4 1195A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1196combine several regexps like this to process a string part-by-part,
c90c0ff4 1197doing different actions depending on which regexp matched. Each
1198regexp tries to match where the previous one leaves off.
e7ea3e70 1199
3fe9a6f1 1200 \$_ = <<'EOL';
e7ea3e70 1201 \$url = new URI::URL "http://www/"; die if \$url eq "xXx";
3fe9a6f1 1202 EOL
1203 LOOP:
e7ea3e70 1204 {
c90c0ff4 1205 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
1206 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
1207 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
1208 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
1209 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
1210 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
1211 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
e7ea3e70
IZ
1212 print ". That's all!\n";
1213 }
1214
1215Here is the output (split into several lines):
1216
1217 line-noise lowercase line-noise lowercase UPPERCASE line-noise
1218 UPPERCASE line-noise lowercase line-noise lowercase line-noise
1219 lowercase lowercase line-noise lowercase lowercase line-noise
1220 MiXeD line-noise. That's all!
44a8e56a 1221
a0d0e21e 1222=item q/STRING/
d74e8afc 1223X<q> X<quote, double> X<'> X<''>
a0d0e21e
LW
1224
1225=item C<'STRING'>
1226
19799a22 1227A single-quoted, literal string. A backslash represents a backslash
68dc0745 1228unless followed by the delimiter or another backslash, in which case
1229the delimiter or backslash is interpolated.
a0d0e21e
LW
1230
1231 \$foo = q!I said, "You said, 'She said it.'"!;
1232 \$bar = q('This is it.');
68dc0745 1233 \$baz = '\n'; # a two-character string
a0d0e21e
LW
1234
1235=item qq/STRING/
d74e8afc 1236X<qq> X<quote, double> X<"> X<"">
a0d0e21e
LW
1237
1238=item "STRING"
1239
1240A double-quoted, interpolated string.
1241
1242 \$_ .= qq
1243 (*** The previous line contains the naughty word "\$1".\n)
19799a22 1244 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 1245 \$baz = "\n"; # a one-character string
a0d0e21e 1246
eec2d3df 1247=item qr/STRING/imosx
d74e8afc 1248X<qr> X</i> X</m> X</o> X</s> X</x>
eec2d3df 1249
322edccd 1250This operator quotes (and possibly compiles) its I<STRING> as a regular
19799a22
GS
1251expression. I<STRING> is interpolated the same way as I<PATTERN>
1252in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1253is done. Returns a Perl value which may be used instead of the
1254corresponding C</STRING/imosx> expression.
4b6a7270
IZ
1255
1256For example,
1257
1258 \$rex = qr/my.STRING/is;
1259 s/\$rex/foo/;
1260
1261is equivalent to
1262
1263 s/my.STRING/foo/is;
1264
1265The result may be used as a subpattern in a match:
eec2d3df
GS
1266
1267 \$re = qr/\$pattern/;
0a92e3a8
GS
1268 \$string =~ /foo\${re}bar/; # can be interpolated in other patterns
1269 \$string =~ \$re; # or used standalone
4b6a7270
IZ
1270 \$string =~ /\$re/; # or this way
1271
1272Since Perl may compile the pattern at the moment of execution of qr()
19799a22 1273operator, using qr() may have speed advantages in some situations,
4b6a7270
IZ
1274notably if the result of qr() is used standalone:
1275
1276 sub match {
1277 my \$patterns = shift;
1278 my @compiled = map qr/\$_/i, @\$patterns;
1279 grep {
1280 my \$success = 0;
a7665c5e 1281 foreach my \$pat (@compiled) {
4b6a7270
IZ
1282 \$success = 1, last if /\$pat/;
1283 }
1284 \$success;
1285 } @_;
1286 }
1287
19799a22
GS
1288Precompilation of the pattern into an internal representation at
1289the moment of qr() avoids a need to recompile the pattern every
1290time a match C</\$pat/> is attempted. (Perl has many other internal
1291optimizations, but none would be triggered in the above example if
1292we did not use qr() operator.)
eec2d3df
GS
1293
1294Options are:
1295
1296 i Do case-insensitive pattern matching.
1297 m Treat string as multiple lines.
1298 o Compile pattern only once.
1299 s Treat string as single line.
1300 x Use extended regular expressions.
1301
0a92e3a8
GS
1302See L<perlre> for additional information on valid syntax for STRING, and
1303for a detailed look at the semantics of regular expressions.
1304
a0d0e21e 1305=item qx/STRING/
d74e8afc 1306X<qx> X<`> X<``> X<backtick>
a0d0e21e
LW
1307
1308=item `STRING`
1309
43dd4d21
JH
1310A string which is (possibly) interpolated and then executed as a
1311system command with C</bin/sh> or its equivalent. Shell wildcards,
1312pipes, and redirections will be honored. The collected standard
1313output of the command is returned; standard error is unaffected. In
1314scalar context, it comes back as a single (potentially multi-line)
1315string, or undef if the command failed. In list context, returns a
1316list of lines (however you've defined lines with \$/ or
1317\$INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
5a964f20
TC
1318
1319Because backticks do not affect standard error, use shell file descriptor
1320syntax (assuming the shell supports this) if you care to address this.
1321To capture a command's STDERR and STDOUT together:
a0d0e21e 1322
5a964f20
TC
1323 \$output = `cmd 2>&1`;
1324
1325To capture a command's STDOUT but discard its STDERR:
1326
1327 \$output = `cmd 2>/dev/null`;
1328
1329To capture a command's STDERR but discard its STDOUT (ordering is
1330important here):
1331
1332 \$output = `cmd 2>&1 1>/dev/null`;
1333
1334To exchange a command's STDOUT and STDERR in order to capture the STDERR
1335but leave its STDOUT to come out the old STDERR:
1336
1337 \$output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1338
1339To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
1340to redirect them separately to files, and then read from those files
1341when the program is done:
5a964f20 1342
2359510d 1343 system("program args 1>program.stdout 2>program.stderr");
5a964f20 1344
30398227
SP
1345The STDIN filehandle used by the command is inherited from Perl's STDIN.
1346For example:
1347
1348 open BLAM, "blam" || die "Can't open: \$!";
1349 open STDIN, "<&BLAM";
1350 print `sort`;
1351
1352will print the sorted contents of the file "blam".
