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
89d205f2 8=head2 Operator Precedence and Associativity
d74e8afc 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
60Many operators can be overloaded for objects. See L<overload>.
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
124=head2 The Arrow Operator
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
5f05dabc 140=head2 Auto-increment and Auto-decrement
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 152Note that just as in C, Perl doesn't define B<when> the variable is
89d205f2 153incremented or decremented. You just know it will be done sometime
b033823e
A
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
190=head2 Symbolic Unary Operators
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
210example, C<0666 & ~027> is 0640. (See also L<Integer Arithmetic> and
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
89d205f2 239operator. See L</"Regexp Quote-Like Operators"> for details and
d7782e69 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
89d205f2
YO
244time. Note that this means that its contents will be interpolated twice, so
245
246 '\\' =~ q'\\';
247
248is not ok, as the regex engine will end up trying to compile the
249pattern C<\>, which it will consider a syntax error.
a0d0e21e
LW
250
251Binary "!~" is just like "=~" except the return value is negated in
252the logical sense.
253
d74e8afc 255X<operator, multiplicative>
a0d0e21e
LW
256
257Binary "*" multiplies two numbers.
d74e8afc 258X<*>
a0d0e21e
LW
259
260Binary "/" divides two numbers.
d74e8afc 261X</> X<slash>
a0d0e21e 262
54310121 263Binary "%" computes the modulus of two numbers. Given integer
264operands C<\$a> and C<\$b>: If C<\$b> is positive, then C<\$a % \$b> is
265C<\$a> minus the largest multiple of C<\$b> that is not greater than
266C<\$a>. If C<\$b> is negative, then C<\$a % \$b> is C<\$a> minus the
267smallest multiple of C<\$b> that is not less than C<\$a> (i.e. the
89b4f0ad 268result will be less than or equal to zero). If the operands
4848a83b
TS
269C<\$a> and C<\$b> are floating point values and the absolute value of
270C<\$b> (that is C<abs(\$b)>) is less than C<(UV_MAX + 1)>, only
271the integer portion of C<\$a> and C<\$b> will be used in the operation
272(Note: here C<UV_MAX> means the maximum of the unsigned integer type).
273If the absolute value of the right operand (C<abs(\$b)>) is greater than
274or equal to C<(UV_MAX + 1)>, "%" computes the floating-point remainder
275C<\$r> in the equation C<(\$r = \$a - \$i*\$b)> where C<\$i> is a certain
276integer that makes C<\$r> should have the same sign as the right operand
277C<\$b> (B<not> as the left operand C<\$a> like C function C<fmod()>)
278and the absolute value less than that of C<\$b>.
0412d526 279Note that when C<use integer> is in scope, "%" gives you direct access
55d729e4
GS
280to the modulus operator as implemented by your C compiler. This
281operator is not as well defined for negative operands, but it will
282execute faster.
d74e8afc 283X<%> X<remainder> X<modulus> X<mod>
55d729e4 284
62d10b70
GS
285Binary "x" is the repetition operator. In scalar context or if the left
286operand is not enclosed in parentheses, it returns a string consisting
287of the left operand repeated the number of times specified by the right
288operand. In list context, if the left operand is enclosed in
3585017f
YST
289parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
290If the right operand is zero or negative, it returns an empty string
291or an empty list, depending on the context.
d74e8afc 292X<x>
a0d0e21e
LW
293
294 print '-' x 80; # print row of dashes
295
296 print "\t" x (\$tab/8), ' ' x (\$tab%8); # tab over
297
298 @ones = (1) x 80; # a list of 80 1's
299 @ones = (5) x @ones; # set all elements to 5
300
301
a0d0e21e
LW
304
305Binary "+" returns the sum of two numbers.
d74e8afc 306X<+>
a0d0e21e
LW
307
308Binary "-" returns the difference of two numbers.
d74e8afc 309X<->
a0d0e21e
LW
310
311Binary "." concatenates two strings.
d74e8afc
ITB
312X<string, concatenation> X<concatenation>
313X<cat> X<concat> X<concatenate> X<.>
a0d0e21e
LW
314
d74e8afc
ITB
316X<shift operator> X<operator, shift> X<<< << >>>
317X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
318X<shl> X<shr> X<shift, right> X<shift, left>
a0d0e21e 319
55497cff 320Binary "<<" returns the value of its left argument shifted left by the
321number of bits specified by the right argument. Arguments should be
a0d0e21e 323
55497cff 324Binary ">>" returns the value of its left argument shifted right by
325the number of bits specified by the right argument. Arguments should
982ce180 326be integers. (See also L<Integer Arithmetic>.)
a0d0e21e 327
b16cf6df
JH
328Note that both "<<" and ">>" in Perl are implemented directly using
329"<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
330in force then signed C integers are used, else unsigned C integers are
331used. Either way, the implementation isn't going to generate results
332larger than the size of the integer type Perl was built with (32 bits
333or 64 bits).
334
335The result of overflowing the range of the integers is undefined
336because it is undefined also in C. In other words, using 32-bit
337integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
338of bits is also undefined.
339
a0d0e21e 340=head2 Named Unary Operators
d74e8afc 341X<operator, named unary>
a0d0e21e
LW
342
343The various named unary operators are treated as functions with one
568e6d8b 344argument, with optional parentheses.
a0d0e21e
LW
345
346If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
347is followed by a left parenthesis as the next token, the operator and
348arguments within parentheses are taken to be of highest precedence,
3981b0eb
JA
349just like a normal function call. For example,
350because named unary operators are higher precedence than ||:
a0d0e21e
LW
351
352 chdir \$foo || die; # (chdir \$foo) || die
353 chdir(\$foo) || die; # (chdir \$foo) || die
354 chdir (\$foo) || die; # (chdir \$foo) || die
355 chdir +(\$foo) || die; # (chdir \$foo) || die
356
3981b0eb 357but, because * is higher precedence than named operators:
a0d0e21e
LW
358
359 chdir \$foo * 20; # chdir (\$foo * 20)
360 chdir(\$foo) * 20; # (chdir \$foo) * 20
361 chdir (\$foo) * 20; # (chdir \$foo) * 20
362 chdir +(\$foo) * 20; # chdir (\$foo * 20)
363
364 rand 10 * 20; # rand (10 * 20)
365 rand(10) * 20; # (rand 10) * 20
366 rand (10) * 20; # (rand 10) * 20
367 rand +(10) * 20; # rand (10 * 20)
368
568e6d8b
RGS
369Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
370treated like named unary operators, but they don't follow this functional
371parenthesis rule. That means, for example, that C<-f(\$file).".bak"> is
372equivalent to C<-f "\$file.bak">.
d74e8afc 373X<-X> X<filetest> X<operator, filetest>
568e6d8b 374
5ba421f6 375See also L<"Terms and List Operators (Leftward)">.
a0d0e21e
LW
376
d74e8afc 378X<relational operator> X<operator, relational>
a0d0e21e 379
35f2feb0 380Binary "<" returns true if the left argument is numerically less than
a0d0e21e 381the right argument.
d74e8afc 382X<< < >>
a0d0e21e 383
35f2feb0 384Binary ">" returns true if the left argument is numerically greater
a0d0e21e 385than the right argument.
d74e8afc 386X<< > >>
a0d0e21e 387
35f2feb0 388Binary "<=" returns true if the left argument is numerically less than
a0d0e21e 389or equal to the right argument.
d74e8afc 390X<< <= >>
a0d0e21e 391
35f2feb0 392Binary ">=" returns true if the left argument is numerically greater
a0d0e21e 393than or equal to the right argument.
d74e8afc 394X<< >= >>
a0d0e21e
LW
395
396Binary "lt" returns true if the left argument is stringwise less than
397the right argument.
d74e8afc 398X<< lt >>
a0d0e21e
LW
399
400Binary "gt" returns true if the left argument is stringwise greater
401than the right argument.
d74e8afc 402X<< gt >>
a0d0e21e
LW
403
404Binary "le" returns true if the left argument is stringwise less than
405or equal to the right argument.
d74e8afc 406X<< le >>
a0d0e21e
LW
407
408Binary "ge" returns true if the left argument is stringwise greater
409than or equal to the right argument.
d74e8afc 410X<< ge >>
a0d0e21e
LW
411
d74e8afc 413X<equality> X<equal> X<equals> X<operator, equality>
a0d0e21e
LW
414
415Binary "==" returns true if the left argument is numerically equal to
416the right argument.
d74e8afc 417X<==>
a0d0e21e
LW
418
419Binary "!=" returns true if the left argument is numerically not equal
420to the right argument.
d74e8afc 421X<!=>
a0d0e21e 422
35f2feb0 423Binary "<=>" returns -1, 0, or 1 depending on whether the left
6ee5d4e7 424argument is numerically less than, equal to, or greater than the right
d4ad863d 425argument. If your platform supports NaNs (not-a-numbers) as numeric
7d3a9d88
NC
426values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
427"<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
428returns true, as does NaN != anything else. If your platform doesn't
429support NaNs then NaN is just a string with numeric value 0.
d74e8afc 430X<< <=> >> X<spaceship>
7d3a9d88 431
2b54f59f
YST
432 perl -le '\$a = "NaN"; print "No NaN support here" if \$a == \$a'
433 perl -le '\$a = "NaN"; print "NaN support here" if \$a != \$a'
a0d0e21e
LW
434
435Binary "eq" returns true if the left argument is stringwise equal to
436the right argument.
d74e8afc 437X<eq>
a0d0e21e
LW
438
439Binary "ne" returns true if the left argument is stringwise not equal
440to the right argument.
d74e8afc 441X<ne>
a0d0e21e 442
JH
443Binary "cmp" returns -1, 0, or 1 depending on whether the left
444argument is stringwise less than, equal to, or greater than the right
445argument.
d74e8afc 446X<cmp>
a0d0e21e 447
0d863452 448Binary "~~" does a smart match between its arguments. Smart matching
0f7107a0 449is described in L<perlsyn/"Smart matching in detail">.
0d863452
RH
450This operator is only available if you enable the "~~" feature:
452X<~~>
453
a034a98d
DD
454"lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
455by the current locale if C<use locale> is in effect. See L<perllocale>.
456
a0d0e21e 457=head2 Bitwise And
d74e8afc 458X<operator, bitwise, and> X<bitwise and> X<&>
a0d0e21e 459
2cdc098b 460Binary "&" returns its operands ANDed together bit by bit.
a0d0e21e 462
2cdc098b
MG
463Note that "&" has lower priority than relational operators, so for example
464the brackets are essential in a test like
465
466 print "Even\n" if (\$x & 1) == 0;
467
a0d0e21e 468=head2 Bitwise Or and Exclusive Or
d74e8afc
ITB
469X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
470X<bitwise xor> X<^>
a0d0e21e 471
2cdc098b 472Binary "|" returns its operands ORed together bit by bit.
a0d0e21e 474
2cdc098b 475Binary "^" returns its operands XORed together bit by bit.
a0d0e21e 477
2cdc098b
MG
478Note that "|" and "^" have lower priority than relational operators, so
479for example the brackets are essential in a test like
480
481 print "false\n" if (8 | 2) != 10;
482
a0d0e21e 483=head2 C-style Logical And
d74e8afc 484X<&&> X<logical and> X<operator, logical, and>
a0d0e21e
LW
485
486Binary "&&" performs a short-circuit logical AND operation. That is,
487if the left operand is false, the right operand is not even evaluated.
488Scalar or list context propagates down to the right operand if it
489is evaluated.
490
491=head2 C-style Logical Or
d74e8afc 492X<||> X<operator, logical, or>
a0d0e21e
LW
493
494Binary "||" performs a short-circuit logical OR operation. That is,
495if the left operand is true, the right operand is not even evaluated.
496Scalar or list context propagates down to the right operand if it
497is evaluated.
