This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
(perl #127780) point backtick users at the open pragma
[perl5.git] / pod / perlop.pod
CommitLineData
a0d0e21e 1=head1 NAME
d74e8afc 2X<operator>
a0d0e21e
LW
3
4perlop - Perl operators and precedence
5
d042e63d
MS
6=head1 DESCRIPTION
7
ae3f7391 8In Perl, the operator determines what operation is performed,
ba7f043c 9independent of the type of the operands. For example S<C<$x + $y>>
db691027 10is always a numeric addition, and if C<$x> or C<$y> do not contain
ae3f7391
ML
11numbers, an attempt is made to convert them to numbers first.
12
13This is in contrast to many other dynamic languages, where the
46f8a5ea 14operation is determined by the type of the first argument. It also
ae3f7391 15means that Perl has two versions of some operators, one for numeric
ba7f043c
KW
16and one for string comparison. For example S<C<$x == $y>> compares
17two numbers for equality, and S<C<$x eq $y>> compares two strings.
ae3f7391
ML
18
19There are a few exceptions though: C<x> can be either string
20repetition or list repetition, depending on the type of the left
0b55efd7 21operand, and C<&>, C<|>, C<^> and C<~> can be either string or numeric bit
ae3f7391
ML
22operations.
23
89d205f2 24=head2 Operator Precedence and Associativity
d74e8afc 25X<operator, precedence> X<precedence> X<associativity>
d042e63d
MS
26
27Operator precedence and associativity work in Perl more or less like
28they do in mathematics.
29
30I<Operator precedence> means some operators are evaluated before
ba7f043c
KW
31others. For example, in S<C<2 + 4 * 5>>, the multiplication has higher
32precedence so S<C<4 * 5>> is evaluated first yielding S<C<2 + 20 ==
3322>> and not S<C<6 * 5 == 30>>.
d042e63d
MS
34
35I<Operator associativity> defines what happens if a sequence of the
36same operators is used one after another: whether the evaluator will
ba7f043c
KW
37evaluate the left operations first, or the right first. For example, in
38S<C<8 - 4 - 2>>, subtraction is left associative so Perl evaluates the
39expression left to right. S<C<8 - 4>> is evaluated first making the
40expression S<C<4 - 2 == 2>> and not S<C<8 - 2 == 6>>.
a0d0e21e
LW
41
42Perl operators have the following associativity and precedence,
19799a22
GS
43listed from highest precedence to lowest. Operators borrowed from
44C keep the same precedence relationship with each other, even where
45C's precedence is slightly screwy. (This makes learning Perl easier
46for C folks.) With very few exceptions, these all operate on scalar
47values only, not array values.
a0d0e21e
LW
48
49 left terms and list operators (leftward)
50 left ->
51 nonassoc ++ --
52 right **
53 right ! ~ \ and unary + and -
54310121 54 left =~ !~
a0d0e21e
LW
55 left * / % x
56 left + - .
57 left << >>
58 nonassoc named unary operators
59 nonassoc < > <= >= lt gt le ge
0d863452 60 nonassoc == != <=> eq ne cmp ~~
a0d0e21e
LW
61 left &
62 left | ^
63 left &&
c963b151 64 left || //
137443ea 65 nonassoc .. ...
a0d0e21e 66 right ?:
2ba1f20a 67 right = += -= *= etc. goto last next redo dump
a0d0e21e
LW
68 left , =>
69 nonassoc list operators (rightward)
a5f75d66 70 right not
a0d0e21e 71 left and
f23102e2 72 left or xor
a0d0e21e
LW
73
74In the following sections, these operators are covered in precedence order.
75
5a964f20
TC
76Many operators can be overloaded for objects. See L<overload>.
77
a0d0e21e 78=head2 Terms and List Operators (Leftward)
d74e8afc 79X<list operator> X<operator, list> X<term>
a0d0e21e 80
62c18ce2 81A TERM has the highest precedence in Perl. They include variables,
5f05dabc 82quote and quote-like operators, any expression in parentheses,
a0d0e21e
LW
83and any function whose arguments are parenthesized. Actually, there
84aren't really functions in this sense, just list operators and unary
85operators behaving as functions because you put parentheses around
86the arguments. These are all documented in L<perlfunc>.
87
ba7f043c 88If any list operator (C<print()>, etc.) or any unary operator (C<chdir()>, etc.)
a0d0e21e
LW
89is followed by a left parenthesis as the next token, the operator and
90arguments within parentheses are taken to be of highest precedence,
91just like a normal function call.
92
93In the absence of parentheses, the precedence of list operators such as
94C<print>, C<sort>, or C<chmod> is either very high or very low depending on
54310121 95whether you are looking at the left side or the right side of the operator.
a0d0e21e
LW
96For example, in
97
98 @ary = (1, 3, sort 4, 2);
99 print @ary; # prints 1324
100
ba7f043c 101the commas on the right of the C<sort> are evaluated before the C<sort>,
19799a22
GS
102but the commas on the left are evaluated after. In other words,
103list operators tend to gobble up all arguments that follow, and
a0d0e21e 104then act like a simple TERM with regard to the preceding expression.
19799a22 105Be careful with parentheses:
a0d0e21e
LW
106
107 # These evaluate exit before doing the print:
108 print($foo, exit); # Obviously not what you want.
109 print $foo, exit; # Nor is this.
110
111 # These do the print before evaluating exit:
112 (print $foo), exit; # This is what you want.
113 print($foo), exit; # Or this.
114 print ($foo), exit; # Or even this.
115
116Also note that
117
118 print ($foo & 255) + 1, "\n";
119
d042e63d
MS
120probably doesn't do what you expect at first glance. The parentheses
121enclose the argument list for C<print> which is evaluated (printing
ba7f043c 122the result of S<C<$foo & 255>>). Then one is added to the return value
d042e63d
MS
123of C<print> (usually 1). The result is something like this:
124
125 1 + 1, "\n"; # Obviously not what you meant.
126
127To do what you meant properly, you must write:
128
129 print(($foo & 255) + 1, "\n");
130
131See L<Named Unary Operators> for more discussion of this.
a0d0e21e 132
ba7f043c 133Also parsed as terms are the S<C<do {}>> and S<C<eval {}>> constructs, as
54310121 134well as subroutine and method calls, and the anonymous
a0d0e21e
LW
135constructors C<[]> and C<{}>.
136
2ae324a7 137See also L<Quote and Quote-like Operators> toward the end of this section,
da87341d 138as well as L</"I/O Operators">.
a0d0e21e
LW
139
140=head2 The Arrow Operator
d74e8afc 141X<arrow> X<dereference> X<< -> >>
a0d0e21e 142
35f2feb0 143"C<< -> >>" is an infix dereference operator, just as it is in C
19799a22
GS
144and C++. If the right side is either a C<[...]>, C<{...}>, or a
145C<(...)> subscript, then the left side must be either a hard or
146symbolic reference to an array, a hash, or a subroutine respectively.