1353
5a964f20
TC
1354Using single-quote as a delimiter protects the command from Perl's
1355double-quote interpolation, passing it on to the shell instead:
1356
1357 \$perl_info = qx(ps \$\$); # that's Perl's \$\$
1358 \$shell_info = qx'ps \$\$'; # that's the new shell's \$\$
1359
19799a22 1360How that string gets evaluated is entirely subject to the command
5a964f20
TC
1361interpreter on your system. On most platforms, you will have to protect
1362shell metacharacters if you want them treated literally. This is in
1363practice difficult to do, as it's unclear how to escape which characters.
1364See L<perlsec> for a clean and safe example of a manual fork() and exec()
1365to emulate backticks safely.
a0d0e21e 1366
bb32b41a
GS
1367On some platforms (notably DOS-like ones), the shell may not be
1368capable of dealing with multiline commands, so putting newlines in
1369the string may not get you what you want. You may be able to evaluate
1370multiple commands in a single line by separating them with the command
1371separator character, if your shell supports that (e.g. C<;> on many Unix
1372shells; C<&> on the Windows NT C<cmd> shell).
1373
0f897271
GS
1374Beginning with v5.6.0, Perl will attempt to flush all files opened for
1375output before starting the child process, but this may not be supported
1376on some platforms (see L<perlport>). To be safe, you may need to set
1377C<\$|> (\$AUTOFLUSH in English) or call the C<autoflush()> method of
1378C<IO::Handle> on any open handles.
1379
bb32b41a
GS
1380Beware that some command shells may place restrictions on the length
1381of the command line. You must ensure your strings don't exceed this
1382limit after any necessary interpolations. See the platform-specific
1384
5a964f20
TC
1385Using this operator can lead to programs that are difficult to port,
1386because the shell commands called vary between systems, and may in
1387fact not be present at all. As one example, the C<type> command under
1388the POSIX shell is very different from the C<type> command under DOS.
1389That doesn't mean you should go out of your way to avoid backticks
1390when they're the right way to get something done. Perl was made to be
1391a glue language, and one of the things it glues together is commands.
1392Just understand what you're getting yourself into.
bb32b41a 1393
da87341d 1394See L</"I/O Operators"> for more discussion.
a0d0e21e 1395
945c54fd 1396=item qw/STRING/
d74e8afc 1397X<qw> X<quote, list> X<quote, words>
945c54fd
JH
1398
1399Evaluates to a list of the words extracted out of STRING, using embedded
1400whitespace as the word delimiters. It can be understood as being roughly
1401equivalent to:
1402
1403 split(' ', q/STRING/);
1404
efb1e162
CW
1405the differences being that it generates a real list at compile time, and
1406in scalar context it returns the last element in the list. So
945c54fd
JH
1407this expression:
1408
1409 qw(foo bar baz)
1410
1411is semantically equivalent to the list:
1412
1413 'foo', 'bar', 'baz'
1414
1415Some frequently seen examples:
1416
1417 use POSIX qw( setlocale localeconv )
1418 @EXPORT = qw( foo bar baz );
1419
1420A common mistake is to try to separate the words with comma or to
1421put comments into a multi-line C<qw>-string. For this reason, the
1422C<use warnings> pragma and the B<-w> switch (that is, the C<\$^W> variable)
1423produces warnings if the STRING contains the "," or the "#" character.
1424
a0d0e21e 1425=item s/PATTERN/REPLACEMENT/egimosx
d74e8afc
ITB
1426X<substitute> X<substitution> X<replace> X<regexp, replace>
1427X<regexp, substitute> X</e> X</g> X</i> X</m> X</o> X</s> X</x>
a0d0e21e
LW
1428
1429Searches a string for a pattern, and if found, replaces that pattern
1430with the replacement text and returns the number of substitutions
e37d713d 1431made. Otherwise it returns false (specifically, the empty string).
a0d0e21e
LW
1432
1433If no string is specified via the C<=~> or C<!~> operator, the C<\$_>
1434variable is searched and modified. (The string specified with C<=~> must
5a964f20 1435be scalar variable, an array element, a hash element, or an assignment
5f05dabc 1436to one of those, i.e., an lvalue.)
a0d0e21e 1437
19799a22 1438If the delimiter chosen is a single quote, no interpolation is
a0d0e21e
LW
1439done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1440PATTERN contains a \$ that looks like a variable rather than an
1441end-of-string test, the variable will be interpolated into the pattern
5f05dabc 1442at run-time. If you want the pattern compiled only once the first time
a0d0e21e 1443the variable is interpolated, use the C</o> option. If the pattern
5a964f20 1444evaluates to the empty string, the last successfully executed regular
a0d0e21e 1445expression is used instead. See L<perlre> for further explanation on these.
5a964f20 1446See L<perllocale> for discussion of additional considerations that apply
a034a98d 1447when C<use locale> is in effect.
a0d0e21e
LW
1448
1449Options are:
1450
1451 e Evaluate the right side as an expression.
5f05dabc 1452 g Replace globally, i.e., all occurrences.
a0d0e21e
LW
1453 i Do case-insensitive pattern matching.
1454 m Treat string as multiple lines.
5f05dabc 1455 o Compile pattern only once.
a0d0e21e
LW
1456 s Treat string as single line.
1457 x Use extended regular expressions.
1458
1459Any non-alphanumeric, non-whitespace delimiter may replace the
1460slashes. If single quotes are used, no interpretation is done on the
e37d713d 1461replacement string (the C</e> modifier overrides this, however). Unlike
54310121 1462Perl 4, Perl 5 treats backticks as normal delimiters; the replacement
e37d713d 1463text is not evaluated as a command. If the
a0d0e21e 1464PATTERN is delimited by bracketing quotes, the REPLACEMENT has its own
5f05dabc 1465pair of quotes, which may or may not be bracketing quotes, e.g.,
35f2feb0 1466C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
cec88af6
GS
1467replacement portion to be treated as a full-fledged Perl expression
1468and evaluated right then and there. It is, however, syntax checked at
1469compile-time. A second C<e> modifier will cause the replacement portion
1470to be C<eval>ed before being run as a Perl expression.