498
c963b151 499=head2 C-style Logical Defined-Or
d74e8afc 500X<//> X<operator, logical, defined-or>
c963b151
BD
501
502Although it has no direct equivalent in C, Perl's C<//> operator is related
89d205f2 503to its C-style or. In fact, it's exactly the same as C<||>, except that it
c963b151 504tests the left hand side's definedness instead of its truth. Thus, C<\$a // \$b>
89d205f2
YO
505is similar to C<defined(\$a) || \$b> (except that it returns the value of C<\$a>
506rather than the value of C<defined(\$a)>) and is exactly equivalent to
c963b151 507C<defined(\$a) ? \$a : \$b>. This is very useful for providing default values
89d205f2 508for variables. If you actually want to test if at least one of C<\$a> and
d042e63d 509C<\$b> is defined, use C<defined(\$a // \$b)>.
c963b151 510
d042e63d
MS
511The C<||>, C<//> and C<&&> operators return the last value evaluated
512(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
513portable way to find out the home directory might be:
a0d0e21e 514
c963b151
BD
515 \$home = \$ENV{'HOME'} // \$ENV{'LOGDIR'} //
516 (getpwuid(\$<))[7] // die "You're homeless!\n";
a0d0e21e 517
5a964f20
TC
518In particular, this means that you shouldn't use this
519for selecting between two aggregates for assignment:
520
521 @a = @b || @c; # this is wrong
522 @a = scalar(@b) || @c; # really meant this
523 @a = @b ? @b : @c; # this works fine, though
524
c963b151
BD
525As more readable alternatives to C<&&>, C<//> and C<||> when used for
526control flow, Perl provides C<and>, C<err> and C<or> operators (see below).
89d205f2 527The short-circuit behavior is identical. The precedence of "and", "err"
c963b151 528and "or" is much lower, however, so that you can safely use them after a
5a964f20 529list operator without the need for parentheses:
a0d0e21e
LW
530
531 unlink "alpha", "beta", "gamma"
532 or gripe(), next LINE;
533
534With the C-style operators that would have been written like this:
535
536 unlink("alpha", "beta", "gamma")
537 || (gripe(), next LINE);
538
eeb6a2c9 539Using "or" for assignment is unlikely to do what you want; see below.
5a964f20
TC
540
d74e8afc 542X<operator, range> X<range> X<..> X<...>
a0d0e21e
LW
543
544Binary ".." is the range operator, which is really two different
fb53bbb2 545operators depending on the context. In list context, it returns a
54ae734e 546list of values counting (up by ones) from the left value to the right
2cdbc966 547value. If the left value is greater than the right value then it
fb53bbb2 548returns the empty list. The range operator is useful for writing
54ae734e 549C<foreach (1..10)> loops and for doing slice operations on arrays. In
2cdbc966
JD
550the current implementation, no temporary array is created when the
551range operator is used as the expression in C<foreach> loops, but older
552versions of Perl might burn a lot of memory when you write something
553like this:
a0d0e21e
LW
554
555 for (1 .. 1_000_000) {
556 # code
54310121 557 }
a0d0e21e 558
54ae734e
MG
559The range operator also works on strings, using the magical auto-increment,
560see below.
561
5a964f20 562In scalar context, ".." returns a boolean value. The operator is
a0d0e21e
LW
563bistable, like a flip-flop, and emulates the line-range (comma) operator
564of B<sed>, B<awk>, and various editors. Each ".." operator maintains its
565own boolean state. It is false as long as its left operand is false.
566Once the left operand is true, the range operator stays true until the
567right operand is true, I<AFTER> which the range operator becomes false
19799a22 568again. It doesn't become false till the next time the range operator is
a0d0e21e
LW
569evaluated. It can test the right operand and become false on the same
570evaluation it became true (as in B<awk>), but it still returns true once.
19799a22
GS
571If you don't want it to test the right operand till the next
572evaluation, as in B<sed>, just use three dots ("...") instead of
573two. In all other regards, "..." behaves just like ".." does.
574
575The right operand is not evaluated while the operator is in the
576"false" state, and the left operand is not evaluated while the
577operator is in the "true" state. The precedence is a little lower
578than || and &&. The value returned is either the empty string for
579false, or a sequence number (beginning with 1) for true. The
580sequence number is reset for each range encountered. The final
581sequence number in a range has the string "E0" appended to it, which
582doesn't affect its numeric value, but gives you something to search
583for if you want to exclude the endpoint. You can exclude the
584beginning point by waiting for the sequence number to be greater
df5f8116
CW
585than 1.
586
587If either operand of scalar ".." is a constant expression,
588that operand is considered true if it is equal (C<==>) to the current
589input line number (the C<\$.> variable).
590
591To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
592but that is only an issue if you use a floating point expression; when
593implicitly using C<\$.> as described in the previous paragraph, the
594comparison is C<int(EXPR) == int(\$.)> which is only an issue when C<\$.>
595is set to a floating point value and you are not reading from a file.
596Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
597you want in scalar context because each of the operands are evaluated
598using their integer representation.
599
600Examples:
a0d0e21e
LW
601
602As a scalar operator:
603
df5f8116
CW
604 if (101 .. 200) { print; } # print 2nd hundred lines, short for
605 # if (\$. == 101 .. \$. == 200) ...
9f10b797
RGS
606
607 next LINE if (1 .. /^\$/); # skip header lines, short for
df5f8116 608 # ... if (\$. == 1 .. /^\$/);
9f10b797
RGS
609 # (typically in a loop labeled LINE)
610
611 s/^/> / if (/^\$/ .. eof()); # quote body
a0d0e21e 612
5a964f20
TC
613 # parse mail messages
614 while (<>) {
615 \$in_header = 1 .. /^\$/;
df5f8116
CW
616 \$in_body = /^\$/ .. eof;
617 if (\$in_header) {
618 # ...
619 } else { # in body
620 # ...
621 }
5a964f20 622 } continue {
df5f8116 623 close ARGV if eof; # reset \$. each file
5a964f20
TC
624 }
625
acf31ca5
SF
626Here's a simple example to illustrate the difference between
627the two range operators:
628
629 @lines = (" - Foo",
630 "01 - Bar",
631 "1 - Baz",
632 " - Quux");
633
9f10b797
RGS
634 foreach (@lines) {
635 if (/0/ .. /1/) {
acf31ca5
SF
636 print "\$_\n";
637 }
638 }
639
9f10b797
RGS
640This program will print only the line containing "Bar". If
641the range operator is changed to C<...>, it will also print the
acf31ca5
SF
642"Baz" line.
643
644And now some examples as a list operator:
a0d0e21e
LW
645
646 for (101 .. 200) { print; } # print \$_ 100 times
3e3baf6d 647 @foo = @foo[0 .. \$#foo]; # an expensive no-op
a0d0e21e
LW
648 @foo = @foo[\$#foo-4 .. \$#foo]; # slice last 5 items
649
5a964f20 650The range operator (in list context) makes use of the magical
5f05dabc 651auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
652can say
653
654 @alphabet = ('A' .. 'Z');
655
54ae734e 656to get all normal letters of the English alphabet, or
a0d0e21e
LW
657
658 \$hexdigit = (0 .. 9, 'a' .. 'f')[\$num & 15];
659
660to get a hexadecimal digit, or
661
662 @z2 = ('01' .. '31'); print \$z2[\$mday];
663
ea4f5703
YST
664to get dates with leading zeros.
665
666If the final value specified is not in the sequence that the magical
667increment would produce, the sequence goes until the next value would
668be longer than the final value specified.
669
670If the initial value specified isn't part of a magical increment
671sequence (that is, a non-empty string matching "/^[a-zA-Z]*[0-9]*\z/"),
672only the initial value will be returned. So the following will only
673return an alpha:
674
675 use charnames 'greek';
676 my @greek_small = ("\N{alpha}" .. "\N{omega}");
677
678To get lower-case greek letters, use this instead:
679
680 my @greek_small = map { chr } ( ord("\N{alpha}") .. ord("\N{omega}") );
a0d0e21e 681
df5f8116
CW
682Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
683return two elements in list context.
684
685 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
686
a0d0e21e 687=head2 Conditional Operator
d74e8afc 688X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
a0d0e21e
LW
689
690Ternary "?:" is the conditional operator, just as in C. It works much
691like an if-then-else. If the argument before the ? is true, the
692argument before the : is returned, otherwise the argument after the :
cb1a09d0
693is returned. For example:
694
54310121 695 printf "I have %d dog%s.\n", \$n,
cb1a09d0
696 (\$n == 1) ? '' : "s";
697
698Scalar or list context propagates downward into the 2nd
54310121 699or 3rd argument, whichever is selected.
cb1a09d0
700
701 \$a = \$ok ? \$b : \$c; # get a scalar
702 @a = \$ok ? @b : @c; # get an array
703 \$a = \$ok ? @b : @c; # oops, that's just a count!
704
705The operator may be assigned to if both the 2nd and 3rd arguments are
706legal lvalues (meaning that you can assign to them):
a0d0e21e
LW
707
708 (\$a_or_b ? \$a : \$b) = \$c;
709
5a964f20
TC
710Because this operator produces an assignable result, using assignments
711without parentheses will get you in trouble. For example, this:
712
713 \$a % 2 ? \$a += 10 : \$a += 2
714
715Really means this:
716
717 ((\$a % 2) ? (\$a += 10) : \$a) += 2
718
719Rather than this:
720
721 (\$a % 2) ? (\$a += 10) : (\$a += 2)
722
19799a22
GS
723That should probably be written more simply as:
724
725 \$a += (\$a % 2) ? 10 : 2;
726
4633a7c4 727=head2 Assignment Operators
d74e8afc 728X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
5ac3b81c 729X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
d74e8afc 730X<%=> X<^=> X<x=>
a0d0e21e
LW
731
732"=" is the ordinary assignment operator.
733
734Assignment operators work as in C. That is,
735
736 \$a += 2;
737
738is equivalent to
739
740 \$a = \$a + 2;
741
742although without duplicating any side effects that dereferencing the lvalue
54310121 743might trigger, such as from tie(). Other assignment operators work similarly.
744The following are recognized:
a0d0e21e
LW
745
746 **= += *= &= <<= &&=
9f10b797
RGS
747 -= /= |= >>= ||=
748 .= %= ^= //=
749 x=
a0d0e21e 750
19799a22 751Although these are grouped by family, they all have the precedence
a0d0e21e
LW
752of assignment.
753
b350dd2f
GS
754Unlike in C, the scalar assignment operator produces a valid lvalue.
755Modifying an assignment is equivalent to doing the assignment and
756then modifying the variable that was assigned to. This is useful
757for modifying a copy of something, like this:
a0d0e21e
LW
758
759 (\$tmp = \$global) =~ tr [A-Z] [a-z];
760
761Likewise,
762
763 (\$a += 2) *= 3;
764
765is equivalent to
766
767 \$a += 2;
768 \$a *= 3;
769
b350dd2f
GS
770Similarly, a list assignment in list context produces the list of
771lvalues assigned to, and a list assignment in scalar context returns
772the number of elements produced by the expression on the right hand
773side of the assignment.
774
748a9306 775=head2 Comma Operator
d74e8afc 776X<comma> X<operator, comma> X<,>
a0d0e21e 777
5a964f20 778Binary "," is the comma operator. In scalar context it evaluates
a0d0e21e
LW
779its left argument, throws that value away, then evaluates its right
780argument and returns that value. This is just like C's comma operator.
781
5a964f20 782In list context, it's just the list argument separator, and inserts
ed5c6d31
PJ
783both its arguments into the list. These arguments are also evaluated
784from left to right.
a0d0e21e 785
d042e63d 786The C<< => >> operator is a synonym for the comma, but forces any word
719b43e8 787(consisting entirely of word characters) to its left to be interpreted
a44e5664
MS
788as a string (as of 5.001). This includes words that might otherwise be
789considered a constant or function call.
790
791 use constant FOO => "something";
792
793 my %h = ( FOO => 23 );
794
795is equivalent to:
796
797 my %h = ("FOO", 23);
798
799It is I<NOT>:
800
801 my %h = ("something", 23);
802
803If the argument on the left is not a word, it is first interpreted as
804an expression, and then the string value of that is used.
719b43e8
RGS
805
806The C<< => >> operator is helpful in documenting the correspondence
807between keys and values in hashes, and other paired elements in lists.
748a9306 808
a44e5664
MS
809 %hash = ( \$key => \$value );
811
a0d0e21e 812=head2 List Operators (Rightward)
d74e8afc 813X<operator, list, rightward> X<list operator>
a0d0e21e
LW
814
815On the right side of a list operator, it has very low precedence,
816such that it controls all comma-separated expressions found there.
817The only operators with lower precedence are the logical operators
818"and", "or", and "not", which may be used to evaluate calls to list
819operators without the need for extra parentheses:
820
821 open HANDLE, "filename"
822 or die "Can't open: \$!\n";
823
5ba421f6 824See also discussion of list operators in L<Terms and List Operators (Leftward)>.
a0d0e21e
LW
825
d74e8afc 827X<operator, logical, not> X<not>
a0d0e21e
LW
828
829Unary "not" returns the logical negation of the expression to its right.