147(Or technically speaking, a location capable of holding a hard
148reference, if it's an array or hash reference being used for
149assignment.) See L<perlreftut> and L<perlref>.
a0d0e21e 150
19799a22
GS
151Otherwise, the right side is a method name or a simple scalar
152variable containing either the method name or a subroutine reference,
153and the left side must be either an object (a blessed reference)
154or a class name (that is, a package name). See L<perlobj>.
a0d0e21e 155
821361b6 156The dereferencing cases (as opposed to method-calling cases) are
2ad792cd 157somewhat extended by the C<postderef> feature. For the
821361b6
RS
158details of that feature, consult L<perlref/Postfix Dereference Syntax>.
159
5f05dabc 160=head2 Auto-increment and Auto-decrement
d74e8afc 161X<increment> X<auto-increment> X<++> X<decrement> X<auto-decrement> X<-->
a0d0e21e 162
ba7f043c 163C<"++"> and C<"--"> work as in C. That is, if placed before a variable,
d042e63d
MS
164they increment or decrement the variable by one before returning the
165value, and if placed after, increment or decrement after returning the
166value.
167
168 $i = 0; $j = 0;
169 print $i++; # prints 0
170 print ++$j; # prints 1
a0d0e21e 171
b033823e 172Note that just as in C, Perl doesn't define B<when> the variable is
46f8a5ea
FC
173incremented or decremented. You just know it will be done sometime
174before or after the value is returned. This also means that modifying
c543c01b 175a variable twice in the same statement will lead to undefined behavior.
b033823e
A
176Avoid statements like:
177
178 $i = $i ++;
179 print ++ $i + $i ++;
180
181Perl will not guarantee what the result of the above statements is.
182
54310121 183The auto-increment operator has a little extra builtin magic to it. If
a0d0e21e
LW
184you increment a variable that is numeric, or that has ever been used in
185a numeric context, you get a normal increment. If, however, the
5f05dabc 186variable has been used in only string contexts since it was set, and
5a964f20 187has a value that is not the empty string and matches the pattern
9c0670e1 188C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
a0d0e21e
LW
189character within its range, with carry:
190
c543c01b
TC
191 print ++($foo = "99"); # prints "100"
192 print ++($foo = "a0"); # prints "a1"
193 print ++($foo = "Az"); # prints "Ba"
194 print ++($foo = "zz"); # prints "aaa"
a0d0e21e 195
6a61d433
HS
196C<undef> is always treated as numeric, and in particular is changed
197to C<0> before incrementing (so that a post-increment of an undef value
198will return C<0> rather than C<undef>).
199
5f05dabc 200The auto-decrement operator is not magical.
a0d0e21e
LW
201
202=head2 Exponentiation
d74e8afc 203X<**> X<exponentiation> X<power>
a0d0e21e 204
ba7f043c
KW
205Binary C<"**"> is the exponentiation operator. It binds even more
206tightly than unary minus, so C<-2**4> is C<-(2**4)>, not C<(-2)**4>.
207(This is
208implemented using C's C<pow(3)> function, which actually works on doubles
cb1a09d0 209internally.)
a0d0e21e 210
44a465b3
JH
211Note that certain exponentiation expressions are ill-defined:
212these include C<0**0>, C<1**Inf>, and C<Inf**0>. Do not expect
213any particular results from these special cases, the results
214are platform-dependent.
215
a0d0e21e 216=head2 Symbolic Unary Operators
d74e8afc 217X<unary operator> X<operator, unary>
a0d0e21e 218
ba7f043c 219Unary C<"!"> performs logical negation, that is, "not". See also C<not> for a lower
a0d0e21e 220precedence version of this.
d74e8afc 221X<!>
a0d0e21e 222
ba7f043c 223Unary C<"-"> performs arithmetic negation if the operand is numeric,
da2f94c5
FC
224including any string that looks like a number. If the operand is
225an identifier, a string consisting of a minus sign concatenated
226with the identifier is returned. Otherwise, if the string starts
227with a plus or minus, a string starting with the opposite sign is
ba7f043c
KW
228returned. One effect of these rules is that C<-bareword> is equivalent
229to the string C<"-bareword">. If, however, the string begins with a
230non-alphabetic character (excluding C<"+"> or C<"-">), Perl will attempt
231to convert
232the string to a numeric, and the arithmetic negation is performed. If the
06705523
SP
233string cannot be cleanly converted to a numeric, Perl will give the warning
234B<Argument "the string" isn't numeric in negation (-) at ...>.
d74e8afc 235X<-> X<negation, arithmetic>
a0d0e21e 236
ba7f043c
KW
237Unary C<"~"> performs bitwise negation, that is, 1's complement. For
238example, S<C<0666 & ~027>> is 0640. (See also L<Integer Arithmetic> and
972b05a9 239L<Bitwise String Operators>.) Note that the width of the result is
ba7f043c 240platform-dependent: C<~0> is 32 bits wide on a 32-bit platform, but 64
972b05a9 241bits wide on a 64-bit platform, so if you are expecting a certain bit
ba7f043c 242width, remember to use the C<"&"> operator to mask off the excess bits.
d74e8afc 243X<~> X<negation, binary>
a0d0e21e 244
f113cf86
TC
245When complementing strings, if all characters have ordinal values under
246256, then their complements will, also. But if they do not, all
247characters will be in either 32- or 64-bit complements, depending on your
248architecture. So for example, C<~"\x{3B1}"> is C<"\x{FFFF_FC4E}"> on
24932-bit machines and C<"\x{FFFF_FFFF_FFFF_FC4E}"> on 64-bit machines.
250
ba7f043c
KW
251If the experimental "bitwise" feature is enabled via S<C<use feature
252'bitwise'>>, then unary C<"~"> always treats its argument as a number, and an
253alternate form of the operator, C<"~.">, always treats its argument as a
fb7054ba
FC
254string. So C<~0> and C<~"0"> will both give 2**32-1 on 32-bit platforms,
255whereas C<~.0> and C<~."0"> will both yield C<"\xff">. This feature
ba7f043c 256produces a warning unless you use S<C<no warnings 'experimental::bitwise'>>.
fb7054ba 257
ba7f043c 258Unary C<"+"> has no effect whatsoever, even on strings. It is useful
a0d0e21e
LW
259syntactically for separating a function name from a parenthesized expression
260that would otherwise be interpreted as the complete list of function
a95b3d6a 261arguments. (See examples above under L</Terms and List Operators (Leftward)>.)
d74e8afc 262X<+>
a0d0e21e 263
ba7f043c 264Unary C<"\"> creates a reference to whatever follows it. See L<perlreftut>
19799a22
GS
265and L<perlref>. Do not confuse this behavior with the behavior of
266backslash within a string, although both forms do convey the notion
267of protecting the next thing from interpolation.
d74e8afc 268X<\> X<reference> X<backslash>
a0d0e21e
LW
269
270=head2 Binding Operators
d74e8afc 271X<binding> X<operator, binding> X<=~> X<!~>
a0d0e21e 272
ba7f043c
KW
273Binary C<"=~"> binds a scalar expression to a pattern match. Certain operations
274search or modify the string C<$_> by default. This operator makes that kind
cb1a09d0 275of operation work on some other string. The right argument is a search
2c268ad5
TP
276pattern, substitution, or transliteration. The left argument is what is
277supposed to be searched, substituted, or transliterated instead of the default
ba7f043c
KW
278C<$_>. When used in scalar context, the return value generally indicates the
279success of the operation. The exceptions are substitution (C<s///>)
280and transliteration (C<y///>) with the C</r> (non-destructive) option,
8ff32507
FC
281which cause the B<r>eturn value to be the result of the substitution.