a0d0e21e
LW
1471
1472Examples:
1473
1474 s/\bgreen\b/mauve/g; # don't change wintergreen
1475
1476 \$path =~ s|/usr/bin|/usr/local/bin|;
1477
1479
5a964f20 1480 (\$foo = \$bar) =~ s/this/that/; # copy first, then change
a0d0e21e 1481
5a964f20 1482 \$count = (\$paragraph =~ s/Mister\b/Mr./g); # get change-count
a0d0e21e
LW
1483
1484 \$_ = 'abc123xyz';
1485 s/\d+/\$&*2/e; # yields 'abc246xyz'
1486 s/\d+/sprintf("%5d",\$&)/e; # yields 'abc 246xyz'
1487 s/\w/\$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1488
1489 s/%(.)/\$percent{\$1}/g; # change percent escapes; no /e
1490 s/%(.)/\$percent{\$1} || \$&/ge; # expr now, so /e
1491 s/^=(\w+)/&pod(\$1)/ge; # use function call
1492
5a964f20
TC
1493 # expand variables in \$_, but dynamics only, using
1494 # symbolic dereferencing
1495 s/\\$(\w+)/\${\$1}/g;
1496
cec88af6
GS
1497 # Add one to the value of any numbers in the string
1498 s/(\d+)/1 + \$1/eg;
1499
1500 # This will expand any embedded scalar variable
1501 # (including lexicals) in \$_ : First \$1 is interpolated
1502 # to the variable name, and then evaluated
a0d0e21e
LW
1503 s/(\\$\w+)/\$1/eeg;
1504
5a964f20 1505 # Delete (most) C comments.
a0d0e21e 1506 \$program =~ s {
4633a7c4
LW
1507 /\* # Match the opening delimiter.
1508 .*? # Match a minimal number of characters.
1509 \*/ # Match the closing delimiter.
a0d0e21e
LW
1510 } []gsx;
1511
6b0ac556 1512 s/^\s*(.*?)\s*\$/\$1/; # trim whitespace in \$_, expensively
5a964f20 1513
6b0ac556 1514 for (\$variable) { # trim whitespace in \$variable, cheap
5a964f20
TC
1515 s/^\s+//;
1516 s/\s+\$//;
1517 }
a0d0e21e
LW
1518
1519 s/([^ ]*) *([^ ]*)/\$2 \$1/; # reverse 1st two fields
1520
54310121 1521Note the use of \$ instead of \ in the last example. Unlike
35f2feb0
GS
1522B<sed>, we use the \<I<digit>> form in only the left hand side.
1523Anywhere else it's \$<I<digit>>.
a0d0e21e 1524
5f05dabc 1525Occasionally, you can't use just a C</g> to get all the changes
19799a22 1526to occur that you might want. Here are two common cases:
a0d0e21e
LW
1527
1528 # put commas in the right places in an integer
19799a22 1529 1 while s/(\d)(\d\d\d)(?!\d)/\$1,\$2/g;
a0d0e21e
LW
1530
1531 # expand tabs to 8-column spacing
1532 1 while s/\t+/' ' x (length(\$&)*8 - length(\$`)%8)/e;
1533
6940069f 1534=item tr/SEARCHLIST/REPLACEMENTLIST/cds
d74e8afc 1535X<tr> X<y> X<transliterate> X</c> X</d> X</s>
a0d0e21e 1536
6940069f 1537=item y/SEARCHLIST/REPLACEMENTLIST/cds
a0d0e21e 1538
2c268ad5 1539Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
1540with the corresponding character in the replacement list. It returns
1541the number of characters replaced or deleted. If no string is
2c268ad5 1542specified via the =~ or !~ operator, the \$_ string is transliterated. (The
54310121 1543string specified with =~ must be a scalar variable, an array element, a
1544hash element, or an assignment to one of those, i.e., an lvalue.)
TP
1546A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
1547does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121 1548For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1549SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1550its own pair of quotes, which may or may not be bracketing quotes,
a0d0e21e 1552
cc255d5f
JH
1553Note that C<tr> does B<not> do regular expression character classes
1554such as C<\d> or C<[:lower:]>. The <tr> operator is not equivalent to
1555the tr(1) utility. If you want to map strings between lower/upper
1556cases, see L<perlfunc/lc> and L<perlfunc/uc>, and in general consider
1557using the C<s> operator if you need regular expressions.
1558
JH
1559Note also that the whole range idea is rather unportable between
1560character sets--and even within character sets they may cause results
1561you probably didn't expect. A sound principle is to use only ranges
1562that begin from and end at either alphabets of equal case (a-e, A-E),
1563or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1564character sets in full.
1565
a0d0e21e
LW
1566Options:
1567
1568 c Complement the SEARCHLIST.
1569 d Delete found but unreplaced characters.
1570 s Squash duplicate replaced characters.
1571
19799a22
GS
1572If the C</c> modifier is specified, the SEARCHLIST character set
1573is complemented. If the C</d> modifier is specified, any characters
1575(Note that this is slightly more flexible than the behavior of some
1576B<tr> programs, which delete anything they find in the SEARCHLIST,
1577period.) If the C</s> modifier is specified, sequences of characters
1578that were transliterated to the same character are squashed down
1579to a single instance of the character.
a0d0e21e
LW
1580
1581If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1582exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1583than the SEARCHLIST, the final character is replicated till it is long
5a964f20 1584enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
a0d0e21e
LW
1585This latter is useful for counting characters in a class or for
1586squashing character sequences in a class.
1587
1588Examples:
1589
1590 \$ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1591
1592 \$cnt = tr/*/*/; # count the stars in \$_
1593
1594 \$cnt = \$sky =~ tr/*/*/; # count the stars in \$sky
1595
1596 \$cnt = tr/0-9//; # count the digits in \$_
1597
1598 tr/a-zA-Z//s; # bookkeeper -> bokeper
1599
1600 (\$HOST = \$host) =~ tr/a-z/A-Z/;
1601
1602 tr/a-zA-Z/ /cs; # change non-alphas to single space
1603
1604 tr [\200-\377]
1605 [\000-\177]; # delete 8th bit
1606
19799a22
GS
1607If multiple transliterations are given for a character, only the
1608first one is used:
748a9306
LW
1609
1610 tr/AAA/XYZ/
1611
2c268ad5 1612will transliterate any A to X.