830It's the equivalent of "!" except for the very low precedence.
831
d74e8afc 833X<operator, logical, and> X<and>
a0d0e21e
LW
834
835Binary "and" returns the logical conjunction of the two surrounding
836expressions. It's equivalent to && except for the very low
5f05dabc 837precedence. This means that it short-circuits: i.e., the right
a0d0e21e
LW
838expression is evaluated only if the left expression is true.
839
c963b151 840=head2 Logical or, Defined or, and Exclusive Or
d74e8afc
ITB
841X<operator, logical, or> X<operator, logical, xor> X<operator, logical, err>
842X<operator, logical, defined or> X<operator, logical, exclusive or>
843X<or> X<xor> X<err>
a0d0e21e
LW
844
845Binary "or" returns the logical disjunction of the two surrounding
5a964f20
TC
846expressions. It's equivalent to || except for the very low precedence.
847This makes it useful for control flow
848
849 print FH \$data or die "Can't write to FH: \$!";
850
851This means that it short-circuits: i.e., the right expression is evaluated
852only if the left expression is false. Due to its precedence, you should
853probably avoid using this for assignment, only for control flow.
854
855 \$a = \$b or \$c; # bug: this is wrong
856 (\$a = \$b) or \$c; # really means this
857 \$a = \$b || \$c; # better written this way
858
19799a22 859However, when it's a list-context assignment and you're trying to use
5a964f20
TC
860"||" for control flow, you probably need "or" so that the assignment
861takes higher precedence.
862
863 @info = stat(\$file) || die; # oops, scalar sense of stat!
864 @info = stat(\$file) or die; # better, now @info gets its due
865
c963b151
BD
866Then again, you could always use parentheses.
867
9f10b797
RGS
868Binary "err" is equivalent to C<//>--it's just like binary "or", except it
869tests its left argument's definedness instead of its truth. There are two
870ways to remember "err": either because many functions return C<undef> on
871an B<err>or, or as a sort of correction: C<\$a = (\$b err 'default')>. This
872keyword is only available when the 'err' feature is enabled: see
a0d0e21e
LW
874
875Binary "xor" returns the exclusive-OR of the two surrounding expressions.
876It cannot short circuit, of course.
877
878=head2 C Operators Missing From Perl
d74e8afc
ITB
879X<operator, missing from perl> X<&> X<*>
880X<typecasting> X<(TYPE)>
a0d0e21e
LW
881
882Here is what C has that Perl doesn't:
883
884=over 8
885
886=item unary &
887
888Address-of operator. (But see the "\" operator for taking a reference.)
889
890=item unary *
891
54310121 892Dereference-address operator. (Perl's prefix dereferencing
a0d0e21e
LW
893operators are typed: \$, @, %, and &.)
894
895=item (TYPE)
896
19799a22 897Type-casting operator.
a0d0e21e
LW
898
899=back
900
5f05dabc 901=head2 Quote and Quote-like Operators
89d205f2 902X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
d74e8afc
ITB
903X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
904X<escape sequence> X<escape>
905
a0d0e21e
LW
906
907While we usually think of quotes as literal values, in Perl they
908function as operators, providing various kinds of interpolating and
909pattern matching capabilities. Perl provides customary quote characters
910for these behaviors, but also provides a way for you to choose your
911quote character for any of them. In the following table, a C<{}> represents
9f10b797 912any pair of delimiters you choose.
a0d0e21e 913
TP
914 Customary Generic Meaning Interpolates
915 '' q{} Literal no
916 "" qq{} Literal yes
af9219ee 917 `` qx{} Command yes*
2c268ad5 918 qw{} Word list no
af9219ee
MG
919 // m{} Pattern match yes*
920 qr{} Pattern yes*
921 s{}{} Substitution yes*
2c268ad5 922 tr{}{} Transliteration no (but see below)
7e3b091d 923 <<EOF here-doc yes*
a0d0e21e 924
af9219ee
MG
925 * unless the delimiter is ''.
926
87275199
GS
927Non-bracketing delimiters use the same character fore and aft, but the four
928sorts of brackets (round, angle, square, curly) will all nest, which means
9f10b797 929that
87275199 930
9f10b797 931 q{foo{bar}baz}
35f2feb0 932
9f10b797 933is the same as
87275199
GS
934
935 'foo{bar}baz'
936
937Note, however, that this does not always work for quoting Perl code:
938
939 \$s = q{ if(\$a eq "}") ... }; # WRONG
940
83df6a1d
JH
941is a syntax error. The C<Text::Balanced> module (from CPAN, and
942starting from Perl 5.8 part of the standard distribution) is able
943to do this properly.
87275199 944
19799a22 945There can be whitespace between the operator and the quoting
fb73857a 946characters, except when C<#> is being used as the quoting character.
19799a22
GS
947C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
948operator C<q> followed by a comment. Its argument will be taken
949from the next line. This allows you to write:
fb73857a 950
951 s {foo} # Replace foo
952 {bar} # with bar.
953
904501ec
MG
954The following escape sequences are available in constructs that interpolate
955and in transliterations.
d74e8afc 956X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N>
a0d0e21e 957
6ee5d4e7 958 \t tab (HT, TAB)
5a964f20 959 \n newline (NL)
6ee5d4e7 960 \r return (CR)
961 \f form feed (FF)
962 \b backspace (BS)
963 \a alarm (bell) (BEL)
964 \e escape (ESC)
ee9f418e
WL
965 \033 octal char (example: ESC)
966 \x1b hex char (example: ESC)
967 \x{263a} wide hex char (example: SMILEY)
968 \c[ control char (example: ESC)
95cc3e0c 969 \N{name} named Unicode character
ee9f418e
WL
971The character following C<\c> is mapped to some other character by
972converting letters to upper case and then (on ASCII systems) by inverting
973the 7th bit (0x40). The most interesting range is from '@' to '_'
974(0x40 through 0x5F), resulting in a control character from 0x00
975through 0x1F. A '?' maps to the DEL character. On EBCDIC systems only
976'@', the letters, '[', '\', ']', '^', '_' and '?' will work, resulting
977in 0x00 through 0x1F and 0x7F.
978
4c77eaa2 979B<NOTE>: Unlike C and other languages, Perl has no \v escape sequence for
ee9f418e 980the vertical tab (VT - ASCII 11), but you may use C<\ck> or C<\x0b>.
4c77eaa2 981
904501ec
MG
982The following escape sequences are available in constructs that interpolate
983but not in transliterations.
d74e8afc 984X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q>
904501ec 985
a0d0e21e
LW
986 \l lowercase next char
987 \u uppercase next char
988 \L lowercase till \E
989 \U uppercase till \E
990 \E end case modification
1d2dff63 991 \Q quote non-word characters till \E
a0d0e21e 992
95cc3e0c
JH
993If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
994C<\u> and C<\U> is taken from the current locale. See L<perllocale>.
995If Unicode (for example, C<\N{}> or wide hex characters of 0x100 or
996beyond) is being used, the case map used by C<\l>, C<\L>, C<\u> and
997C<\U> is as defined by Unicode. For documentation of C<\N{name}>,
998see L<charnames>.
a034a98d 999
5a964f20
TC
1000All systems use the virtual C<"\n"> to represent a line terminator,
1001called a "newline". There is no such thing as an unvarying, physical
19799a22 1002newline character. It is only an illusion that the operating system,
5a964f20
TC
1003device drivers, C libraries, and Perl all conspire to preserve. Not all
1004systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
1005on a Mac, these are reversed, and on systems without line terminator,
1006printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
1007you mean a "newline" for your system, but use the literal ASCII when you
1008need an exact character. For example, most networking protocols expect
2a380090 1009and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
1010and although they often accept just C<"\012">, they seldom tolerate just
1011C<"\015">. If you get in the habit of using C<"\n"> for networking,
1012you may be burned some day.
d74e8afc
ITB
1013X<newline> X<line terminator> X<eol> X<end of line>
1014X<\n> X<\r> X<\r\n>
5a964f20 1015
904501ec
MG
1016For constructs that do interpolate, variables beginning with "C<\$>"
1017or "C<@>" are interpolated. Subscripted variables such as C<\$a[3]> or
A
1018C<< \$href->{key}[0] >> are also interpolated, as are array and hash slices.
1019But method calls such as C<< \$obj->meth >> are not.
af9219ee
MG
1020
1021Interpolating an array or slice interpolates the elements in order,
1022separated by the value of C<\$">, so is equivalent to interpolating
6deea57f
TS
1023C<join \$", @array>. "Punctuation" arrays such as C<@*> are only
1024interpolated if the name is enclosed in braces C<@{*}>, but special
1025arrays C<@_>, C<@+>, and C<@-> are interpolated, even without braces.
af9219ee 1026
89d205f2
YO
1027You cannot include a literal C<\$> or C<@> within a C<\Q> sequence.
1028An unescaped C<\$> or C<@> interpolates the corresponding variable,
1d2dff63 1029while escaping will cause the literal string C<\\$> to be inserted.
89d205f2 1030You'll need to write something like C<m/\Quser\E\@\Qhost/>.
1d2dff63 1031
a0d0e21e
LW
1032Patterns are subject to an additional level of interpretation as a
1033regular expression. This is done as a second pass, after variables are
1034interpolated, so that regular expressions may be incorporated into the
1035pattern from the variables. If this is not what you want, use C<\Q> to
1036interpolate a variable literally.
1037
19799a22
GS
1038Apart from the behavior described above, Perl does not expand
1039multiple levels of interpolation. In particular, contrary to the
1040expectations of shell programmers, back-quotes do I<NOT> interpolate
1041within double quotes, nor do single quotes impede evaluation of
1042variables when used within double quotes.
a0d0e21e 1043
5f05dabc 1044=head2 Regexp Quote-Like Operators
d74e8afc 1045X<operator, regexp>
cb1a09d0 1046
5f05dabc 1047Here are the quote-like operators that apply to pattern
cb1a09d0
1048matching and related activities.
1049
a0d0e21e
LW
1050=over 8
1051
1052=item ?PATTERN?
d74e8afc 1053X<?>
a0d0e21e
LW
1054
1055This is just like the C</pattern/> search, except that it matches only
1056once between calls to the reset() operator. This is a useful
5f05dabc 1057optimization when you want to see only the first occurrence of
a0d0e21e
LW
1058something in each file of a set of files, for instance. Only C<??>
1059patterns local to the current package are reset.
1060
5a964f20
TC
1061 while (<>) {
1062 if (?^\$?) {
1063 # blank line between header and body
1064 }
1065 } continue {
1066 reset if eof; # clear ?? status for next file
1067 }
1068
483b4840 1069This usage is vaguely deprecated, which means it just might possibly
19799a22
GS
1070be removed in some distant future version of Perl, perhaps somewhere
1071around the year 2168.
a0d0e21e 1072
6fcf1495 1073=item m/PATTERN/cgimosxk
89d205f2
YO
1074X<m> X<operator, match>
1075X<regexp, options> X<regexp> X<regex, options> X<regex>
d74e8afc 1076X</c> X</i> X</m> X</o> X</s> X</x>
a0d0e21e 1077
cde0cee5 1078=item /PATTERN/cgimosxk
a0d0e21e 1079
5a964f20 1080Searches a string for a pattern match, and in scalar context returns
19799a22
GS
1081true if it succeeds, false if it fails. If no string is specified
1082via the C<=~> or C<!~> operator, the \$_ string is searched. (The
1083string specified with C<=~> need not be an lvalue--it may be the
1084result of an expression evaluation, but remember the C<=~> binds
1085rather tightly.) See also L<perlre>. See L<perllocale> for
1086discussion of additional considerations that apply when C<use locale>
1087is in effect.
a0d0e21e
LW
1088
1089Options are:
1090
a0d0e21e
LW
1091 i Do case-insensitive pattern matching.
1092 m Treat string as multiple lines.
a0d0e21e
LW
1093 s Treat string as single line.
1094 x Use extended regular expressions.
cde0cee5
YO
1095 g Match globally, i.e., find all occurrences.
1096 c Do not reset search position on a failed match when /g is in effect.
1097 o Compile pattern only once.