282Behavior in list context depends on the particular operator.
000c65fc
DG
283See L</"Regexp Quote-Like Operators"> for details and L<perlretut> for
284examples using these operators.
f8bab1e9
GS
285
286If the right argument is an expression rather than a search pattern,
2c268ad5 287substitution, or transliteration, it is interpreted as a search pattern at run
46f8a5ea
FC
288time. Note that this means that its
289contents will be interpolated twice, so
89d205f2 290
1ca345ed 291 '\\' =~ q'\\';
89d205f2
YO
292
293is not ok, as the regex engine will end up trying to compile the
294pattern C<\>, which it will consider a syntax error.
a0d0e21e 295
ba7f043c 296Binary C<"!~"> is just like C<"=~"> except the return value is negated in
a0d0e21e
LW
297the logical sense.
298
ba7f043c
KW
299Binary C<"!~"> with a non-destructive substitution (C<s///r>) or transliteration
300(C<y///r>) is a syntax error.
4f4d7508 301
a0d0e21e 302=head2 Multiplicative Operators
d74e8afc 303X<operator, multiplicative>
a0d0e21e 304
ba7f043c 305Binary C<"*"> multiplies two numbers.
d74e8afc 306X<*>
a0d0e21e 307
ba7f043c 308Binary C<"/"> divides two numbers.
d74e8afc 309X</> X<slash>
a0d0e21e 310
ba7f043c 311Binary C<"%"> is the modulo operator, which computes the division
f7918450
KW
312remainder of its first argument with respect to its second argument.
313Given integer
ba7f043c 314operands C<$m> and C<$n>: If C<$n> is positive, then S<C<$m % $n>> is
db691027 315C<$m> minus the largest multiple of C<$n> less than or equal to
ba7f043c 316C<$m>. If C<$n> is negative, then S<C<$m % $n>> is C<$m> minus the
db691027 317smallest multiple of C<$n> that is not less than C<$m> (that is, the
89b4f0ad 318result will be less than or equal to zero). If the operands
db691027 319C<$m> and C<$n> are floating point values and the absolute value of
ba7f043c 320C<$n> (that is C<abs($n)>) is less than S<C<(UV_MAX + 1)>>, only
db691027 321the integer portion of C<$m> and C<$n> will be used in the operation
4848a83b 322(Note: here C<UV_MAX> means the maximum of the unsigned integer type).
db691027 323If the absolute value of the right operand (C<abs($n)>) is greater than
ba7f043c
KW
324or equal to S<C<(UV_MAX + 1)>>, C<"%"> computes the floating-point remainder
325C<$r> in the equation S<C<($r = $m - $i*$n)>> where C<$i> is a certain
f7918450 326integer that makes C<$r> have the same sign as the right operand
db691027
SF
327C<$n> (B<not> as the left operand C<$m> like C function C<fmod()>)
328and the absolute value less than that of C<$n>.
ba7f043c 329Note that when S<C<use integer>> is in scope, C<"%"> gives you direct access
f7918450 330to the modulo operator as implemented by your C compiler. This
55d729e4
GS
331operator is not as well defined for negative operands, but it will
332execute faster.
f7918450 333X<%> X<remainder> X<modulo> X<mod>
55d729e4 334
ba7f043c 335Binary C<"x"> is the repetition operator. In scalar context or if the left
62d10b70
GS
336operand is not enclosed in parentheses, it returns a string consisting
337of the left operand repeated the number of times specified by the right
338operand. In list context, if the left operand is enclosed in
ba7f043c 339parentheses or is a list formed by C<qw/I<STRING>/>, it repeats the list.
31201a8e
KW
340If the right operand is zero or negative (raising a warning on
341negative), it returns an empty string
3585017f 342or an empty list, depending on the context.
d74e8afc 343X<x>
a0d0e21e
LW
344
345 print '-' x 80; # print row of dashes
346
347 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
348
349 @ones = (1) x 80; # a list of 80 1's
350 @ones = (5) x @ones; # set all elements to 5
351
352
353=head2 Additive Operators
d74e8afc 354X<operator, additive>
a0d0e21e 355
ba7f043c 356Binary C<"+"> returns the sum of two numbers.
d74e8afc 357X<+>
a0d0e21e 358
ba7f043c 359Binary C<"-"> returns the difference of two numbers.
d74e8afc 360X<->
a0d0e21e 361
ba7f043c 362Binary C<"."> concatenates two strings.
d74e8afc
ITB
363X<string, concatenation> X<concatenation>
364X<cat> X<concat> X<concatenate> X<.>
a0d0e21e
LW
365
366=head2 Shift Operators
d74e8afc
ITB
367X<shift operator> X<operator, shift> X<<< << >>>
368X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
369X<shl> X<shr> X<shift, right> X<shift, left>
a0d0e21e 370
ba7f043c 371Binary C<<< "<<" >>> returns the value of its left argument shifted left by the
55497cff 372number of bits specified by the right argument. Arguments should be
982ce180 373integers. (See also L<Integer Arithmetic>.)
a0d0e21e 374
ba7f043c 375Binary C<<< ">>" >>> returns the value of its left argument shifted right by
55497cff 376the number of bits specified by the right argument. Arguments should
982ce180 377be integers. (See also L<Integer Arithmetic>.)
a0d0e21e 378
a63df121
JH
379If S<C<use integer>> (see L<Integer Arithmetic>) is in force then
380signed C integers are used (I<arithmetic shift>), otherwise unsigned C
381integers are used (I<logical shift>), even for negative shiftees.
382In arithmetic right shift the sign bit is replicated on the left,
383in logical shift zero bits come in from the left.
384
385Either way, the implementation isn't going to generate results larger
386than the size of the integer type Perl was built with (32 bits or 64 bits).
387
388Shifting by negative number of bits means the reverse shift: left
389shift becomes right shift, right shift becomes left shift. This is
390unlike in C, where negative shift is undefined.
391
392Shifting by more bits than the size of the integers means most of the
393time zero (all bits fall off), except that under S<C<use integer>>
394right overshifting a negative shiftee results in -1. This is unlike
395in C, where shifting by too many bits is undefined. A common C
396behavior is "shift by modulo wordbits", so that for example
397
398 1 >> 64 == 1 >> (64 % 64) == 1 >> 0 == 1 # Common C behavior.