748a9306 1613
19799a22 1614Because the transliteration table is built at compile time, neither
a0d0e21e 1615the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
19799a22
GS
1616interpolation. That means that if you want to use variables, you
1617must use an eval():
a0d0e21e
LW
1618
1619 eval "tr/\$oldlist/\$newlist/";
1620 die \$@ if \$@;
1621
1622 eval "tr/\$oldlist/\$newlist/, 1" or die \$@;
1623
7e3b091d 1624=item <<EOF
d74e8afc 1625X<here-doc> X<heredoc> X<here-document> X<<< << >>>
7e3b091d
DA
1626
1627A line-oriented form of quoting is based on the shell "here-document"
1628syntax. Following a C<< << >> you specify a string to terminate
1629the quoted material, and all lines following the current line down to
1630the terminating string are the value of the item. The terminating
1631string may be either an identifier (a word), or some quoted text. If
1632quoted, the type of quotes you use determines the treatment of the
1633text, just as in regular quoting. An unquoted identifier works like
1634double quotes. There must be no space between the C<< << >> and
1635the identifier, unless the identifier is quoted. (If you put a space it
1636will be treated as a null identifier, which is valid, and matches the first
1637empty line.) The terminating string must appear by itself (unquoted and
1638with no surrounding whitespace) on the terminating line.
1639
1640 print <<EOF;
1641 The price is \$Price.
1642 EOF
1643
1644 print << "EOF"; # same as above
1645 The price is \$Price.
1646 EOF
1647
1648 print << `EOC`; # execute commands
1649 echo hi there
1650 echo lo there
1651 EOC
1652
1653 print <<"foo", <<"bar"; # you can stack them
1654 I said foo.
1655 foo
1656 I said bar.
1657 bar
1658
1659 myfunc(<< "THIS", 23, <<'THAT');
1660 Here's a line
1661 or two.
1662 THIS
1663 and here's another.
1664 THAT
1665
1666Just don't forget that you have to put a semicolon on the end
1667to finish the statement, as Perl doesn't know you're not going to
1668try to do this:
1669
1670 print <<ABC
1671 179231
1672 ABC
1673 + 20;
1674
1675If you want your here-docs to be indented with the
1676rest of the code, you'll need to remove leading whitespace
1677from each line manually:
1678
1679 (\$quote = <<'FINIS') =~ s/^\s+//gm;
1680 The Road goes ever on and on,
1681 down from the door where it began.
1682 FINIS
1683
1684If you use a here-doc within a delimited construct, such as in C<s///eg>,
1685the quoted material must come on the lines following the final delimiter.
1687
1688 s/this/<<E . 'that'
1689 the other
1690 E
1691 . 'more '/eg;
1692
1693you have to write
1694
1695 s/this/<<E . 'that'
1696 . 'more '/eg;
1697 the other
1698 E
1699
1700If the terminating identifier is on the last line of the program, you
1701must be sure there is a newline after it; otherwise, Perl will give the
1702warning B<Can't find string terminator "END" anywhere before EOF...>.
1703
1704Additionally, the quoting rules for the identifier are not related to
1705Perl's quoting rules -- C<q()>, C<qq()>, and the like are not supported
1706in place of C<''> and C<"">, and the only interpolation is for backslashing
1707the quoting character:
1708
1709 print << "abc\"def";
1710 testing...
1711 abc"def
1712
1713Finally, quoted strings cannot span multiple lines. The general rule is
1714that the identifier must be a string literal. Stick with that, and you
1715should be safe.
1716
a0d0e21e
LW
1717=back
1718
75e14d17 1719=head2 Gory details of parsing quoted constructs
d74e8afc 1720X<quote, gory details>
75e14d17 1721
19799a22
GS
1722When presented with something that might have several different
1723interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
1724principle to pick the most probable interpretation. This strategy
1725is so successful that Perl programmers often do not suspect the
1726ambivalence of what they write. But from time to time, Perl's
1727notions differ substantially from what the author honestly meant.
1728
1729This section hopes to clarify how Perl handles quoted constructs.
1730Although the most common reason to learn this is to unravel labyrinthine
1731regular expressions, because the initial steps of parsing are the
1732same for all quoting operators, they are all discussed together.
1733
1734The most important Perl parsing rule is the first one discussed
1735below: when processing a quoted construct, Perl first finds the end
1736of that construct, then interprets its contents. If you understand
1737this rule, you may skip the rest of this section on the first
1739expectations much less frequently than this first one.
1740
1741Some passes discussed below are performed concurrently, but because
1742their results are the same, we consider them individually. For different
1743quoting constructs, Perl performs different numbers of passes, from
1744one to five, but these passes are always performed in the same order.
75e14d17 1745
13a2d996 1746=over 4
75e14d17
IZ
1747
1748=item Finding the end
1749
19799a22
GS
1750The first pass is finding the end of the quoted construct, whether
1751it be a multicharacter delimiter C<"\nEOF\n"> in the C<<<EOF>
1752construct, a C</> that terminates a C<qq//> construct, a C<]> which
35f2feb0
GS
1753terminates C<qq[]> construct, or a C<< > >> which terminates a
1754fileglob started with C<< < >>.
75e14d17 1755
19799a22
GS
1756When searching for single-character non-pairing delimiters, such
1757as C</>, combinations of C<\\> and C<\/> are skipped. However,
1758when searching for single-character pairing delimiter like C<[>,
1759combinations of C<\\>, C<\]>, and C<\[> are all skipped, and nested
1760C<[>, C<]> are skipped as well. When searching for multicharacter
1761delimiters, nothing is skipped.
75e14d17 1762
19799a22
GS
1763For constructs with three-part delimiters (C<s///>, C<y///>, and
1764C<tr///>), the search is repeated once more.
75e14d17 1765
19799a22
GS
1766During this search no attention is paid to the semantics of the construct.
1767Thus:
75e14d17
IZ
1768
1769 "\$hash{"\$foo/\$bar"}"
1770
2a94b7ce 1771or:
75e14d17
IZ
1772
1773 m/
2a94b7ce 1774 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
1775 /x
1776
19799a22
GS
1777do not form legal quoted expressions. The quoted part ends on the
1778first C<"> and C</>, and the rest happens to be a syntax error.
1779Because the slash that terminated C<m//> was followed by a C<SPACE>,
1780the example above is not C<m//x>, but rather C<m//> with no C</x>
1781modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 1782
0d594e51
TS
1783Also no attention is paid to C<\c\> during this search.
1784Thus the second C<\> in C<qq/\c\/> is interpreted as a part of C<\/>,
1785and the following C</> is not recognized as a delimiter.
1786Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
1787
75e14d17
IZ
1788=item Removal of backslashes before delimiters
1789
19799a22
GS
1790During the second pass, text between the starting and ending
1791delimiters is copied to a safe location, and the C<\> is removed
1792from combinations consisting of C<\> and delimiter--or delimiters,
1793meaning both starting and ending delimiters will should these differ.