1098 k Keep a copy of the matched string so that \${^MATCH} and friends
1099 will be defined.
a0d0e21e
LW
1100
1101If "/" is the delimiter then the initial C<m> is optional. With the C<m>
89d205f2 1102you can use any pair of non-alphanumeric, non-whitespace characters
19799a22
GS
1103as delimiters. This is particularly useful for matching path names
1104that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
7bac28a0 1105the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
19799a22 1106If "'" is the delimiter, no interpolation is performed on the PATTERN.
a0d0e21e
LW
1107
1108PATTERN may contain variables, which will be interpolated (and the
f70b4f9c 1109pattern recompiled) every time the pattern search is evaluated, except
1f247705
GS
1110for when the delimiter is a single quote. (Note that C<\$(>, C<\$)>, and
1111C<\$|> are not interpolated because they look like end-of-string tests.)
f70b4f9c
AB
1112If you want such a pattern to be compiled only once, add a C</o> after
1113the trailing delimiter. This avoids expensive run-time recompilations,
1114and is useful when the value you are interpolating won't change over
1115the life of the script. However, mentioning C</o> constitutes a promise
1116that you won't change the variables in the pattern. If you change them,
13a2d996 1117Perl won't even notice. See also L<"qr/STRING/imosx">.
a0d0e21e 1118
5a964f20 1119If the PATTERN evaluates to the empty string, the last
d65afb4b
HS
1120I<successfully> matched regular expression is used instead. In this
1121case, only the C<g> and C<c> flags on the empty pattern is honoured -
1122the other flags are taken from the original pattern. If no match has
1123previously succeeded, this will (silently) act instead as a genuine
1124empty pattern (which will always match).
a0d0e21e 1125
89d205f2
YO
1126Note that it's possible to confuse Perl into thinking C<//> (the empty
1127regex) is really C<//> (the defined-or operator). Perl is usually pretty
1128good about this, but some pathological cases might trigger this, such as
1129C<\$a///> (is that C<(\$a) / (//)> or C<\$a // />?) and C<print \$fh //>
1130(C<print \$fh(//> or C<print(\$fh //>?). In all of these examples, Perl
1131will assume you meant defined-or. If you meant the empty regex, just
1132use parentheses or spaces to disambiguate, or even prefix the empty
c963b151
BD
1133regex with an C<m> (so C<//> becomes C<m//>).
1134
19799a22 1135If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1136list consisting of the subexpressions matched by the parentheses in the
f7e33566
GS
1137pattern, i.e., (C<\$1>, C<\$2>, C<\$3>...). (Note that here C<\$1> etc. are
1138also set, and that this differs from Perl 4's behavior.) When there are
1139no parentheses in the pattern, the return value is the list C<(1)> for
1140success. With or without parentheses, an empty list is returned upon
1141failure.
a0d0e21e
LW
1142
1143Examples:
1144
1145 open(TTY, '/dev/tty');
1146 <TTY> =~ /^y/i && foo(); # do foo if desired
1147
1148 if (/Version: *([0-9.]*)/) { \$version = \$1; }
1149
1150 next if m#^/usr/spool/uucp#;
1151
1152 # poor man's grep
1153 \$arg = shift;
1154 while (<>) {
1155 print if /\$arg/o; # compile only once
1156 }
1157
1158 if ((\$F1, \$F2, \$Etc) = (\$foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
1159
1160This last example splits \$foo into the first two words and the
5f05dabc 1161remainder of the line, and assigns those three fields to \$F1, \$F2, and
1162\$Etc. The conditional is true if any variables were assigned, i.e., if
a0d0e21e
LW
1163the pattern matched.
1164
19799a22
GS
1165The C</g> modifier specifies global pattern matching--that is,
1166matching as many times as possible within the string. How it behaves
1167depends on the context. In list context, it returns a list of the
1168substrings matched by any capturing parentheses in the regular
1169expression. If there are no parentheses, it returns a list of all
1170the matched strings, as if there were parentheses around the whole
1171pattern.
a0d0e21e 1172
7e86de3e 1173In scalar context, each execution of C<m//g> finds the next match,
19799a22 1174returning true if it matches, and false if there is no further match.
7e86de3e
MG
1175The position after the last match can be read or set using the pos()
1176function; see L<perlfunc/pos>. A failed match normally resets the
1177search position to the beginning of the string, but you can avoid that
1178by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
1179string also resets the search position.
c90c0ff4 1180
1181You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
1182zero-width assertion that matches the exact position where the previous
5d43e42d
DC
1183C<m//g>, if any, left off. Without the C</g> modifier, the C<\G> assertion
1184still anchors at pos(), but the match is of course only attempted once.
1185Using C<\G> without C</g> on a target string that has not previously had a
1186C</g> match applied to it is the same as using the C<\A> assertion to match
fe4b3f22
RGS
1187the beginning of the string. Note also that, currently, C<\G> is only
1188properly supported when anchored at the very beginning of the pattern.
c90c0ff4 1189
1190Examples:
a0d0e21e
LW
1191
1192 # list context
1193 (\$one,\$five,\$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1194
1195 # scalar context
5d43e42d 1196 \$/ = "";
19799a22
GS
1197 while (defined(\$paragraph = <>)) {
1198 while (\$paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
1199 \$sentences++;
a0d0e21e
LW
1200 }
1201 }
1202 print "\$sentences\n";
1203
c90c0ff4 1204 # using m//gc with \G
137443ea 1205 \$_ = "ppooqppqq";
44a8e56a 1206 while (\$i++ < 2) {
1207 print "1: '";
c90c0ff4 1208 print \$1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1209 print "2: '";
c90c0ff4 1210 print \$1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1211 print "3: '";
c90c0ff4 1212 print \$1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1213 }
5d43e42d 1214 print "Final: '\$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a 1215
1216The last example should print:
1217
1218 1: 'oo', pos=4
137443ea 1219 2: 'q', pos=5
44a8e56a 1220 3: 'pp', pos=7
1221 1: '', pos=7
137443ea 1222 2: 'q', pos=8
1223 3: '', pos=8
5d43e42d
DC
1224 Final: 'q', pos=8
1225
1226Notice that the final match matched C<q> instead of C<p>, which a match
1227without the C<\G> anchor would have done. Also note that the final match
1228did not update C<pos> -- C<pos> is only updated on a C</g> match. If the
1229final match did indeed match C<p>, it's a good bet that you're running an
1230older (pre-5.6.0) Perl.
44a8e56a 1231
c90c0ff4 1232A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1233combine several regexps like this to process a string part-by-part,
c90c0ff4 1234doing different actions depending on which regexp matched. Each
1235regexp tries to match where the previous one leaves off.
e7ea3e70 1236
3fe9a6f1 1237 \$_ = <<'EOL';
63acfd00 1238 \$url = URI::URL->new( "http://www/" ); die if \$url eq "xXx";
3fe9a6f1 1239 EOL
1240 LOOP:
e7ea3e70 1241 {
c90c0ff4 1242 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
1243 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
1244 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
1245 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
1246 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
1247 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
1248 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
e7ea3e70
IZ
1249 print ". That's all!\n";
1250 }
1251
1252Here is the output (split into several lines):
1253
1254 line-noise lowercase line-noise lowercase UPPERCASE line-noise
1255 UPPERCASE line-noise lowercase line-noise lowercase line-noise
1256 lowercase lowercase line-noise lowercase lowercase line-noise
1257 MiXeD line-noise. That's all!
44a8e56a 1258
a0d0e21e 1259=item q/STRING/
5d44bfff 1260X<q> X<quote, single> X<'> X<''>
a0d0e21e 1261
5d44bfff 1262=item 'STRING'
a0d0e21e 1263
19799a22 1264A single-quoted, literal string. A backslash represents a backslash
68dc0745 1265unless followed by the delimiter or another backslash, in which case
1266the delimiter or backslash is interpolated.
a0d0e21e
LW
1267
1268 \$foo = q!I said, "You said, 'She said it.'"!;
1269 \$bar = q('This is it.');
68dc0745 1270 \$baz = '\n'; # a two-character string
a0d0e21e
LW
1271
1272=item qq/STRING/
d74e8afc 1273X<qq> X<quote, double> X<"> X<"">
a0d0e21e
LW
1274
1275=item "STRING"
1276
1277A double-quoted, interpolated string.
1278
1279 \$_ .= qq
1280 (*** The previous line contains the naughty word "\$1".\n)
19799a22 1281 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 1282 \$baz = "\n"; # a one-character string
a0d0e21e 1283
eec2d3df 1284=item qr/STRING/imosx
d74e8afc 1285X<qr> X</i> X</m> X</o> X</s> X</x>
eec2d3df 1286
322edccd 1287This operator quotes (and possibly compiles) its I<STRING> as a regular
19799a22
GS
1288expression. I<STRING> is interpolated the same way as I<PATTERN>
1289in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1290is done. Returns a Perl value which may be used instead of the
1291corresponding C</STRING/imosx> expression.
4b6a7270
IZ
1292
1293For example,
1294
1295 \$rex = qr/my.STRING/is;
1296 s/\$rex/foo/;
1297
1298is equivalent to
1299
1300 s/my.STRING/foo/is;
1301
1302The result may be used as a subpattern in a match:
eec2d3df
GS
1303
1304 \$re = qr/\$pattern/;
0a92e3a8
GS
1305 \$string =~ /foo\${re}bar/; # can be interpolated in other patterns
1306 \$string =~ \$re; # or used standalone
4b6a7270
IZ
1307 \$string =~ /\$re/; # or this way
1308
1309Since Perl may compile the pattern at the moment of execution of qr()
19799a22 1310operator, using qr() may have speed advantages in some situations,
4b6a7270
IZ
1311notably if the result of qr() is used standalone:
1312
1313 sub match {
1314 my \$patterns = shift;
1315 my @compiled = map qr/\$_/i, @\$patterns;
1316 grep {
1317 my \$success = 0;
a7665c5e 1318 foreach my \$pat (@compiled) {
4b6a7270
IZ
1319 \$success = 1, last if /\$pat/;
1320 }
1321 \$success;
1322 } @_;
1323 }
1324
19799a22
GS
1325Precompilation of the pattern into an internal representation at
1326the moment of qr() avoids a need to recompile the pattern every
1327time a match C</\$pat/> is attempted. (Perl has many other internal
1328optimizations, but none would be triggered in the above example if
1329we did not use qr() operator.)
eec2d3df
GS
1330
1331Options are:
1332
1333 i Do case-insensitive pattern matching.
1334 m Treat string as multiple lines.
1335 o Compile pattern only once.
1336 s Treat string as single line.
1337 x Use extended regular expressions.
1338
0a92e3a8
GS
1339See L<perlre> for additional information on valid syntax for STRING, and
1340for a detailed look at the semantics of regular expressions.
1341
a0d0e21e 1342=item qx/STRING/
d74e8afc 1343X<qx> X<`> X<``> X<backtick>
a0d0e21e
LW
1344
1345=item `STRING`
1346
43dd4d21
JH
1347A string which is (possibly) interpolated and then executed as a
1348system command with C</bin/sh> or its equivalent. Shell wildcards,
1349pipes, and redirections will be honored. The collected standard
1350output of the command is returned; standard error is unaffected. In
1351scalar context, it comes back as a single (potentially multi-line)
1352string, or undef if the command failed. In list context, returns a
1353list of lines (however you've defined lines with \$/ or
1354\$INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
5a964f20
TC
1355
1356Because backticks do not affect standard error, use shell file descriptor
1357syntax (assuming the shell supports this) if you care to address this.
1358To capture a command's STDERR and STDOUT together:
a0d0e21e 1359
5a964f20
TC
1360 \$output = `cmd 2>&1`;
1361
1362To capture a command's STDOUT but discard its STDERR:
1363
1364 \$output = `cmd 2>/dev/null`;
1365
1366To capture a command's STDERR but discard its STDOUT (ordering is
1367important here):
1368
1369 \$output = `cmd 2>&1 1>/dev/null`;
1370
1371To exchange a command's STDOUT and STDERR in order to capture the STDERR
1372but leave its STDOUT to come out the old STDERR:
1373
1374 \$output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1375
1376To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
1377to redirect them separately to files, and then read from those files
1378when the program is done:
5a964f20 1379
2359510d 1380 system("program args 1>program.stdout 2>program.stderr");
5a964f20 1381
30398227
SP
1382The STDIN filehandle used by the command is inherited from Perl's STDIN.