399
400but that is completely accidental.
b16cf6df 401
1ca345ed 402If you get tired of being subject to your platform's native integers,
ba7f043c 403the S<C<use bigint>> pragma neatly sidesteps the issue altogether:
1ca345ed
TC
404
405 print 20 << 20; # 20971520
406 print 20 << 40; # 5120 on 32-bit machines,
407 # 21990232555520 on 64-bit machines
408 use bigint;
409 print 20 << 100; # 25353012004564588029934064107520
410
a0d0e21e 411=head2 Named Unary Operators
d74e8afc 412X<operator, named unary>
a0d0e21e
LW
413
414The various named unary operators are treated as functions with one
568e6d8b 415argument, with optional parentheses.
a0d0e21e 416
ba7f043c 417If any list operator (C<print()>, etc.) or any unary operator (C<chdir()>, etc.)
a0d0e21e
LW
418is followed by a left parenthesis as the next token, the operator and
419arguments within parentheses are taken to be of highest precedence,
3981b0eb 420just like a normal function call. For example,
1ca345ed 421because named unary operators are higher precedence than C<||>:
a0d0e21e
LW
422
423 chdir $foo || die; # (chdir $foo) || die
424 chdir($foo) || die; # (chdir $foo) || die
425 chdir ($foo) || die; # (chdir $foo) || die
426 chdir +($foo) || die; # (chdir $foo) || die
427
ba7f043c 428but, because C<"*"> is higher precedence than named operators:
a0d0e21e
LW
429
430 chdir $foo * 20; # chdir ($foo * 20)
431 chdir($foo) * 20; # (chdir $foo) * 20
432 chdir ($foo) * 20; # (chdir $foo) * 20
433 chdir +($foo) * 20; # chdir ($foo * 20)
434
435 rand 10 * 20; # rand (10 * 20)
436 rand(10) * 20; # (rand 10) * 20
437 rand (10) * 20; # (rand 10) * 20
438 rand +(10) * 20; # rand (10 * 20)
439
568e6d8b
RGS
440Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
441treated like named unary operators, but they don't follow this functional
442parenthesis rule. That means, for example, that C<-f($file).".bak"> is
ba7f043c 443equivalent to S<C<-f "$file.bak">>.
d74e8afc 444X<-X> X<filetest> X<operator, filetest>
568e6d8b 445
5ba421f6 446See also L<"Terms and List Operators (Leftward)">.
a0d0e21e
LW
447
448=head2 Relational Operators
d74e8afc 449X<relational operator> X<operator, relational>
a0d0e21e 450
1ca345ed
TC
451Perl operators that return true or false generally return values
452that can be safely used as numbers. For example, the relational
453operators in this section and the equality operators in the next
454one return C<1> for true and a special version of the defined empty
455string, C<"">, which counts as a zero but is exempt from warnings
ba7f043c 456about improper numeric conversions, just as S<C<"0 but true">> is.
1ca345ed 457
ba7f043c 458Binary C<< "<" >> returns true if the left argument is numerically less than
a0d0e21e 459the right argument.
d74e8afc 460X<< < >>
a0d0e21e 461
ba7f043c 462Binary C<< ">" >> returns true if the left argument is numerically greater
a0d0e21e 463than the right argument.
d74e8afc 464X<< > >>
a0d0e21e 465
ba7f043c 466Binary C<< "<=" >> returns true if the left argument is numerically less than
a0d0e21e 467or equal to the right argument.
d74e8afc 468X<< <= >>
a0d0e21e 469
ba7f043c 470Binary C<< ">=" >> returns true if the left argument is numerically greater
a0d0e21e 471than or equal to the right argument.
d74e8afc 472X<< >= >>
a0d0e21e 473
ba7f043c 474Binary C<"lt"> returns true if the left argument is stringwise less than
a0d0e21e 475the right argument.
d74e8afc 476X<< lt >>
a0d0e21e 477
ba7f043c 478Binary C<"gt"> returns true if the left argument is stringwise greater
a0d0e21e 479than the right argument.
d74e8afc 480X<< gt >>
a0d0e21e 481
ba7f043c 482Binary C<"le"> returns true if the left argument is stringwise less than
a0d0e21e 483or equal to the right argument.
d74e8afc 484X<< le >>
a0d0e21e 485
ba7f043c 486Binary C<"ge"> returns true if the left argument is stringwise greater
a0d0e21e 487than or equal to the right argument.
d74e8afc 488X<< ge >>
a0d0e21e
LW
489
490=head2 Equality Operators
d74e8afc 491X<equality> X<equal> X<equals> X<operator, equality>
a0d0e21e 492
ba7f043c 493Binary C<< "==" >> returns true if the left argument is numerically equal to
a0d0e21e 494the right argument.
d74e8afc 495X<==>
a0d0e21e 496
ba7f043c 497Binary C<< "!=" >> returns true if the left argument is numerically not equal
a0d0e21e 498to the right argument.
d74e8afc 499X<!=>
a0d0e21e 500
ba7f043c 501Binary C<< "<=>" >> returns -1, 0, or 1 depending on whether the left
6ee5d4e7 502argument is numerically less than, equal to, or greater than the right
ba7f043c
KW
503argument. If your platform supports C<NaN>'s (not-a-numbers) as numeric
504values, using them with C<< "<=>" >> returns undef. C<NaN> is not
505C<< "<" >>, C<< "==" >>, C<< ">" >>, C<< "<=" >> or C<< ">=" >> anything
506(even C<NaN>), so those 5 return false. S<C<< NaN != NaN >>> returns
507true, as does S<C<NaN !=> I<anything else>>. If your platform doesn't
508support C<NaN>'s then C<NaN> is just a string with numeric value 0.
509X<< <=> >>
510X<spaceship>
7d3a9d88 511
db691027
SF
512 $ perl -le '$x = "NaN"; print "No NaN support here" if $x == $x'
513 $ perl -le '$x = "NaN"; print "NaN support here" if $x != $x'
1ca345ed 514
db691027 515(Note that the L<bigint>, L<bigrat>, and L<bignum> pragmas all
ba7f043c 516support C<"NaN">.)
a0d0e21e 517
ba7f043c 518Binary C<"eq"> returns true if the left argument is stringwise equal to
a0d0e21e 519the right argument.
d74e8afc 520X<eq>
a0d0e21e 521
ba7f043c 522Binary C<"ne"> returns true if the left argument is stringwise not equal
a0d0e21e 523to the right argument.
d74e8afc 524X<ne>
a0d0e21e 525
ba7f043c 526Binary C<"cmp"> returns -1, 0, or 1 depending on whether the left
d4ad863d
JH
527argument is stringwise less than, equal to, or greater than the right
528argument.
d74e8afc 529X<cmp>
a0d0e21e 530
ba7f043c 531Binary C<"~~"> does a smartmatch between its arguments. Smart matching
1ca345ed 532is described in the next section.