1794This removal does not happen for multi-character delimiters.
1795Note that the combination C<\\> is left intact, just as it was.
75e14d17 1796
19799a22
GS
1797Starting from this step no information about the delimiters is
1798used in parsing.
75e14d17
IZ
1799
1800=item Interpolation
d74e8afc 1801X<interpolation>
75e14d17 1802
19799a22
GS
1803The next step is interpolation in the text obtained, which is now
1804delimiter-independent. There are four different cases.
75e14d17 1805
13a2d996 1806=over 4
75e14d17
IZ
1807
1808=item C<<<'EOF'>, C<m''>, C<s'''>, C<tr///>, C<y///>
1809
1810No interpolation is performed.
1811
1812=item C<''>, C<q//>
1813
1814The only interpolation is removal of C<\> from pairs C<\\>.
1815
35f2feb0 1816=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>
75e14d17 1817
19799a22
GS
1818C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
1819converted to corresponding Perl constructs. Thus, C<"\$foo\Qbaz\$bar">
1820is converted to C<\$foo . (quotemeta("baz" . \$bar))> internally.
1821The other combinations are replaced with appropriate expansions.
2a94b7ce 1822
19799a22
GS
1823Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
1824is interpolated in the usual way. Something like C<"\Q\\E"> has
1825no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
1826result is the same as for C<"\\\\E">. As a general rule, backslashes
1827between C<\Q> and C<\E> may lead to counterintuitive results. So,
1828C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
1829as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
1830
1831 \$str = '\t';
1832 return "\Q\$str";
1833
1834may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
1835
19799a22 1836Interpolated scalars and arrays are converted internally to the C<join> and
92d29cee 1837C<.> catenation operations. Thus, C<"\$foo XXX '@arr'"> becomes:
75e14d17 1838
19799a22 1839 \$foo . " XXX '" . (join \$", @arr) . "'";
75e14d17 1840
19799a22 1841All operations above are performed simultaneously, left to right.
75e14d17 1842
19799a22
GS
1843Because the result of C<"\Q STRING \E"> has all metacharacters
1844quoted, there is no way to insert a literal C<\$> or C<@> inside a
1845C<\Q\E> pair. If protected by C<\>, C<\$> will be quoted to became
1846C<"\\\\$">; if not, it is interpreted as the start of an interpolated
1847scalar.
75e14d17 1848
19799a22
GS
1849Note also that the interpolation code needs to make a decision on
1850where the interpolated scalar ends. For instance, whether
35f2feb0 1851C<< "a \$b -> {c}" >> really means:
75e14d17
IZ
1852
1853 "a " . \$b . " -> {c}";
1854
2a94b7ce 1855or:
75e14d17
IZ
1856
1857 "a " . \$b -> {c};
1858
19799a22
GS
1859Most of the time, the longest possible text that does not include
1860spaces between components and which contains matching braces or
1861brackets. because the outcome may be determined by voting based
1862on heuristic estimators, the result is not strictly predictable.
1863Fortunately, it's usually correct for ambiguous cases.
75e14d17
IZ
1864
1865=item C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
1866
19799a22
GS
1867Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
1868happens (almost) as with C<qq//> constructs, but the substitution
1869of C<\> followed by RE-special chars (including C<\>) is not
1870performed. Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
1871a C<#>-comment in a C<//x>-regular expression, no processing is
1872performed whatsoever. This is the first step at which the presence
1873of the C<//x> modifier is relevant.
1874
1875Interpolation has several quirks: C<\$|>, C<\$(>, and C<\$)> are not
1876interpolated, and constructs C<\$var[SOMETHING]> are voted (by several
1877different estimators) to be either an array element or C<\$var>
1878followed by an RE alternative. This is where the notation
1879C<\${arr[\$bar]}> comes handy: C</\${arr[0-9]}/> is interpreted as
1880array element C<-9>, not as a regular expression from the variable
1881C<\$arr> followed by a digit, which would be the interpretation of
1882C</\$arr[0-9]/>. Since voting among different estimators may occur,
1883the result is not predictable.
1884
1885It is at this step that C<\1> is begrudgingly converted to C<\$1> in
1886the replacement text of C<s///> to correct the incorrigible
1887I<sed> hackers who haven't picked up the saner idiom yet. A warning
9f1b1f2d
GS
1888is emitted if the C<use warnings> pragma or the B<-w> command-line flag
1889(that is, the C<\$^W> variable) was set.
19799a22
GS
1890
1891The lack of processing of C<\\> creates specific restrictions on
1892the post-processed text. If the delimiter is C</>, one cannot get
1893the combination C<\/> into the result of this step. C</> will
1894finish the regular expression, C<\/> will be stripped to C</> on
1895the previous step, and C<\\/> will be left as is. Because C</> is
1896equivalent to C<\/> inside a regular expression, this does not
1897matter unless the delimiter happens to be character special to the
1898RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
1899alphanumeric char, as in:
2a94b7ce
IZ
1900
1901 m m ^ a \s* b mmx;
1902
19799a22 1903In the RE above, which is intentionally obfuscated for illustration, the
2a94b7ce 1904delimiter is C<m>, the modifier is C<mx>, and after backslash-removal the
aa863641 1905RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
19799a22
GS
1906reason you're encouraged to restrict your delimiters to non-alphanumeric,
1907non-whitespace choices.
75e14d17
IZ
1908
1909=back
1910
19799a22 1911This step is the last one for all constructs except regular expressions,
75e14d17
IZ
1912which are processed further.
1913
1914=item Interpolation of regular expressions
d74e8afc 1915X<regexp, interpolation>
75e14d17 1916
19799a22
GS
1917Previous steps were performed during the compilation of Perl code,
1918but this one happens at run time--although it may be optimized to
1919be calculated at compile time if appropriate. After preprocessing
1920described above, and possibly after evaluation if catenation,
1921joining, casing translation, or metaquoting are involved, the
1922resulting I<string> is passed to the RE engine for compilation.
1923
1924Whatever happens in the RE engine might be better discussed in L<perlre>,
1925but for the sake of continuity, we shall do so here.
1926
1927This is another step where the presence of the C<//x> modifier is
1928relevant. The RE engine scans the string from left to right and
1929converts it to a finite automaton.