1383For example:
1384
1385 open BLAM, "blam" || die "Can't open: \$!";
1386 open STDIN, "<&BLAM";
1387 print `sort`;
1388
1389will print the sorted contents of the file "blam".
1390
5a964f20
TC
1391Using single-quote as a delimiter protects the command from Perl's
1392double-quote interpolation, passing it on to the shell instead:
1393
1394 \$perl_info = qx(ps \$\$); # that's Perl's \$\$
1395 \$shell_info = qx'ps \$\$'; # that's the new shell's \$\$
1396
19799a22 1397How that string gets evaluated is entirely subject to the command
5a964f20
TC
1398interpreter on your system. On most platforms, you will have to protect
1399shell metacharacters if you want them treated literally. This is in
1400practice difficult to do, as it's unclear how to escape which characters.
1401See L<perlsec> for a clean and safe example of a manual fork() and exec()
1402to emulate backticks safely.
a0d0e21e 1403
bb32b41a
GS
1404On some platforms (notably DOS-like ones), the shell may not be
1405capable of dealing with multiline commands, so putting newlines in
1406the string may not get you what you want. You may be able to evaluate
1407multiple commands in a single line by separating them with the command
1408separator character, if your shell supports that (e.g. C<;> on many Unix
1409shells; C<&> on the Windows NT C<cmd> shell).
1410
0f897271
GS
1411Beginning with v5.6.0, Perl will attempt to flush all files opened for
1412output before starting the child process, but this may not be supported
1413on some platforms (see L<perlport>). To be safe, you may need to set
1414C<\$|> (\$AUTOFLUSH in English) or call the C<autoflush()> method of
1415C<IO::Handle> on any open handles.
1416
bb32b41a
GS
1417Beware that some command shells may place restrictions on the length
1418of the command line. You must ensure your strings don't exceed this
1419limit after any necessary interpolations. See the platform-specific
1420release notes for more details about your particular environment.
1421
5a964f20
TC
1422Using this operator can lead to programs that are difficult to port,
1423because the shell commands called vary between systems, and may in
1424fact not be present at all. As one example, the C<type> command under
1425the POSIX shell is very different from the C<type> command under DOS.
1426That doesn't mean you should go out of your way to avoid backticks
1427when they're the right way to get something done. Perl was made to be
1428a glue language, and one of the things it glues together is commands.
1429Just understand what you're getting yourself into.
bb32b41a 1430
da87341d 1431See L</"I/O Operators"> for more discussion.
a0d0e21e 1432
945c54fd 1433=item qw/STRING/
d74e8afc 1434X<qw> X<quote, list> X<quote, words>
945c54fd
JH
1435
1436Evaluates to a list of the words extracted out of STRING, using embedded
1437whitespace as the word delimiters. It can be understood as being roughly
1438equivalent to:
1439
1440 split(' ', q/STRING/);
1441
efb1e162
CW
1442the differences being that it generates a real list at compile time, and
1443in scalar context it returns the last element in the list. So
945c54fd
JH
1444this expression:
1445
1446 qw(foo bar baz)
1447
1448is semantically equivalent to the list:
1449
1450 'foo', 'bar', 'baz'
1451
1452Some frequently seen examples:
1453
1454 use POSIX qw( setlocale localeconv )
1455 @EXPORT = qw( foo bar baz );
1456
1457A common mistake is to try to separate the words with comma or to
1458put comments into a multi-line C<qw>-string. For this reason, the
89d205f2 1459C<use warnings> pragma and the B<-w> switch (that is, the C<\$^W> variable)
945c54fd
JH
1460produces warnings if the STRING contains the "," or the "#" character.
1461
cde0cee5 1462=item s/PATTERN/REPLACEMENT/egimosxk
d74e8afc
ITB
1463X<substitute> X<substitution> X<replace> X<regexp, replace>
1464X<regexp, substitute> X</e> X</g> X</i> X</m> X</o> X</s> X</x>
a0d0e21e
LW
1465
1466Searches a string for a pattern, and if found, replaces that pattern
1467with the replacement text and returns the number of substitutions
e37d713d 1468made. Otherwise it returns false (specifically, the empty string).
a0d0e21e
LW
1469
1470If no string is specified via the C<=~> or C<!~> operator, the C<\$_>
1471variable is searched and modified. (The string specified with C<=~> must
5a964f20 1472be scalar variable, an array element, a hash element, or an assignment
5f05dabc 1473to one of those, i.e., an lvalue.)
a0d0e21e 1474
19799a22 1475If the delimiter chosen is a single quote, no interpolation is
a0d0e21e
LW
1476done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1477PATTERN contains a \$ that looks like a variable rather than an
1478end-of-string test, the variable will be interpolated into the pattern
5f05dabc 1479at run-time. If you want the pattern compiled only once the first time
a0d0e21e 1480the variable is interpolated, use the C</o> option. If the pattern
5a964f20 1481evaluates to the empty string, the last successfully executed regular
a0d0e21e 1482expression is used instead. See L<perlre> for further explanation on these.
5a964f20 1483See L<perllocale> for discussion of additional considerations that apply
a034a98d 1484when C<use locale> is in effect.
a0d0e21e
LW
1485
1486Options are:
1487
a0d0e21e
LW
1488 i Do case-insensitive pattern matching.
1489 m Treat string as multiple lines.
a0d0e21e
LW
1490 s Treat string as single line.
1491 x Use extended regular expressions.
cde0cee5
YO
1492 g Replace globally, i.e., all occurrences.
1493 o Compile pattern only once.
1494 k Keep a copy of the original string so \${^MATCH} and friends
1495 will be defined.
1496 e Evaluate the right side as an expression.
1497
a0d0e21e
LW
1498
1499Any non-alphanumeric, non-whitespace delimiter may replace the
1500slashes. If single quotes are used, no interpretation is done on the
e37d713d 1501replacement string (the C</e> modifier overrides this, however). Unlike
54310121 1502Perl 4, Perl 5 treats backticks as normal delimiters; the replacement
e37d713d 1503text is not evaluated as a command. If the
a0d0e21e 1504PATTERN is delimited by bracketing quotes, the REPLACEMENT has its own
5f05dabc 1505pair of quotes, which may or may not be bracketing quotes, e.g.,
35f2feb0 1506C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
cec88af6
GS
1507replacement portion to be treated as a full-fledged Perl expression
1508and evaluated right then and there. It is, however, syntax checked at
1509compile-time. A second C<e> modifier will cause the replacement portion
1510to be C<eval>ed before being run as a Perl expression.
a0d0e21e
LW
1511
1512Examples:
1513
1514 s/\bgreen\b/mauve/g; # don't change wintergreen
1515
1516 \$path =~ s|/usr/bin|/usr/local/bin|;
1517
1518 s/Login: \$foo/Login: \$bar/; # run-time pattern
1519
5a964f20 1520 (\$foo = \$bar) =~ s/this/that/; # copy first, then change
a0d0e21e 1521
5a964f20 1522 \$count = (\$paragraph =~ s/Mister\b/Mr./g); # get change-count
a0d0e21e
LW
1523
1524 \$_ = 'abc123xyz';
1525 s/\d+/\$&*2/e; # yields 'abc246xyz'
1526 s/\d+/sprintf("%5d",\$&)/e; # yields 'abc 246xyz'
1527 s/\w/\$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1528
1529 s/%(.)/\$percent{\$1}/g; # change percent escapes; no /e
1530 s/%(.)/\$percent{\$1} || \$&/ge; # expr now, so /e
023a2b40 1531 s/^=(\w+)/pod(\$1)/ge; # use function call
a0d0e21e 1532
5a964f20
TC
1533 # expand variables in \$_, but dynamics only, using
1534 # symbolic dereferencing
1535 s/\\$(\w+)/\${\$1}/g;
1536
cec88af6
GS
1537 # Add one to the value of any numbers in the string
1538 s/(\d+)/1 + \$1/eg;
1539
1540 # This will expand any embedded scalar variable
1541 # (including lexicals) in \$_ : First \$1 is interpolated
1542 # to the variable name, and then evaluated
a0d0e21e
LW
1543 s/(\\$\w+)/\$1/eeg;
1544
5a964f20 1545 # Delete (most) C comments.
a0d0e21e 1546 \$program =~ s {
4633a7c4
LW
1547 /\* # Match the opening delimiter.
1548 .*? # Match a minimal number of characters.
1549 \*/ # Match the closing delimiter.
a0d0e21e
LW
1550 } []gsx;
1551
6b0ac556 1552 s/^\s*(.*?)\s*\$/\$1/; # trim whitespace in \$_, expensively
5a964f20 1553
6b0ac556 1554 for (\$variable) { # trim whitespace in \$variable, cheap
5a964f20
TC
1555 s/^\s+//;
1556 s/\s+\$//;
1557 }
a0d0e21e
LW
1558
1559 s/([^ ]*) *([^ ]*)/\$2 \$1/; # reverse 1st two fields
1560
54310121 1561Note the use of \$ instead of \ in the last example. Unlike
35f2feb0
GS
1562B<sed>, we use the \<I<digit>> form in only the left hand side.
1563Anywhere else it's \$<I<digit>>.
a0d0e21e 1564
5f05dabc 1565Occasionally, you can't use just a C</g> to get all the changes
19799a22 1566to occur that you might want. Here are two common cases:
a0d0e21e
LW
1567
1568 # put commas in the right places in an integer
89d205f2 1569 1 while s/(\d)(\d\d\d)(?!\d)/\$1,\$2/g;
a0d0e21e
LW
1570
1571 # expand tabs to 8-column spacing
1572 1 while s/\t+/' ' x (length(\$&)*8 - length(\$`)%8)/e;
1573
6940069f 1574=item tr/SEARCHLIST/REPLACEMENTLIST/cds
d74e8afc 1575X<tr> X<y> X<transliterate> X</c> X</d> X</s>
a0d0e21e 1576
6940069f 1577=item y/SEARCHLIST/REPLACEMENTLIST/cds
a0d0e21e 1578
2c268ad5 1579Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
1580with the corresponding character in the replacement list. It returns
1581the number of characters replaced or deleted. If no string is
2c268ad5 1582specified via the =~ or !~ operator, the \$_ string is transliterated. (The
54310121 1583string specified with =~ must be a scalar variable, an array element, a
1584hash element, or an assignment to one of those, i.e., an lvalue.)
89d205f2 1586A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
2c268ad5 1587does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121 1588For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1589SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1590its own pair of quotes, which may or may not be bracketing quotes,
2c268ad5 1591e.g., C<tr[A-Z][a-z]> or C<tr(+\-*/)/ABCD/>.
a0d0e21e 1592
cc255d5f 1593Note that C<tr> does B<not> do regular expression character classes
e0c83546 1594such as C<\d> or C<[:lower:]>. The C<tr> operator is not equivalent to
cc255d5f
JH
1595the tr(1) utility. If you want to map strings between lower/upper
1596cases, see L<perlfunc/lc> and L<perlfunc/uc>, and in general consider
1597using the C<s> operator if you need regular expressions.
1598
JH
1599Note also that the whole range idea is rather unportable between
1600character sets--and even within character sets they may cause results
1601you probably didn't expect. A sound principle is to use only ranges
1602that begin from and end at either alphabets of equal case (a-e, A-E),
1603or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1604character sets in full.
1605
a0d0e21e
LW
1606Options:
1607
1608 c Complement the SEARCHLIST.
1609 d Delete found but unreplaced characters.
1610 s Squash duplicate replaced characters.
1611
19799a22
GS
1612If the C</c> modifier is specified, the SEARCHLIST character set
1613is complemented. If the C</d> modifier is specified, any characters
1614specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
1615(Note that this is slightly more flexible than the behavior of some
1616B<tr> programs, which delete anything they find in the SEARCHLIST,
1617period.) If the C</s> modifier is specified, sequences of characters
1618that were transliterated to the same character are squashed down
1619to a single instance of the character.
a0d0e21e
LW
1620
1621If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1622exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1623than the SEARCHLIST, the final character is replicated till it is long
5a964f20 1624enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
a0d0e21e
LW
1625This latter is useful for counting characters in a class or for
1626squashing character sequences in a class.