0d863452
RH
533X<~~>
534
ba7f043c
KW
535C<"lt">, C<"le">, C<"ge">, C<"gt"> and C<"cmp"> use the collation (sort)
536order specified by the current C<LC_COLLATE> locale if a S<C<use
537locale>> form that includes collation is in effect. See L<perllocale>.
538Do not mix these with Unicode,
539only use them with legacy 8-bit locale encodings.
540The standard C<L<Unicode::Collate>> and
541C<L<Unicode::Collate::Locale>> modules offer much more powerful
542solutions to collation issues.
1ca345ed 543
82365311
DG
544For case-insensitive comparisions, look at the L<perlfunc/fc> case-folding
545function, available in Perl v5.16 or later:
546
547 if ( fc($x) eq fc($y) ) { ... }
548
1ca345ed
TC
549=head2 Smartmatch Operator
550
551First available in Perl 5.10.1 (the 5.10.0 version behaved differently),
552binary C<~~> does a "smartmatch" between its arguments. This is mostly
553used implicitly in the C<when> construct described in L<perlsyn>, although
554not all C<when> clauses call the smartmatch operator. Unique among all of
cc08d69f
RS
555Perl's operators, the smartmatch operator can recurse. The smartmatch
556operator is L<experimental|perlpolicy/experimental> and its behavior is
557subject to change.
1ca345ed
TC
558
559It is also unique in that all other Perl operators impose a context
560(usually string or numeric context) on their operands, autoconverting
561those operands to those imposed contexts. In contrast, smartmatch
562I<infers> contexts from the actual types of its operands and uses that
563type information to select a suitable comparison mechanism.
564
565The C<~~> operator compares its operands "polymorphically", determining how
566to compare them according to their actual types (numeric, string, array,
567hash, etc.) Like the equality operators with which it shares the same
568precedence, C<~~> returns 1 for true and C<""> for false. It is often best
569read aloud as "in", "inside of", or "is contained in", because the left
570operand is often looked for I<inside> the right operand. That makes the
40bec8a5 571order of the operands to the smartmatch operand often opposite that of
1ca345ed
TC
572the regular match operator. In other words, the "smaller" thing is usually
573placed in the left operand and the larger one in the right.
574
575The behavior of a smartmatch depends on what type of things its arguments
576are, as determined by the following table. The first row of the table
577whose types apply determines the smartmatch behavior. Because what
578actually happens is mostly determined by the type of the second operand,
579the table is sorted on the right operand instead of on the left.
580
581 Left Right Description and pseudocode
582 ===============================================================
583 Any undef check whether Any is undefined
584 like: !defined Any
585
586 Any Object invoke ~~ overloading on Object, or die
587
588 Right operand is an ARRAY:
589
590 Left Right Description and pseudocode
591 ===============================================================
592 ARRAY1 ARRAY2 recurse on paired elements of ARRAY1 and ARRAY2[2]
593 like: (ARRAY1[0] ~~ ARRAY2[0])
594 && (ARRAY1[1] ~~ ARRAY2[1]) && ...
595 HASH ARRAY any ARRAY elements exist as HASH keys
596 like: grep { exists HASH->{$_} } ARRAY
597 Regexp ARRAY any ARRAY elements pattern match Regexp
598 like: grep { /Regexp/ } ARRAY
599 undef ARRAY undef in ARRAY
600 like: grep { !defined } ARRAY
40bec8a5 601 Any ARRAY smartmatch each ARRAY element[3]
1ca345ed
TC
602 like: grep { Any ~~ $_ } ARRAY
603
604 Right operand is a HASH:
605
606 Left Right Description and pseudocode
607 ===============================================================
608 HASH1 HASH2 all same keys in both HASHes
609 like: keys HASH1 ==
610 grep { exists HASH2->{$_} } keys HASH1
611 ARRAY HASH any ARRAY elements exist as HASH keys
612 like: grep { exists HASH->{$_} } ARRAY
613 Regexp HASH any HASH keys pattern match Regexp
614 like: grep { /Regexp/ } keys HASH
615 undef HASH always false (undef can't be a key)
616 like: 0 == 1
617 Any HASH HASH key existence
618 like: exists HASH->{Any}
619
620 Right operand is CODE:
f703fc96 621
1ca345ed
TC
622 Left Right Description and pseudocode
623 ===============================================================
624 ARRAY CODE sub returns true on all ARRAY elements[1]
625 like: !grep { !CODE->($_) } ARRAY
626 HASH CODE sub returns true on all HASH keys[1]
627 like: !grep { !CODE->($_) } keys HASH
628 Any CODE sub passed Any returns true
629 like: CODE->(Any)
630
631Right operand is a Regexp:
632
633 Left Right Description and pseudocode
634 ===============================================================
635 ARRAY Regexp any ARRAY elements match Regexp
636 like: grep { /Regexp/ } ARRAY
637 HASH Regexp any HASH keys match Regexp
638 like: grep { /Regexp/ } keys HASH
639 Any Regexp pattern match
640 like: Any =~ /Regexp/
641
642 Other:
643
644 Left Right Description and pseudocode
645 ===============================================================
646 Object Any invoke ~~ overloading on Object,
647 or fall back to...
648
649 Any Num numeric equality
650 like: Any == Num
651 Num nummy[4] numeric equality
652 like: Num == nummy
653 undef Any check whether undefined
654 like: !defined(Any)
655 Any Any string equality
656 like: Any eq Any
657
658
659Notes:
660
661=over
662
663=item 1.
664Empty hashes or arrays match.
665
666=item 2.
40bec8a5 667That is, each element smartmatches the element of the same index in the other array.[3]
1ca345ed
TC
668
669=item 3.
670If a circular reference is found, fall back to referential equality.
671
672=item 4.
673Either an actual number, or a string that looks like one.
674
675=back
676
677The smartmatch implicitly dereferences any non-blessed hash or array
678reference, so the C<I<HASH>> and C<I<ARRAY>> entries apply in those cases.
679For blessed references, the C<I<Object>> entries apply. Smartmatches
680involving hashes only consider hash keys, never hash values.
681
682The "like" code entry is not always an exact rendition. For example, the
40bec8a5 683smartmatch operator short-circuits whenever possible, but C<grep> does
1ca345ed
TC
684not. Also, C<grep> in scalar context returns the number of matches, but
685C<~~> returns only true or false.