1930
1931Backslashed characters are either replaced with corresponding
1932literal strings (as with C<\{>), or else they generate special nodes
1933in the finite automaton (as with C<\b>). Characters special to the
1934RE engine (such as C<|>) generate corresponding nodes or groups of
1935nodes. C<(?#...)> comments are ignored. All the rest is either
1936converted to literal strings to match, or else is ignored (as is
1937whitespace and C<#>-style comments if C<//x> is present).
1938
1939Parsing of the bracketed character class construct, C<[...]>, is
1940rather different than the rule used for the rest of the pattern.
1941The terminator of this construct is found using the same rules as
1942for finding the terminator of a C<{}>-delimited construct, the only
1943exception being that C<]> immediately following C<[> is treated as
1944though preceded by a backslash. Similarly, the terminator of
1945C<(?{...})> is found using the same rules as for finding the
1946terminator of a C<{}>-delimited construct.
1947
1948It is possible to inspect both the string given to RE engine and the
1949resulting finite automaton. See the arguments C<debug>/C<debugcolor>
1950in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 1951switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
1952
1953=item Optimization of regular expressions
d74e8afc 1954X<regexp, optimization>
75e14d17 1955
7522fed5 1956This step is listed for completeness only. Since it does not change
75e14d17 1957semantics, details of this step are not documented and are subject
19799a22
GS
1958to change without notice. This step is performed over the finite
1959automaton that was generated during the previous pass.
2a94b7ce 1960
19799a22
GS
1961It is at this stage that C<split()> silently optimizes C</^/> to
1962mean C</^/m>.
75e14d17
IZ
1963
1964=back
1965
d74e8afc
ITB
1967X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
1968X<< <> >> X<@ARGV>
a0d0e21e 1969
54310121 1970There are several I/O operators you should know about.
7b8d334a 1972A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
1973double-quote interpolation. It is then interpreted as an external
1974command, and the output of that command is the value of the
e9c56f9b
JH
1975backtick string, like in a shell. In scalar context, a single string
1976consisting of all output is returned. In list context, a list of
1977values is returned, one per line of output. (You can set C<\$/> to use
1978a different line terminator.) The command is executed each time the
1979pseudo-literal is evaluated. The status value of the command is
1980returned in C<\$?> (see L<perlvar> for the interpretation of C<\$?>).
1981Unlike in B<csh>, no translation is done on the return data--newlines
1982remain newlines. Unlike in any of the shells, single quotes do not
1983hide variable names in the command from interpretation. To pass a
1984literal dollar-sign through to the shell you need to hide it with a
1985backslash. The generalized form of backticks is C<qx//>. (Because
1986backticks always undergo shell expansion as well, see L<perlsec> for
1987security concerns.)
d74e8afc 1988X<qx> X<`> X<``> X<backtick> X<glob>
19799a22
GS
1989
1990In scalar context, evaluating a filehandle in angle brackets yields
1991the next line from that file (the newline, if any, included), or
1992C<undef> at end-of-file or on error. When C<\$/> is set to C<undef>
1993(sometimes known as file-slurp mode) and the file is empty, it
1994returns C<''> the first time, followed by C<undef> subsequently.
1995
1996Ordinarily you must assign the returned value to a variable, but
1997there is one situation where an automatic assignment happens. If
1998and only if the input symbol is the only thing inside the conditional
1999of a C<while> statement (even if disguised as a C<for(;;)> loop),
2000the value is automatically assigned to the global variable \$_,
2001destroying whatever was there previously. (This may seem like an
2002odd thing to you, but you'll use the construct in almost every Perl
17b829fa 2003script you write.) The \$_ variable is not implicitly localized.
19799a22
GS
2004You'll have to put a C<local \$_;> before the loop if you want that
2005to happen.
2006
2007The following lines are equivalent:
a0d0e21e 2008
748a9306 2009 while (defined(\$_ = <STDIN>)) { print; }
7b8d334a 2010 while (\$_ = <STDIN>) { print; }
a0d0e21e
LW
2011 while (<STDIN>) { print; }
2012 for (;<STDIN>;) { print; }
748a9306 2013 print while defined(\$_ = <STDIN>);
7b8d334a 2014 print while (\$_ = <STDIN>);
a0d0e21e
LW
2015 print while <STDIN>;
2016
19799a22 2017This also behaves similarly, but avoids \$_ :
7b8d334a
GS
2018
2019 while (my \$line = <STDIN>) { print \$line }
2020
19799a22
GS
2021In these loop constructs, the assigned value (whether assignment
2022is automatic or explicit) is then tested to see whether it is
2023defined. The defined test avoids problems where line has a string
2024value that would be treated as false by Perl, for example a "" or
2025a "0" with no trailing newline. If you really mean for such values
2026to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
2027
2028 while ((\$_ = <STDIN>) ne '0') { ... }
2029 while (<STDIN>) { last unless \$_; ... }
2030
35f2feb0 2031In other boolean contexts, C<< <I<filehandle>> >> without an
9f1b1f2d
GS
2032explicit C<defined> test or comparison elicit a warning if the
2033C<use warnings> pragma or the B<-w>
19799a22 2034command-line switch (the C<\$^W> variable) is in effect.
7b8d334a 2035
5f05dabc 2036The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
2037filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
2038in packages, where they would be interpreted as local identifiers
2039rather than global.) Additional filehandles may be created with
2040the open() function, amongst others. See L<perlopentut> and
2041L<perlfunc/open> for details on this.
d74e8afc 2042X<stdin> X<stdout> X<sterr>
a0d0e21e 2043
35f2feb0 2044If a <FILEHANDLE> is used in a context that is looking for
19799a22
GS
2045a list, a list comprising all input lines is returned, one line per
2046list element. It's easy to grow to a rather large data space this
2047way, so use with care.
a0d0e21e 2048
35f2feb0 2049<FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
35f2feb0
GS
2052The null filehandle <> is special: it can be used to emulate the
2053behavior of B<sed> and B<awk>. Input from <> comes either from
a0d0e21e 2054standard input, or from each file listed on the command line. Here's
35f2feb0 2055how it works: the first time <> is evaluated, the @ARGV array is
5a964f20 2056checked, and if it is empty, C<\$ARGV[0]> is set to "-", which when opened
a0d0e21e
LW
2057gives you standard input. The @ARGV array is then processed as a list
2058of filenames. The loop
2059
2060 while (<>) {
2061 ... # code for each line
2062 }
2063
2064is equivalent to the following Perl-like pseudo code:
2065
3e3baf6d 2066 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
2067 while (\$ARGV = shift) {
2068 open(ARGV, \$ARGV);
2069 while (<ARGV>) {
2070 ... # code for each line
2071 }
2072 }
2073
19799a22
GS
2074except that it isn't so cumbersome to say, and will actually work.