1627
1628Examples:
1629
1630 \$ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1631
1632 \$cnt = tr/*/*/; # count the stars in \$_
1633
1634 \$cnt = \$sky =~ tr/*/*/; # count the stars in \$sky
1635
1636 \$cnt = tr/0-9//; # count the digits in \$_
1637
1638 tr/a-zA-Z//s; # bookkeeper -> bokeper
1639
1640 (\$HOST = \$host) =~ tr/a-z/A-Z/;
1641
1642 tr/a-zA-Z/ /cs; # change non-alphas to single space
1643
1644 tr [\200-\377]
1645 [\000-\177]; # delete 8th bit
1646
19799a22
GS
1647If multiple transliterations are given for a character, only the
1648first one is used:
748a9306
LW
1649
1650 tr/AAA/XYZ/
1651
2c268ad5 1652will transliterate any A to X.
748a9306 1653
19799a22 1654Because the transliteration table is built at compile time, neither
a0d0e21e 1655the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
19799a22
GS
1656interpolation. That means that if you want to use variables, you
1657must use an eval():
a0d0e21e
LW
1658
1659 eval "tr/\$oldlist/\$newlist/";
1660 die \$@ if \$@;
1661
1662 eval "tr/\$oldlist/\$newlist/, 1" or die \$@;
1663
7e3b091d 1664=item <<EOF
d74e8afc 1665X<here-doc> X<heredoc> X<here-document> X<<< << >>>
7e3b091d
DA
1666
1667A line-oriented form of quoting is based on the shell "here-document"
1668syntax. Following a C<< << >> you specify a string to terminate
1669the quoted material, and all lines following the current line down to
89d205f2
YO
1670the terminating string are the value of the item.
1671
1672The terminating string may be either an identifier (a word), or some
1673quoted text. An unquoted identifier works like double quotes.
1674There may not be a space between the C<< << >> and the identifier,
1675unless the identifier is explicitly quoted. (If you put a space it
1676will be treated as a null identifier, which is valid, and matches the
1677first empty line.) The terminating string must appear by itself
1678(unquoted and with no surrounding whitespace) on the terminating line.
1679
1680If the terminating string is quoted, the type of quotes used determine
1681the treatment of the text.
1682
1683=over 4
1684
1685=item Double Quotes
1686
1687Double quotes indicate that the text will be interpolated using exactly
1688the same rules as normal double quoted strings.
7e3b091d
DA
1689
1690 print <<EOF;
1691 The price is \$Price.
1692 EOF
1693
1694 print << "EOF"; # same as above
1695 The price is \$Price.
1696 EOF
1697
89d205f2
YO
1698
1699=item Single Quotes
1700
1701Single quotes indicate the text is to be treated literally with no
1702interpolation of its content. This is similar to single quoted
1703strings except that backslashes have no special meaning, with C<\\>
1704being treated as two backslashes and not one as they would in every
1705other quoting construct.
1706
1707This is the only form of quoting in perl where there is no need
1708to worry about escaping content, something that code generators
1709can and do make good use of.
1710
1711=item Backticks
1712
1713The content of the here doc is treated just as it would be if the
1714string were embedded in backticks. Thus the content is interpolated
1715as though it were double quoted and then executed via the shell, with
1716the results of the execution returned.
1717
1718 print << `EOC`; # execute command and get results
7e3b091d 1719 echo hi there
7e3b091d
DA
1720 EOC
1721
89d205f2
YO
1722=back
1723
1724It is possible to stack multiple here-docs in a row:
1725
7e3b091d
DA
1726 print <<"foo", <<"bar"; # you can stack them
1727 I said foo.
1728 foo
1729 I said bar.
1730 bar
1731
1732 myfunc(<< "THIS", 23, <<'THAT');
1733 Here's a line
1734 or two.
1735 THIS
1736 and here's another.
1737 THAT
1738
1739Just don't forget that you have to put a semicolon on the end
1740to finish the statement, as Perl doesn't know you're not going to
1741try to do this:
1742
1743 print <<ABC
1744 179231
1745 ABC
1746 + 20;
1747
872d7e53
TS
1748If you want to remove the line terminator from your here-docs,
1749use C<chomp()>.
1750
1751 chomp(\$string = <<'END');
1752 This is a string.
1753 END
1754
1755If you want your here-docs to be indented with the rest of the code,
1756you'll need to remove leading whitespace from each line manually:
7e3b091d
DA
1757
1758 (\$quote = <<'FINIS') =~ s/^\s+//gm;
89d205f2 1759 The Road goes ever on and on,
7e3b091d
DA
1760 down from the door where it began.
1761 FINIS
1762
1763If you use a here-doc within a delimited construct, such as in C<s///eg>,
1764the quoted material must come on the lines following the final delimiter.
1766
1767 s/this/<<E . 'that'
1768 the other
1769 E
1770 . 'more '/eg;
1771
1772you have to write
1773
89d205f2
YO
1774 s/this/<<E . 'that'
1775 . 'more '/eg;
1776 the other
1777 E
7e3b091d
DA
1778
1779If the terminating identifier is on the last line of the program, you
1780must be sure there is a newline after it; otherwise, Perl will give the
1781warning B<Can't find string terminator "END" anywhere before EOF...>.
1782
89d205f2
YO
1783Additionally, the quoting rules for the end of string identifier are not
1784related to Perl's quoting rules -- C<q()>, C<qq()>, and the like are not
1785supported in place of C<''> and C<"">, and the only interpolation is for
1786backslashing the quoting character:
7e3b091d
DA
1787
1788 print << "abc\"def";
1789 testing...
1790 abc"def
1791
1792Finally, quoted strings cannot span multiple lines. The general rule is
1793that the identifier must be a string literal. Stick with that, and you
1794should be safe.
1795
a0d0e21e
LW
1796=back
1797
75e14d17 1798=head2 Gory details of parsing quoted constructs
d74e8afc 1799X<quote, gory details>
75e14d17 1800
19799a22
GS
1801When presented with something that might have several different
1802interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
1803principle to pick the most probable interpretation. This strategy
1804is so successful that Perl programmers often do not suspect the
1805ambivalence of what they write. But from time to time, Perl's
1806notions differ substantially from what the author honestly meant.
1807
1808This section hopes to clarify how Perl handles quoted constructs.
1809Although the most common reason to learn this is to unravel labyrinthine
1810regular expressions, because the initial steps of parsing are the
1811same for all quoting operators, they are all discussed together.
1812
1813The most important Perl parsing rule is the first one discussed
1814below: when processing a quoted construct, Perl first finds the end
1815of that construct, then interprets its contents. If you understand
1816this rule, you may skip the rest of this section on the first
1817reading. The other rules are likely to contradict the user's
1818expectations much less frequently than this first one.
1819
1820Some passes discussed below are performed concurrently, but because
1821their results are the same, we consider them individually. For different
1822quoting constructs, Perl performs different numbers of passes, from
6deea57f 1823one to four, but these passes are always performed in the same order.
75e14d17 1824
13a2d996 1825=over 4
75e14d17
IZ
1826
1827=item Finding the end
1828
6deea57f
TS
1829The first pass is finding the end of the quoted construct, where
1830the information about the delimiters is used in parsing.
1831During this search, text between the starting and ending delimiters
1832is copied to a safe location. The text copied gets delimiter-independent.
1833
1834If the construct is a here-doc, the ending delimiter is a line
1835that has a terminating string as the content. Therefore C<<<EOF> is
1836terminated by C<EOF> immediately followed by C<"\n"> and starting
1837from the first column of the terminating line.
1838When searching for the terminating line of a here-doc, nothing
1839is skipped. In other words, lines after the here-doc syntax
1840are compared with the terminating string line by line.
1841
1842For the constructs except here-docs, single characters are used as starting
1843and ending delimiters. If the starting delimiter is an opening punctuation
1844(that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
1845corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
1846If the starting delimiter is an unpaired character like C</> or a closing
1847punctuation, the ending delimiter is same as the starting delimiter.
1848Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
1849C<qq[]> and C<qq]]> constructs.
1850
1851When searching for single-character delimiters, escaped delimiters
1852and C<\\> are skipped. For example, while searching for terminating C</>,
1853combinations of C<\\> and C<\/> are skipped. If the delimiters are
1854bracketing, nested pairs are also skipped. For example, while searching
1855for closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
1856and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
1857However, when backslashes are used as the delimiters (like C<qq\\> and
1858C<tr\\\>), nothing is skipped.
1859During the search for the end, backslashes that escape delimiters
1860are removed (exactly speaking, they are not copied to the safe location).
75e14d17 1861
19799a22
GS
1862For constructs with three-part delimiters (C<s///>, C<y///>, and
1863C<tr///>), the search is repeated once more.
6deea57f
TS
1864If the first delimiter is not an opening punctuation, three delimiters must
1865be same such as C<s!!!> and C<tr)))>, in which case the second delimiter
1866terminates the left part and starts the right part at once.
1867If the left part is delimited by bracketing punctuations (that is C<()>,
1868C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
1869delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespaces
1870and comments are allowed between both parts, though the comment must follow
1871at least one whitespace; otherwise a character expected as the start of
1872the comment may be regarded as the starting delimiter of the right part.
75e14d17 1873
19799a22
GS
1874During this search no attention is paid to the semantics of the construct.
1875Thus:
75e14d17
IZ
1876
1877 "\$hash{"\$foo/\$bar"}"
1878
2a94b7ce 1879or:
75e14d17 1880
89d205f2 1881 m/
2a94b7ce 1882 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
1883 /x
1884
19799a22
GS
1885do not form legal quoted expressions. The quoted part ends on the
1886first C<"> and C</>, and the rest happens to be a syntax error.
1887Because the slash that terminated C<m//> was followed by a C<SPACE>,
1888the example above is not C<m//x>, but rather C<m//> with no C</x>
1889modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 1890
89d205f2
YO
1891Also no attention is paid to C<\c\> (multichar control char syntax) during
1892this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
1893of C<\/>, and the following C</> is not recognized as a delimiter.
0d594e51
TS
1894Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
1895
75e14d17 1896=item Interpolation
d74e8afc 1897X<interpolation>
75e14d17 1898
19799a22 1899The next step is interpolation in the text obtained, which is now
89d205f2 1900delimiter-independent. There are multiple cases.
75e14d17 1901
13a2d996 1902=over 4
75e14d17 1903
89d205f2 1904=item C<<<'EOF'>
75e14d17
IZ
1905
1906No interpolation is performed.
6deea57f
TS
1907Note that the combination C<\\> is left intact, since escaped delimiters
1908are not available for here-docs.
75e14d17 1909
6deea57f 1910=item C<m''>, the pattern of C<s'''>
89d205f2 1911
6deea57f
TS
1912No interpolation is performed at this stage.
1913Any backslashed sequences including C<\\> are treated at the stage
1914to L</"parsing regular expressions">.
89d205f2 1915
6deea57f 1916=item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
75e14d17 1917
89d205f2 1918The only interpolation is removal of C<\> from pairs of C<\\>.
6deea57f
TS
1919Therefore C<-> in C<tr'''> and C<y'''> is treated literally
1920as a hyphen and no character range is available.
1921C<\1> in the replacement of C<s'''> does not work as C<\$1>.
89d205f2
YO
1922
1923=item C<tr///>, C<y///>
1924
6deea57f
TS
1925No variable interpolation occurs. String modifying combinations for
1926case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
1927The other escape sequences such as C<\200> and C<\t> and backslashed
1928characters such as C<\\> and C<\-> are converted to appropriate literals.
89d205f2
YO
1929The character C<-> is treated specially and therefore C<\-> is treated
1930as a literal C<->.
75e14d17 1931
89d205f2 1932=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
75e14d17 1933
19799a22
GS
1934C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
1935converted to corresponding Perl constructs. Thus, C<"\$foo\Qbaz\$bar">
1936is converted to C<\$foo . (quotemeta("baz" . \$bar))> internally.
6deea57f
TS
1937The other escape sequences such as C<\200> and C<\t> and backslashed
1938characters such as C<\\> and C<\-> are replaced with appropriate
1939expansions.
2a94b7ce 1940
19799a22
GS
1941Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
1942is interpolated in the usual way. Something like C<"\Q\\E"> has
1943no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
1944result is the same as for C<"\\\\E">. As a general rule, backslashes
1945between C<\Q> and C<\E> may lead to counterintuitive results. So,
1946C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
1947as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
1948
1949 \$str = '\t';
1950 return "\Q\$str";
1951
1952may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
1953
19799a22 1954Interpolated scalars and arrays are converted internally to the C<join> and
92d29cee 1955C<.> catenation operations. Thus, C<"\$foo XXX '@arr'"> becomes:
75e14d17 1956
19799a22 1957 \$foo . " XXX '" . (join \$", @arr) . "'";
75e14d17 1958
19799a22 1959All operations above are performed simultaneously, left to right.