686
687Unlike most operators, the smartmatch operator knows to treat C<undef>
688specially:
689
690 use v5.10.1;
691 @array = (1, 2, 3, undef, 4, 5);
692 say "some elements undefined" if undef ~~ @array;
693
694Each operand is considered in a modified scalar context, the modification
695being that array and hash variables are passed by reference to the
696operator, which implicitly dereferences them. Both elements
697of each pair are the same:
698
699 use v5.10.1;
700
701 my %hash = (red => 1, blue => 2, green => 3,
702 orange => 4, yellow => 5, purple => 6,
703 black => 7, grey => 8, white => 9);
704
705 my @array = qw(red blue green);
706
707 say "some array elements in hash keys" if @array ~~ %hash;
708 say "some array elements in hash keys" if \@array ~~ \%hash;
709
710 say "red in array" if "red" ~~ @array;
711 say "red in array" if "red" ~~ \@array;
712
713 say "some keys end in e" if /e$/ ~~ %hash;
714 say "some keys end in e" if /e$/ ~~ \%hash;
715
40bec8a5
TC
716Two arrays smartmatch if each element in the first array smartmatches
717(that is, is "in") the corresponding element in the second array,
718recursively.
1ca345ed
TC
719
720 use v5.10.1;
721 my @little = qw(red blue green);
722 my @bigger = ("red", "blue", [ "orange", "green" ] );
723 if (@little ~~ @bigger) { # true!
724 say "little is contained in bigger";
725 }
726
727Because the smartmatch operator recurses on nested arrays, this
728will still report that "red" is in the array.
729
730 use v5.10.1;
731 my @array = qw(red blue green);
732 my $nested_array = [[[[[[[ @array ]]]]]]];
733 say "red in array" if "red" ~~ $nested_array;
734
735If two arrays smartmatch each other, then they are deep
736copies of each others' values, as this example reports:
737
738 use v5.12.0;
739 my @a = (0, 1, 2, [3, [4, 5], 6], 7);
740 my @b = (0, 1, 2, [3, [4, 5], 6], 7);
741
742 if (@a ~~ @b && @b ~~ @a) {
743 say "a and b are deep copies of each other";
744 }
745 elsif (@a ~~ @b) {
746 say "a smartmatches in b";
747 }
748 elsif (@b ~~ @a) {
749 say "b smartmatches in a";
750 }
751 else {
752 say "a and b don't smartmatch each other at all";
753 }
754
755
ba7f043c
KW
756If you were to set S<C<$b[3] = 4>>, then instead of reporting that "a and b
757are deep copies of each other", it now reports that C<"b smartmatches in a">.
758That's because the corresponding position in C<@a> contains an array that
1ca345ed
TC
759(eventually) has a 4 in it.
760
761Smartmatching one hash against another reports whether both contain the
46f8a5ea 762same keys, no more and no less. This could be used to see whether two
1ca345ed
TC
763records have the same field names, without caring what values those fields
764might have. For example:
765
766 use v5.10.1;
767 sub make_dogtag {
768 state $REQUIRED_FIELDS = { name=>1, rank=>1, serial_num=>1 };
769
770 my ($class, $init_fields) = @_;
771
772 die "Must supply (only) name, rank, and serial number"
773 unless $init_fields ~~ $REQUIRED_FIELDS;
774
775 ...
776 }
777
1b590b38
LM
778However, this only does what you mean if C<$init_fields> is indeed a hash
779reference. The condition C<$init_fields ~~ $REQUIRED_FIELDS> also allows the
780strings C<"name">, C<"rank">, C<"serial_num"> as well as any array reference
781that contains C<"name"> or C<"rank"> or C<"serial_num"> anywhere to pass
782through.
1ca345ed
TC
783
784The smartmatch operator is most often used as the implicit operator of a
785C<when> clause. See the section on "Switch Statements" in L<perlsyn>.
786
787=head3 Smartmatching of Objects
788
40bec8a5
TC
789To avoid relying on an object's underlying representation, if the
790smartmatch's right operand is an object that doesn't overload C<~~>,
791it raises the exception "C<Smartmatching a non-overloaded object
46f8a5ea
FC
792breaks encapsulation>". That's because one has no business digging
793around to see whether something is "in" an object. These are all
40bec8a5 794illegal on objects without a C<~~> overload:
1ca345ed
TC
795
796 %hash ~~ $object
797 42 ~~ $object
798 "fred" ~~ $object
799
800However, you can change the way an object is smartmatched by overloading
46f8a5ea
FC
801the C<~~> operator. This is allowed to
802extend the usual smartmatch semantics.
1ca345ed
TC
803For objects that do have an C<~~> overload, see L<overload>.
804
805Using an object as the left operand is allowed, although not very useful.
806Smartmatching rules take precedence over overloading, so even if the
807object in the left operand has smartmatch overloading, this will be
808ignored. A left operand that is a non-overloaded object falls back on a
809string or numeric comparison of whatever the C<ref> operator returns. That
810means that
811
812 $object ~~ X
813
814does I<not> invoke the overload method with C<I<X>> as an argument.
815Instead the above table is consulted as normal, and based on the type of
816C<I<X>>, overloading may or may not be invoked. For simple strings or
ba7f043c 817numbers, "in" becomes equivalent to this:
1ca345ed
TC
818
819 $object ~~ $number ref($object) == $number
820 $object ~~ $string ref($object) eq $string
821
822For example, this reports that the handle smells IOish
823(but please don't really do this!):
824
825 use IO::Handle;
826 my $fh = IO::Handle->new();
827 if ($fh ~~ /\bIO\b/) {
828 say "handle smells IOish";
829 }
830
831That's because it treats C<$fh> as a string like
832C<"IO::Handle=GLOB(0x8039e0)">, then pattern matches against that.
a034a98d 833
a0d0e21e 834=head2 Bitwise And
d74e8afc 835X<operator, bitwise, and> X<bitwise and> X<&>
a0d0e21e 836
ba7f043c 837Binary C<"&"> returns its operands ANDed together bit by bit. Although no
c791a246
KW
838warning is currently raised, the result is not well defined when this operation
839is performed on operands that aren't either numbers (see
ba7f043c 840L<Integer Arithmetic>) nor bitstrings (see L<Bitwise String Operators>).
a0d0e21e 841
ba7f043c 842Note that C<"&"> has lower priority than relational operators, so for example
1ca345ed 843the parentheses are essential in a test like
2cdc098b 844
1ca345ed 845 print "Even\n" if ($x & 1) == 0;
2cdc098b 846
ba7f043c
KW
847If the experimental "bitwise" feature is enabled via S<C<use feature
848'bitwise'>>, then this operator always treats its operand as numbers. This
849feature produces a warning unless you also use C<S<no warnings
850'experimental::bitwise'>>.
fb7054ba 851
a0d0e21e 852=head2 Bitwise Or and Exclusive Or
d74e8afc
ITB
853X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
854X<bitwise xor> X<^>
a0d0e21e 855
ba7f043c 856Binary C<"|"> returns its operands ORed together bit by bit.
a0d0e21e 857
ba7f043c 858Binary C<"^"> returns its operands XORed together bit by bit.
c791a246
KW
859
860Although no warning is currently raised, the results are not well
861defined when these operations are performed on operands that aren't either
ba7f043c 862numbers (see L<Integer Arithmetic>) nor bitstrings (see L<Bitwise String
c791a246 863Operators>).
a0d0e21e 864
ba7f043c
KW
865Note that C<"|"> and C<"^"> have lower priority than relational operators, so
866for example the parentheses are essential in a test like
2cdc098b 867
1ca345ed 868 print "false\n" if (8 | 2) != 10;
2cdc098b 869
ba7f043c
KW
870If the experimental "bitwise" feature is enabled via S<C<use feature
871'bitwise'>>, then this operator always treats its operand as numbers. This
872feature produces a warning unless you also use S<C<no warnings
873'experimental::bitwise'>>.
fb7054ba 874
a0d0e21e 875=head2 C-style Logical And
d74e8afc 876X<&&> X<logical and> X<operator, logical, and>
a0d0e21e 877
ba7f043c 878Binary C<"&&"> performs a short-circuit logical AND operation. That is,
a0d0e21e
LW
879if the left operand is false, the right operand is not even evaluated.