2075It really does shift the @ARGV array and put the current filename
2076into the \$ARGV variable. It also uses filehandle I<ARGV>
35f2feb0 2077internally--<> is just a synonym for <ARGV>, which
19799a22 2078is magical. (The pseudo code above doesn't work because it treats
35f2feb0 2079<ARGV> as non-magical.)
a0d0e21e 2080
35f2feb0 2081You can modify @ARGV before the first <> as long as the array ends up
a0d0e21e 2082containing the list of filenames you really want. Line numbers (C<\$.>)
19799a22
GS
2083continue as though the input were one big happy file. See the example
2084in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20
TC
2085
2086If you want to set @ARGV to your own list of files, go right ahead.
2087This sets @ARGV to all plain text files if no @ARGV was given:
2088
2089 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 2090
5a964f20
TC
2091You can even set them to pipe commands. For example, this automatically
2092filters compressed arguments through B<gzip>:
2093
2094 @ARGV = map { /\.(gz|Z)\$/ ? "gzip -dc < \$_ |" : \$_ } @ARGV;
2095
2096If you want to pass switches into your script, you can use one of the
a0d0e21e
LW
2097Getopts modules or put a loop on the front like this:
2098
2099 while (\$_ = \$ARGV[0], /^-/) {
2100 shift;
2101 last if /^--\$/;
2102 if (/^-D(.*)/) { \$debug = \$1 }
2103 if (/^-v/) { \$verbose++ }
5a964f20 2104 # ... # other switches
a0d0e21e 2105 }
5a964f20 2106
a0d0e21e 2107 while (<>) {
5a964f20 2108 # ... # code for each line
a0d0e21e
LW
2109 }
2110
35f2feb0 2111The <> symbol will return C<undef> for end-of-file only once.
19799a22
GS
2112If you call it again after this, it will assume you are processing another
2113@ARGV list, and if you haven't set @ARGV, will read input from STDIN.
a0d0e21e 2114
b159ebd3 2115If what the angle brackets contain is a simple scalar variable (e.g.,
35f2feb0 2116<\$foo>), then that variable contains the name of the
19799a22
GS
2117filehandle to input from, or its typeglob, or a reference to the
2118same. For example:
cb1a09d0
2119
2120 \$fh = \*STDIN;
2121 \$line = <\$fh>;
a0d0e21e 2122
5a964f20
TC
2123If what's within the angle brackets is neither a filehandle nor a simple
2124scalar variable containing a filehandle name, typeglob, or typeglob
2125reference, it is interpreted as a filename pattern to be globbed, and
2126either a list of filenames or the next filename in the list is returned,
19799a22 2127depending on context. This distinction is determined on syntactic
35f2feb0
GS
2128grounds alone. That means C<< <\$x> >> is always a readline() from
2129an indirect handle, but C<< <\$hash{key}> >> is always a glob().
5a964f20 2130That's because \$x is a simple scalar variable, but C<\$hash{key}> is
ef191992
YST
2131not--it's a hash element. Even C<< <\$x > >> (note the extra space)
2132is treated as C<glob("\$x ")>, not C<readline(\$x)>.
5a964f20
TC
2133
2134One level of double-quote interpretation is done first, but you can't
35f2feb0 2135say C<< <\$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
2136in the previous paragraph. (In older versions of Perl, programmers
2137would insert curly brackets to force interpretation as a filename glob:
35f2feb0 2138C<< <\${foo}> >>. These days, it's considered cleaner to call the
5a964f20 2139internal function directly as C<glob(\$foo)>, which is probably the right
19799a22 2140way to have done it in the first place.) For example:
a0d0e21e
LW
2141
2142 while (<*.c>) {
2143 chmod 0644, \$_;
2144 }
2145
3a4b19e4 2146is roughly equivalent to:
a0d0e21e
LW
2147
2148 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
2149 while (<FOO>) {
5b3eff12 2150 chomp;
a0d0e21e
LW
2151 chmod 0644, \$_;
2152 }
2153
3a4b19e4
GS
2154except that the globbing is actually done internally using the standard
2155C<File::Glob> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
2156
2157 chmod 0644, <*.c>;
2158
19799a22
GS
2159A (file)glob evaluates its (embedded) argument only when it is
2160starting a new list. All values must be read before it will start
2161over. In list context, this isn't important because you automatically
2162get them all anyway. However, in scalar context the operator returns
069e01df 2163the next value each time it's called, or C<undef> when the list has
19799a22
GS
2164run out. As with filehandle reads, an automatic C<defined> is
2165generated when the glob occurs in the test part of a C<while>,
2166because legal glob returns (e.g. a file called F<0>) would otherwise
2167terminate the loop. Again, C<undef> is returned only once. So if
2168you're expecting a single value from a glob, it is much better to
2169say
4633a7c4
LW
2170
2171 (\$file) = <blurch*>;
2172
2173than
2174
2175 \$file = <blurch*>;
2176
2177because the latter will alternate between returning a filename and
19799a22 2178returning false.
4633a7c4 2179
b159ebd3 2180If you're trying to do variable interpolation, it's definitely better
4633a7c4 2181to use the glob() function, because the older notation can cause people
e37d713d 2182to become confused with the indirect filehandle notation.
4633a7c4
LW
2183
2184 @files = glob("\$dir/*.[ch]");
2185 @files = glob(\$files[\$i]);
2186
d74e8afc 2188X<constant folding> X<folding>
a0d0e21e
LW
2189
2190Like C, Perl does a certain amount of expression evaluation at
19799a22 2191compile time whenever it determines that all arguments to an
a0d0e21e
LW
2192operator are static and have no side effects. In particular, string
2193concatenation happens at compile time between literals that don't do
19799a22 2194variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
2195compile time. You can say
2196
2197 'Now is the time for all' . "\n" .
2198 'good men to come to.'
2199
54310121 2200and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
2201you say
2202
2203 foreach \$file (@filenames) {
5a964f20 2204 if (-s \$file > 5 + 100 * 2**16) { }
54310121 2205 }
a0d0e21e 2206
19799a22
GS
2207the compiler will precompute the number which that expression
2208represents so that the interpreter won't have to.
a0d0e21e 2209
d74e8afc 2211X<no-op> X<nop>
fd1abbef
DN
2212
2213Perl doesn't officially have a no-op operator, but the bare constants
2214C<0> and C<1> are special-cased to not produce a warning in a void
2215context, so you can for example safely do
2216
2217 1 while foo();
2218
d74e8afc 2220X<operator, bitwise, string>
TP
2221
2222Bitstrings of any size may be manipulated by the bitwise operators
2223(C<~ | & ^>).