75e14d17 1960
19799a22
GS
1961Because the result of C<"\Q STRING \E"> has all metacharacters
1962quoted, there is no way to insert a literal C<\$> or C<@> inside a
1963C<\Q\E> pair. If protected by C<\>, C<\$> will be quoted to became
1964C<"\\\\$">; if not, it is interpreted as the start of an interpolated
1965scalar.
75e14d17 1966
19799a22 1967Note also that the interpolation code needs to make a decision on
89d205f2 1968where the interpolated scalar ends. For instance, whether
35f2feb0 1969C<< "a \$b -> {c}" >> really means:
75e14d17
IZ
1970
1971 "a " . \$b . " -> {c}";
1972
2a94b7ce 1973or:
75e14d17
IZ
1974
1975 "a " . \$b -> {c};
1976
19799a22
GS
1977Most of the time, the longest possible text that does not include
1978spaces between components and which contains matching braces or
1979brackets. because the outcome may be determined by voting based
1980on heuristic estimators, the result is not strictly predictable.
1981Fortunately, it's usually correct for ambiguous cases.
75e14d17 1982
6deea57f 1983=item the replacement of C<s///>
75e14d17 1984
19799a22 1985Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
6deea57f
TS
1986happens as with C<qq//> constructs.
1987
1988It is at this step that C<\1> is begrudgingly converted to C<\$1> in
1989the replacement text of C<s///>, in order to correct the incorrigible
1990I<sed> hackers who haven't picked up the saner idiom yet. A warning
1991is emitted if the C<use warnings> pragma or the B<-w> command-line flag
1992(that is, the C<\$^W> variable) was set.
1993
1994=item C<RE> in C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
1995
cc74c5bd
TS
1996Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\E>,
1997and interpolation happens (almost) as with C<qq//> constructs.
1998
1999However any other combinations of C<\> followed by a character
2000are not substituted but only skipped, in order to parse them
2001as regular expressions at the following step.
6deea57f 2002As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
1749ea0d 2003treated as an array symbol (for example C<@foo>),
6deea57f 2004even though the same text in C<qq//> gives interpolation of C<\c@>.
6deea57f
TS
2005
2006Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
19799a22
GS
2007a C<#>-comment in a C<//x>-regular expression, no processing is
2008performed whatsoever. This is the first step at which the presence
2009of the C<//x> modifier is relevant.
2010
1749ea0d
TS
2011Interpolation in patterns has several quirks: C<\$|>, C<\$(>, C<\$)>, C<@+>
2012and C<@-> are not interpolated, and constructs C<\$var[SOMETHING]> are
2013voted (by several different estimators) to be either an array element
2014or C<\$var> followed by an RE alternative. This is where the notation
19799a22
GS
2015C<\${arr[\$bar]}> comes handy: C</\${arr[0-9]}/> is interpreted as
2016array element C<-9>, not as a regular expression from the variable
2017C<\$arr> followed by a digit, which would be the interpretation of
2018C</\$arr[0-9]/>. Since voting among different estimators may occur,
2019the result is not predictable.
2020
19799a22
GS
2021The lack of processing of C<\\> creates specific restrictions on
2022the post-processed text. If the delimiter is C</>, one cannot get
2023the combination C<\/> into the result of this step. C</> will
2024finish the regular expression, C<\/> will be stripped to C</> on
2025the previous step, and C<\\/> will be left as is. Because C</> is
2026equivalent to C<\/> inside a regular expression, this does not
2027matter unless the delimiter happens to be character special to the
2028RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
2029alphanumeric char, as in:
2a94b7ce
IZ
2030
2031 m m ^ a \s* b mmx;
2032
19799a22 2033In the RE above, which is intentionally obfuscated for illustration, the
6deea57f 2034delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
89d205f2 2035RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
19799a22
GS
2036reason you're encouraged to restrict your delimiters to non-alphanumeric,
2037non-whitespace choices.
75e14d17
IZ
2038
2039=back
2040
19799a22 2041This step is the last one for all constructs except regular expressions,
75e14d17
IZ
2042which are processed further.
2043
6deea57f
TS
2044=item parsing regular expressions
2045X<regexp, parse>
75e14d17 2046
19799a22
GS
2047Previous steps were performed during the compilation of Perl code,
2048but this one happens at run time--although it may be optimized to
2049be calculated at compile time if appropriate. After preprocessing
6deea57f 2050described above, and possibly after evaluation if concatenation,
19799a22
GS
2051joining, casing translation, or metaquoting are involved, the
2052resulting I<string> is passed to the RE engine for compilation.
2053
2054Whatever happens in the RE engine might be better discussed in L<perlre>,
2055but for the sake of continuity, we shall do so here.
2056
2057This is another step where the presence of the C<//x> modifier is
2058relevant. The RE engine scans the string from left to right and
2059converts it to a finite automaton.
2060
2061Backslashed characters are either replaced with corresponding
2062literal strings (as with C<\{>), or else they generate special nodes
2063in the finite automaton (as with C<\b>). Characters special to the
2064RE engine (such as C<|>) generate corresponding nodes or groups of
2065nodes. C<(?#...)> comments are ignored. All the rest is either
2066converted to literal strings to match, or else is ignored (as is
2067whitespace and C<#>-style comments if C<//x> is present).
2068
2069Parsing of the bracketed character class construct, C<[...]>, is
2070rather different than the rule used for the rest of the pattern.
2071The terminator of this construct is found using the same rules as
2072for finding the terminator of a C<{}>-delimited construct, the only
2073exception being that C<]> immediately following C<[> is treated as
2074though preceded by a backslash. Similarly, the terminator of
2075C<(?{...})> is found using the same rules as for finding the
2076terminator of a C<{}>-delimited construct.
2077
2078It is possible to inspect both the string given to RE engine and the
2079resulting finite automaton. See the arguments C<debug>/C<debugcolor>
2080in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 2081switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
2082
2083=item Optimization of regular expressions
d74e8afc 2084X<regexp, optimization>
75e14d17 2085
7522fed5 2086This step is listed for completeness only. Since it does not change
75e14d17 2087semantics, details of this step are not documented and are subject
19799a22
GS
2088to change without notice. This step is performed over the finite
2089automaton that was generated during the previous pass.
2a94b7ce 2090
19799a22
GS
2091It is at this stage that C<split()> silently optimizes C</^/> to
2092mean C</^/m>.
75e14d17
IZ
2093
2094=back
2095
a0d0e21e 2096=head2 I/O Operators
d74e8afc
ITB
2097X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
2098X<< <> >> X<@ARGV>
a0d0e21e 2099
54310121 2100There are several I/O operators you should know about.
7b8d334a 2102A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
2103double-quote interpolation. It is then interpreted as an external
2104command, and the output of that command is the value of the
e9c56f9b
JH
2105backtick string, like in a shell. In scalar context, a single string
2106consisting of all output is returned. In list context, a list of
2107values is returned, one per line of output. (You can set C<\$/> to use
2108a different line terminator.) The command is executed each time the
2109pseudo-literal is evaluated. The status value of the command is
2110returned in C<\$?> (see L<perlvar> for the interpretation of C<\$?>).
2111Unlike in B<csh>, no translation is done on the return data--newlines
2112remain newlines. Unlike in any of the shells, single quotes do not
2113hide variable names in the command from interpretation. To pass a
2114literal dollar-sign through to the shell you need to hide it with a
2115backslash. The generalized form of backticks is C<qx//>. (Because
2116backticks always undergo shell expansion as well, see L<perlsec> for
2117security concerns.)
d74e8afc 2118X<qx> X<`> X<``> X<backtick> X<glob>
19799a22
GS
2119
2120In scalar context, evaluating a filehandle in angle brackets yields
2121the next line from that file (the newline, if any, included), or
2122C<undef> at end-of-file or on error. When C<\$/> is set to C<undef>
2123(sometimes known as file-slurp mode) and the file is empty, it
2124returns C<''> the first time, followed by C<undef> subsequently.
2125
2126Ordinarily you must assign the returned value to a variable, but
2127there is one situation where an automatic assignment happens. If
2128and only if the input symbol is the only thing inside the conditional
2129of a C<while> statement (even if disguised as a C<for(;;)> loop),
2130the value is automatically assigned to the global variable \$_,
2131destroying whatever was there previously. (This may seem like an
2132odd thing to you, but you'll use the construct in almost every Perl
17b829fa 2133script you write.) The \$_ variable is not implicitly localized.
19799a22
GS
2134You'll have to put a C<local \$_;> before the loop if you want that
2135to happen.
2136
2137The following lines are equivalent:
a0d0e21e 2138
748a9306 2139 while (defined(\$_ = <STDIN>)) { print; }
7b8d334a 2140 while (\$_ = <STDIN>) { print; }
a0d0e21e
LW
2141 while (<STDIN>) { print; }
2142 for (;<STDIN>;) { print; }
748a9306 2143 print while defined(\$_ = <STDIN>);
7b8d334a 2144 print while (\$_ = <STDIN>);
a0d0e21e
LW
2145 print while <STDIN>;
2146
19799a22 2147This also behaves similarly, but avoids \$_ :
7b8d334a 2148
89d205f2 2149 while (my \$line = <STDIN>) { print \$line }
7b8d334a 2150
19799a22
GS
2151In these loop constructs, the assigned value (whether assignment
2152is automatic or explicit) is then tested to see whether it is
2153defined. The defined test avoids problems where line has a string
2154value that would be treated as false by Perl, for example a "" or
2155a "0" with no trailing newline. If you really mean for such values
2156to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
2157
2158 while ((\$_ = <STDIN>) ne '0') { ... }
2159 while (<STDIN>) { last unless \$_; ... }
2160
35f2feb0 2161In other boolean contexts, C<< <I<filehandle>> >> without an
89d205f2 2162explicit C<defined> test or comparison elicit a warning if the
9f1b1f2d 2163C<use warnings> pragma or the B<-w>
19799a22 2164command-line switch (the C<\$^W> variable) is in effect.
7b8d334a 2165
5f05dabc 2166The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
2167filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
2168in packages, where they would be interpreted as local identifiers
2169rather than global.) Additional filehandles may be created with
2170the open() function, amongst others. See L<perlopentut> and
2171L<perlfunc/open> for details on this.
d74e8afc 2172X<stdin> X<stdout> X<sterr>
a0d0e21e 2173
35f2feb0 2174If a <FILEHANDLE> is used in a context that is looking for
19799a22
GS
2175a list, a list comprising all input lines is returned, one line per
2176list element. It's easy to grow to a rather large data space this
2177way, so use with care.
a0d0e21e 2178
35f2feb0 2179<FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
35f2feb0
GS
2182The null filehandle <> is special: it can be used to emulate the
2183behavior of B<sed> and B<awk>. Input from <> comes either from
a0d0e21e 2184standard input, or from each file listed on the command line. Here's
35f2feb0 2185how it works: the first time <> is evaluated, the @ARGV array is
5a964f20 2186checked, and if it is empty, C<\$ARGV[0]> is set to "-", which when opened
a0d0e21e
LW
2187gives you standard input. The @ARGV array is then processed as a list
2188of filenames. The loop
2189
2190 while (<>) {
2191 ... # code for each line
2192 }
2193
2194is equivalent to the following Perl-like pseudo code:
2195
3e3baf6d 2196 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
2197 while (\$ARGV = shift) {
2198 open(ARGV, \$ARGV);
2199 while (<ARGV>) {
2200 ... # code for each line
2201 }
2202 }
2203
19799a22
GS
2204except that it isn't so cumbersome to say, and will actually work.
2205It really does shift the @ARGV array and put the current filename
2206into the \$ARGV variable. It also uses filehandle I<ARGV>
35f2feb0 2207internally--<> is just a synonym for <ARGV>, which
19799a22 2208is magical. (The pseudo code above doesn't work because it treats
35f2feb0 2209<ARGV> as non-magical.)
a0d0e21e 2210
35f2feb0 2211You can modify @ARGV before the first <> as long as the array ends up
a0d0e21e 2212containing the list of filenames you really want. Line numbers (C<\$.>)
19799a22
GS
2213continue as though the input were one big happy file. See the example
2214in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20 2215
89d205f2 2216If you want to set @ARGV to your own list of files, go right ahead.