880Scalar or list context propagates down to the right operand if it
881is evaluated.
882
883=head2 C-style Logical Or
d74e8afc 884X<||> X<operator, logical, or>
a0d0e21e 885
ba7f043c 886Binary C<"||"> performs a short-circuit logical OR operation. That is,
a0d0e21e
LW
887if the left operand is true, the right operand is not even evaluated.
888Scalar or list context propagates down to the right operand if it
889is evaluated.
890
26d9d83b 891=head2 Logical Defined-Or
d74e8afc 892X<//> X<operator, logical, defined-or>
c963b151
BD
893
894Although it has no direct equivalent in C, Perl's C<//> operator is related
ba7f043c 895to its C-style "or". In fact, it's exactly the same as C<||>, except that it
95bee9ba 896tests the left hand side's definedness instead of its truth. Thus,
ba7f043c 897S<C<< EXPR1 // EXPR2 >>> returns the value of C<< EXPR1 >> if it's defined,
46f8a5ea
FC
898otherwise, the value of C<< EXPR2 >> is returned.
899(C<< EXPR1 >> is evaluated in scalar context, C<< EXPR2 >>
900in the context of C<< // >> itself). Usually,
ba7f043c
KW
901this is the same result as S<C<< defined(EXPR1) ? EXPR1 : EXPR2 >>> (except that
902the ternary-operator form can be used as a lvalue, while S<C<< EXPR1 // EXPR2 >>>
46f8a5ea 903cannot). This is very useful for
bdc7923b 904providing default values for variables. If you actually want to test if
ba7f043c 905at least one of C<$x> and C<$y> is defined, use S<C<defined($x // $y)>>.
c963b151 906
d042e63d 907The C<||>, C<//> and C<&&> operators return the last value evaluated
46f8a5ea 908(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
d042e63d 909portable way to find out the home directory might be:
a0d0e21e 910
c543c01b
TC
911 $home = $ENV{HOME}
912 // $ENV{LOGDIR}
913 // (getpwuid($<))[7]
914 // die "You're homeless!\n";
a0d0e21e 915
5a964f20
TC
916In particular, this means that you shouldn't use this
917for selecting between two aggregates for assignment:
918
bf55d65d
LTC
919 @a = @b || @c; # This doesn't do the right thing
920 @a = scalar(@b) || @c; # because it really means this.
921 @a = @b ? @b : @c; # This works fine, though.
5a964f20 922
1ca345ed 923As alternatives to C<&&> and C<||> when used for
f23102e2 924control flow, Perl provides the C<and> and C<or> operators (see below).
ba7f043c
KW
925The short-circuit behavior is identical. The precedence of C<"and">
926and C<"or"> is much lower, however, so that you can safely use them after a
5a964f20 927list operator without the need for parentheses:
a0d0e21e
LW
928
929 unlink "alpha", "beta", "gamma"
930 or gripe(), next LINE;
931
932With the C-style operators that would have been written like this:
933
934 unlink("alpha", "beta", "gamma")
935 || (gripe(), next LINE);
936
1ca345ed
TC
937It would be even more readable to write that this way:
938
939 unless(unlink("alpha", "beta", "gamma")) {
940 gripe();
941 next LINE;
942 }
943
ba7f043c 944Using C<"or"> for assignment is unlikely to do what you want; see below.
5a964f20
TC
945
946=head2 Range Operators
d74e8afc 947X<operator, range> X<range> X<..> X<...>
a0d0e21e 948
ba7f043c 949Binary C<".."> is the range operator, which is really two different
fb53bbb2 950operators depending on the context. In list context, it returns a
54ae734e 951list of values counting (up by ones) from the left value to the right
2cdbc966 952value. If the left value is greater than the right value then it
fb53bbb2 953returns the empty list. The range operator is useful for writing
ba7f043c 954S<C<foreach (1..10)>> loops and for doing slice operations on arrays. In
2cdbc966
JD
955the current implementation, no temporary array is created when the
956range operator is used as the expression in C<foreach> loops, but older
957versions of Perl might burn a lot of memory when you write something
958like this:
a0d0e21e
LW
959
960 for (1 .. 1_000_000) {
961 # code
54310121 962 }
a0d0e21e 963
8f0f46f8 964The range operator also works on strings, using the magical
965auto-increment, see below.
54ae734e 966
ba7f043c 967In scalar context, C<".."> returns a boolean value. The operator is
8f0f46f8 968bistable, like a flip-flop, and emulates the line-range (comma)
ba7f043c 969operator of B<sed>, B<awk>, and various editors. Each C<".."> operator
8f0f46f8 970maintains its own boolean state, even across calls to a subroutine
46f8a5ea 971that contains it. It is false as long as its left operand is false.
a0d0e21e
LW
972Once the left operand is true, the range operator stays true until the
973right operand is true, I<AFTER> which the range operator becomes false
8f0f46f8 974again. It doesn't become false till the next time the range operator
975is evaluated. It can test the right operand and become false on the
976same evaluation it became true (as in B<awk>), but it still returns
46f8a5ea 977true once. If you don't want it to test the right operand until the
ba7f043c
KW
978next evaluation, as in B<sed>, just use three dots (C<"...">) instead of
979two. In all other regards, C<"..."> behaves just like C<".."> does.
19799a22
GS
980
981The right operand is not evaluated while the operator is in the
982"false" state, and the left operand is not evaluated while the
983operator is in the "true" state. The precedence is a little lower
984than || and &&. The value returned is either the empty string for
8f0f46f8 985false, or a sequence number (beginning with 1) for true. The sequence
986number is reset for each range encountered. The final sequence number
ba7f043c 987in a range has the string C<"E0"> appended to it, which doesn't affect
8f0f46f8 988its numeric value, but gives you something to search for if you want
989to exclude the endpoint. You can exclude the beginning point by
990waiting for the sequence number to be greater than 1.
df5f8116 991
ba7f043c 992If either operand of scalar C<".."> is a constant expression,
df5f8116
CW
993that operand is considered true if it is equal (C<==>) to the current
994input line number (the C<$.> variable).