2224
19799a22
GS
2225If the operands to a binary bitwise op are strings of different
2226sizes, B<|> and B<^> ops act as though the shorter operand had
2227additional zero bits on the right, while the B<&> op acts as though
2228the longer operand were truncated to the length of the shorter.
2229The granularity for such extension or truncation is one or more
2230bytes.
TP
2231
2232 # ASCII-based examples
2233 print "j p \n" ^ " a h"; # prints "JAPH\n"
2234 print "JA" | " ph\n"; # prints "japh\n"
2235 print "japh\nJunk" & '_____'; # prints "JAPH\n";
2236 print 'p N\$' ^ " E<H\n"; # prints "Perl\n";
2237
19799a22 2238If you are intending to manipulate bitstrings, be certain that
2c268ad5 2239you're supplying bitstrings: If an operand is a number, that will imply
19799a22 2240a B<numeric> bitwise operation. You may explicitly show which type of
TP
2241operation you intend by using C<""> or C<0+>, as in the examples below.
2242
4358a253
SS
2243 \$foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
2244 \$foo = '150' | 105; # yields 255
TP
2245 \$foo = 150 | '105'; # yields 255
2246 \$foo = '150' | '105'; # yields string '155' (under ASCII)
2247
2248 \$baz = 0+\$foo & 0+\$bar; # both ops explicitly numeric
2249 \$biz = "\$foo" ^ "\$bar"; # both ops explicitly stringy
a0d0e21e 2250
1ae175c8
GS
2251See L<perlfunc/vec> for information on how to manipulate individual bits
2252in a bit vector.
2253
d74e8afc 2255X<integer>
a0d0e21e 2256
19799a22 2257By default, Perl assumes that it must do most of its arithmetic in
a0d0e21e
LW
2258floating point. But by saying
2259
2260 use integer;
2261
2262you may tell the compiler that it's okay to use integer operations
19799a22
GS
2263(if it feels like it) from here to the end of the enclosing BLOCK.
2264An inner BLOCK may countermand this by saying
a0d0e21e
LW
2265
2266 no integer;
2267
19799a22
GS
2268which lasts until the end of that BLOCK. Note that this doesn't
2269mean everything is only an integer, merely that Perl may use integer
2270operations if it is so inclined. For example, even under C<use
2271integer>, if you take the C<sqrt(2)>, you'll still get C<1.4142135623731>
2272or so.
2273
2274Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
13a2d996
SP
2276L<Bitwise String Operators>.) However, C<use integer> still has meaning for
19799a22
GS
2277them. By default, their results are interpreted as unsigned integers, but
2278if C<use integer> is in effect, their results are interpreted
2279as signed integers. For example, C<~0> usually evaluates to a large
0be96356 2280integral value. However, C<use integer; ~0> is C<-1> on two's-complement
19799a22 2281machines.
68dc0745 2282
d74e8afc 2284X<floating-point> X<floating point> X<float> X<real>
68dc0745 2285
2286While C<use integer> provides integer-only arithmetic, there is no
19799a22
GS
2287analogous mechanism to provide automatic rounding or truncation to a
2288certain number of decimal places. For rounding to a certain number
2289of digits, sprintf() or printf() is usually the easiest route.
2290See L<perlfaq4>.
68dc0745 2291
5a964f20
TC
2292Floating-point numbers are only approximations to what a mathematician
2293would call real numbers. There are infinitely more reals than floats,
2294so some corners must be cut. For example:
2295
2296 printf "%.20g\n", 123456789123456789;
2297 # produces 123456789123456784
2298
2299Testing for exact equality of floating-point equality or inequality is
2300not a good idea. Here's a (relatively expensive) work-around to compare
2301whether two floating-point numbers are equal to a particular number of
2302decimal places. See Knuth, volume II, for a more robust treatment of
2303this topic.
2304
2305 sub fp_equal {
2306 my (\$X, \$Y, \$POINTS) = @_;
2307 my (\$tX, \$tY);
2308 \$tX = sprintf("%.\${POINTS}g", \$X);
2309 \$tY = sprintf("%.\${POINTS}g", \$Y);
2310 return \$tX eq \$tY;
2311 }
2312
68dc0745 2313The POSIX module (part of the standard perl distribution) implements
19799a22
GS
2314ceil(), floor(), and other mathematical and trigonometric functions.
2315The Math::Complex module (part of the standard perl distribution)
2316defines mathematical functions that work on both the reals and the
2317imaginary numbers. Math::Complex not as efficient as POSIX, but
68dc0745 2318POSIX can't work with complex numbers.
2319
2320Rounding in financial applications can have serious implications, and
2321the rounding method used should be specified precisely. In these
2322cases, it probably pays not to trust whichever system rounding is
2323being used by Perl, but to instead implement the rounding function you
2324need yourself.
5a964f20
TC
2325
d74e8afc 2327X<number, arbitrary precision>
5a964f20
TC
2328
2329The standard Math::BigInt and Math::BigFloat modules provide
19799a22 2330variable-precision arithmetic and overloaded operators, although
cd5c4fce 2331they're currently pretty slow. At the cost of some space and
19799a22
GS
2332considerable speed, they avoid the normal pitfalls associated with
2333limited-precision representations.
5a964f20
TC
2334
2335 use Math::BigInt;
2336 \$x = Math::BigInt->new('123456789123456789');
2337 print \$x * \$x;
2338
2339 # prints +15241578780673678515622620750190521
19799a22 2340
cd5c4fce
T
2341There are several modules that let you calculate with (bound only by
2342memory and cpu-time) unlimited or fixed precision. There are also
2343some non-standard modules that provide faster implementations via
2344external C libraries.
2345
2346Here is a short, but incomplete summary:
2347
2348 Math::Fraction big, unlimited fractions like 9973 / 12967
2349 Math::String treat string sequences like numbers
2350 Math::FixedPrecision calculate with a fixed precision
2351 Math::Currency for currency calculations
2352 Bit::Vector manipulate bit vectors fast (uses C)
2353 Math::BigIntFast Bit::Vector wrapper for big numbers