5a964f20
TC
2217This sets @ARGV to all plain text files if no @ARGV was given:
2218
2219 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 2220
5a964f20
TC
2221You can even set them to pipe commands. For example, this automatically
2222filters compressed arguments through B<gzip>:
2223
2224 @ARGV = map { /\.(gz|Z)\$/ ? "gzip -dc < \$_ |" : \$_ } @ARGV;
2225
2226If you want to pass switches into your script, you can use one of the
a0d0e21e
LW
2227Getopts modules or put a loop on the front like this:
2228
2229 while (\$_ = \$ARGV[0], /^-/) {
2230 shift;
2231 last if /^--\$/;
2232 if (/^-D(.*)/) { \$debug = \$1 }
2233 if (/^-v/) { \$verbose++ }
5a964f20 2234 # ... # other switches
a0d0e21e 2235 }
5a964f20 2236
a0d0e21e 2237 while (<>) {
5a964f20 2238 # ... # code for each line
a0d0e21e
LW
2239 }
2240
89d205f2
YO
2241The <> symbol will return C<undef> for end-of-file only once.
2242If you call it again after this, it will assume you are processing another
19799a22 2243@ARGV list, and if you haven't set @ARGV, will read input from STDIN.
a0d0e21e 2244
b159ebd3 2245If what the angle brackets contain is a simple scalar variable (e.g.,
35f2feb0 2246<\$foo>), then that variable contains the name of the
19799a22
GS
2247filehandle to input from, or its typeglob, or a reference to the
2248same. For example:
cb1a09d0
2249
2250 \$fh = \*STDIN;
2251 \$line = <\$fh>;
a0d0e21e 2252
5a964f20
TC
2253If what's within the angle brackets is neither a filehandle nor a simple
2254scalar variable containing a filehandle name, typeglob, or typeglob
2255reference, it is interpreted as a filename pattern to be globbed, and
2256either a list of filenames or the next filename in the list is returned,
19799a22 2257depending on context. This distinction is determined on syntactic
35f2feb0
GS
2258grounds alone. That means C<< <\$x> >> is always a readline() from
2259an indirect handle, but C<< <\$hash{key}> >> is always a glob().
5a964f20 2260That's because \$x is a simple scalar variable, but C<\$hash{key}> is
ef191992
YST
2261not--it's a hash element. Even C<< <\$x > >> (note the extra space)
2262is treated as C<glob("\$x ")>, not C<readline(\$x)>.
5a964f20
TC
2263
2264One level of double-quote interpretation is done first, but you can't
35f2feb0 2265say C<< <\$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
2266in the previous paragraph. (In older versions of Perl, programmers
2267would insert curly brackets to force interpretation as a filename glob:
35f2feb0 2268C<< <\${foo}> >>. These days, it's considered cleaner to call the
5a964f20 2269internal function directly as C<glob(\$foo)>, which is probably the right
19799a22 2270way to have done it in the first place.) For example:
a0d0e21e
LW
2271
2272 while (<*.c>) {
2273 chmod 0644, \$_;
2274 }
2275
3a4b19e4 2276is roughly equivalent to:
a0d0e21e
LW
2277
2278 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
2279 while (<FOO>) {
5b3eff12 2280 chomp;
a0d0e21e
LW
2281 chmod 0644, \$_;
2282 }
2283
3a4b19e4
GS
2284except that the globbing is actually done internally using the standard
2285C<File::Glob> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
2286
2287 chmod 0644, <*.c>;
2288
19799a22
GS
2289A (file)glob evaluates its (embedded) argument only when it is
2290starting a new list. All values must be read before it will start
2291over. In list context, this isn't important because you automatically
2292get them all anyway. However, in scalar context the operator returns
069e01df 2293the next value each time it's called, or C<undef> when the list has
19799a22
GS
2294run out. As with filehandle reads, an automatic C<defined> is
2295generated when the glob occurs in the test part of a C<while>,
2296because legal glob returns (e.g. a file called F<0>) would otherwise
2297terminate the loop. Again, C<undef> is returned only once. So if
2298you're expecting a single value from a glob, it is much better to
2299say
4633a7c4
LW
2300
2301 (\$file) = <blurch*>;
2302
2303than
2304
2305 \$file = <blurch*>;
2306
2307because the latter will alternate between returning a filename and
19799a22 2308returning false.
4633a7c4 2309
b159ebd3 2310If you're trying to do variable interpolation, it's definitely better
4633a7c4 2311to use the glob() function, because the older notation can cause people
e37d713d 2312to become confused with the indirect filehandle notation.
4633a7c4
LW
2313
2314 @files = glob("\$dir/*.[ch]");
2315 @files = glob(\$files[\$i]);
2316
a0d0e21e 2317=head2 Constant Folding
d74e8afc 2318X<constant folding> X<folding>
a0d0e21e
LW
2319
2320Like C, Perl does a certain amount of expression evaluation at
19799a22 2321compile time whenever it determines that all arguments to an
a0d0e21e
LW
2322operator are static and have no side effects. In particular, string
2323concatenation happens at compile time between literals that don't do
19799a22 2324variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
2325compile time. You can say
2326
2327 'Now is the time for all' . "\n" .
2328 'good men to come to.'
2329
54310121 2330and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
2331you say
2332
2333 foreach \$file (@filenames) {
5a964f20 2334 if (-s \$file > 5 + 100 * 2**16) { }
54310121 2335 }
a0d0e21e 2336
19799a22
GS
2337the compiler will precompute the number which that expression
2338represents so that the interpreter won't have to.
a0d0e21e 2339
d74e8afc 2341X<no-op> X<nop>
fd1abbef
DN
2342
2343Perl doesn't officially have a no-op operator, but the bare constants
2344C<0> and C<1> are special-cased to not produce a warning in a void
2345context, so you can for example safely do
2346
2347 1 while foo();
2348
d74e8afc 2350X<operator, bitwise, string>
TP
2351
2352Bitstrings of any size may be manipulated by the bitwise operators
2353(C<~ | & ^>).
2354
19799a22
GS
2355If the operands to a binary bitwise op are strings of different
2356sizes, B<|> and B<^> ops act as though the shorter operand had
2357additional zero bits on the right, while the B<&> op acts as though
2358the longer operand were truncated to the length of the shorter.
2359The granularity for such extension or truncation is one or more
2360bytes.
89d205f2 2362 # ASCII-based examples
TP
2363 print "j p \n" ^ " a h"; # prints "JAPH\n"
2364 print "JA" | " ph\n"; # prints "japh\n"
2365 print "japh\nJunk" & '_____'; # prints "JAPH\n";
2366 print 'p N\$' ^ " E<H\n"; # prints "Perl\n";
2367
19799a22 2368If you are intending to manipulate bitstrings, be certain that
2c268ad5 2369you're supplying bitstrings: If an operand is a number, that will imply
19799a22 2370a B<numeric> bitwise operation. You may explicitly show which type of
TP
2371operation you intend by using C<""> or C<0+>, as in the examples below.
2372
4358a253
SS
2373 \$foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
2374 \$foo = '150' | 105; # yields 255
TP
2375 \$foo = 150 | '105'; # yields 255
2376 \$foo = '150' | '105'; # yields string '155' (under ASCII)
2377
2378 \$baz = 0+\$foo & 0+\$bar; # both ops explicitly numeric
2379 \$biz = "\$foo" ^ "\$bar"; # both ops explicitly stringy
a0d0e21e 2380
1ae175c8
GS
2381See L<perlfunc/vec> for information on how to manipulate individual bits
2382in a bit vector.
2383
55497cff 2384=head2 Integer Arithmetic
d74e8afc 2385X<integer>
a0d0e21e 2386
19799a22 2387By default, Perl assumes that it must do most of its arithmetic in
a0d0e21e
LW
2388floating point. But by saying
2389
2390 use integer;
2391
2392you may tell the compiler that it's okay to use integer operations
19799a22
GS
2393(if it feels like it) from here to the end of the enclosing BLOCK.
2394An inner BLOCK may countermand this by saying
a0d0e21e
LW
2395
2396 no integer;
2397
19799a22
GS
2398which lasts until the end of that BLOCK. Note that this doesn't
2399mean everything is only an integer, merely that Perl may use integer
2400operations if it is so inclined. For example, even under C<use
2401integer>, if you take the C<sqrt(2)>, you'll still get C<1.4142135623731>
2402or so.
2403
2404Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
89d205f2 2405and ">>") always produce integral results. (But see also
13a2d996 2406L<Bitwise String Operators>.) However, C<use integer> still has meaning for
19799a22
GS
2407them. By default, their results are interpreted as unsigned integers, but
2408if C<use integer> is in effect, their results are interpreted
2409as signed integers. For example, C<~0> usually evaluates to a large
0be96356 2410integral value. However, C<use integer; ~0> is C<-1> on two's-complement
19799a22 2411machines.
68dc0745 2412
d74e8afc 2414X<floating-point> X<floating point> X<float> X<real>
68dc0745 2415
2416While C<use integer> provides integer-only arithmetic, there is no
19799a22
GS
2417analogous mechanism to provide automatic rounding or truncation to a
2418certain number of decimal places. For rounding to a certain number
2419of digits, sprintf() or printf() is usually the easiest route.
2420See L<perlfaq4>.
68dc0745 2421
5a964f20
TC
2422Floating-point numbers are only approximations to what a mathematician
2423would call real numbers. There are infinitely more reals than floats,
2424so some corners must be cut. For example:
2425
2426 printf "%.20g\n", 123456789123456789;
2427 # produces 123456789123456784
2428
2429Testing for exact equality of floating-point equality or inequality is
2430not a good idea. Here's a (relatively expensive) work-around to compare
2431whether two floating-point numbers are equal to a particular number of
2432decimal places. See Knuth, volume II, for a more robust treatment of
2433this topic.
2434
2435 sub fp_equal {
2436 my (\$X, \$Y, \$POINTS) = @_;
2437 my (\$tX, \$tY);
2438 \$tX = sprintf("%.\${POINTS}g", \$X);
2439 \$tY = sprintf("%.\${POINTS}g", \$Y);
2440 return \$tX eq \$tY;
2441 }
2442
68dc0745 2443The POSIX module (part of the standard perl distribution) implements
19799a22
GS
2444ceil(), floor(), and other mathematical and trigonometric functions.
2445The Math::Complex module (part of the standard perl distribution)
2446defines mathematical functions that work on both the reals and the
2447imaginary numbers. Math::Complex not as efficient as POSIX, but
68dc0745 2448POSIX can't work with complex numbers.
2449
2450Rounding in financial applications can have serious implications, and
2451the rounding method used should be specified precisely. In these
2452cases, it probably pays not to trust whichever system rounding is
2453being used by Perl, but to instead implement the rounding function you
2454need yourself.
5a964f20
TC
2455
d74e8afc 2457X<number, arbitrary precision>
5a964f20
TC
2458
2459The standard Math::BigInt and Math::BigFloat modules provide
19799a22 2460variable-precision arithmetic and overloaded operators, although
cd5c4fce 2461they're currently pretty slow. At the cost of some space and
19799a22
GS
2462considerable speed, they avoid the normal pitfalls associated with
2463limited-precision representations.
5a964f20
TC
2464
2465 use Math::BigInt;
2466 \$x = Math::BigInt->new('123456789123456789');
2467 print \$x * \$x;
2468
2469 # prints +15241578780673678515622620750190521
19799a22 2470
cd5c4fce
T
2471There are several modules that let you calculate with (bound only by
2472memory and cpu-time) unlimited or fixed precision. There are also
2473some non-standard modules that provide faster implementations via
2474external C libraries.
2475
2476Here is a short, but incomplete summary:
2477
2478 Math::Fraction big, unlimited fractions like 9973 / 12967
2479 Math::String treat string sequences like numbers
2480 Math::FixedPrecision calculate with a fixed precision
2481 Math::Currency for currency calculations
2482 Bit::Vector manipulate bit vectors fast (uses C)
2483 Math::BigIntFast Bit::Vector wrapper for big numbers
2484 Math::Pari provides access to the Pari C library
2485 Math::BigInteger uses an external C library
2486 Math::Cephes uses external Cephes C library (no big numbers)
2487 Math::Cephes::Fraction fractions via the Cephes library
2488 Math::GMP another one using an external C library
2489
2490Choose wisely.
16070b82
GS
2491
2492=cut