995
ba7f043c 996To be pedantic, the comparison is actually S<C<int(EXPR) == int(EXPR)>>,
df5f8116
CW
997but that is only an issue if you use a floating point expression; when
998implicitly using C<$.> as described in the previous paragraph, the
ba7f043c 999comparison is S<C<int(EXPR) == int($.)>> which is only an issue when C<$.>
df5f8116 1000is set to a floating point value and you are not reading from a file.
ba7f043c 1001Furthermore, S<C<"span" .. "spat">> or S<C<2.18 .. 3.14>> will not do what
df5f8116
CW
1002you want in scalar context because each of the operands are evaluated
1003using their integer representation.
1004
1005Examples:
a0d0e21e
LW
1006
1007As a scalar operator:
1008
df5f8116 1009 if (101 .. 200) { print; } # print 2nd hundred lines, short for
950b09ed 1010 # if ($. == 101 .. $. == 200) { print; }
9f10b797
RGS
1011
1012 next LINE if (1 .. /^$/); # skip header lines, short for
f343f960 1013 # next LINE if ($. == 1 .. /^$/);
9f10b797
RGS
1014 # (typically in a loop labeled LINE)
1015
1016 s/^/> / if (/^$/ .. eof()); # quote body
a0d0e21e 1017
5a964f20
TC
1018 # parse mail messages
1019 while (<>) {
1020 $in_header = 1 .. /^$/;
df5f8116
CW
1021 $in_body = /^$/ .. eof;
1022 if ($in_header) {
f343f960 1023 # do something
df5f8116 1024 } else { # in body
f343f960 1025 # do something else
df5f8116 1026 }
5a964f20 1027 } continue {
df5f8116 1028 close ARGV if eof; # reset $. each file
5a964f20
TC
1029 }
1030
acf31ca5
SF
1031Here's a simple example to illustrate the difference between
1032the two range operators:
1033
1034 @lines = (" - Foo",
1035 "01 - Bar",
1036 "1 - Baz",
1037 " - Quux");
1038
9f10b797
RGS
1039 foreach (@lines) {
1040 if (/0/ .. /1/) {
acf31ca5
SF
1041 print "$_\n";
1042 }
1043 }
1044
46f8a5ea 1045This program will print only the line containing "Bar". If
9f10b797 1046the range operator is changed to C<...>, it will also print the
acf31ca5
SF
1047"Baz" line.
1048
1049And now some examples as a list operator:
a0d0e21e 1050
1ca345ed
TC
1051 for (101 .. 200) { print } # print $_ 100 times
1052 @foo = @foo[0 .. $#foo]; # an expensive no-op
1053 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
a0d0e21e 1054
5a964f20 1055The range operator (in list context) makes use of the magical
5f05dabc 1056auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
1057can say
1058
c543c01b 1059 @alphabet = ("A" .. "Z");
a0d0e21e 1060
54ae734e 1061to get all normal letters of the English alphabet, or
a0d0e21e 1062
c543c01b 1063 $hexdigit = (0 .. 9, "a" .. "f")[$num & 15];
a0d0e21e
LW
1064
1065to get a hexadecimal digit, or
1066
1ca345ed
TC
1067 @z2 = ("01" .. "31");
1068 print $z2[$mday];
a0d0e21e 1069
ea4f5703
YST
1070to get dates with leading zeros.
1071
1072If the final value specified is not in the sequence that the magical
1073increment would produce, the sequence goes until the next value would
1074be longer than the final value specified.
1075
1076If the initial value specified isn't part of a magical increment
c543c01b 1077sequence (that is, a non-empty string matching C</^[a-zA-Z]*[0-9]*\z/>),
ea4f5703
YST
1078only the initial value will be returned. So the following will only
1079return an alpha:
1080
c543c01b 1081 use charnames "greek";
ea4f5703
YST
1082 my @greek_small = ("\N{alpha}" .. "\N{omega}");
1083
c543c01b
TC
1084To get the 25 traditional lowercase Greek letters, including both sigmas,
1085you could use this instead:
ea4f5703 1086
c543c01b 1087 use charnames "greek";
1ca345ed
TC
1088 my @greek_small = map { chr } ( ord("\N{alpha}")
1089 ..
1090 ord("\N{omega}")
1091 );
c543c01b
TC
1092
1093However, because there are I<many> other lowercase Greek characters than
1094just those, to match lowercase Greek characters in a regular expression,
47c56cc8
KW
1095you could use the pattern C</(?:(?=\p{Greek})\p{Lower})+/> (or the
1096L<experimental feature|perlrecharclass/Extended Bracketed Character
1097Classes> C<S</(?[ \p{Greek} & \p{Lower} ])+/>>).
a0d0e21e 1098
ba7f043c 1099Because each operand is evaluated in integer form, S<C<2.18 .. 3.14>> will
df5f8116
CW
1100return two elements in list context.
1101
1102 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
1103
a0d0e21e 1104=head2 Conditional Operator
d74e8afc 1105X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
a0d0e21e 1106
ba7f043c
KW
1107Ternary C<"?:"> is the conditional operator, just as in C. It works much
1108like an if-then-else. If the argument before the C<?> is true, the
1109argument before the C<:> is returned, otherwise the argument after the
1110C<:> is returned. For example:
cb1a09d0 1111
54310121 1112 printf "I have %d dog%s.\n", $n,
c543c01b 1113 ($n == 1) ? "" : "s";
cb1a09d0
AD
1114
1115Scalar or list context propagates downward into the 2nd
54310121 1116or 3rd argument, whichever is selected.
cb1a09d0 1117
db691027
SF
1118 $x = $ok ? $y : $z; # get a scalar
1119 @x = $ok ? @y : @z; # get an array
1120 $x = $ok ? @y : @z; # oops, that's just a count!
cb1a09d0
AD
1121
1122The operator may be assigned to if both the 2nd and 3rd arguments are
1123legal lvalues (meaning that you can assign to them):
a0d0e21e 1124
db691027 1125 ($x_or_y ? $x : $y) = $z;
a0d0e21e 1126
5a964f20
TC
1127Because this operator produces an assignable result, using assignments
1128without parentheses will get you in trouble. For example, this:
1129
db691027 1130 $x % 2 ? $x += 10 : $x += 2
5a964f20
TC
1131
1132Really means this:
1133
db691027 1134 (($x % 2) ? ($x += 10) : $x) += 2
5a964f20
TC
1135
1136Rather than this:
1137
db691027 1138 ($x % 2) ? ($x += 10) : ($x += 2)
5a964f20 1139
19799a22
GS
1140That should probably be written more simply as:
1141
db691027 1142 $x += ($x % 2) ? 10 : 2;
19799a22 1143
4633a7c4 1144=head2 Assignment Operators
d74e8afc 1145X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
5ac3b81c 1146X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
fb7054ba 1147X<%=> X<^=> X<x=> X<&.=> X<|.=> X<^.=>
a0d0e21e 1148
ba7f043c 1149C<"="> is the ordinary assignment operator.
a0d0e21e
LW
1150
1151Assignment operators work as in C. That is,
1152
db691027 1153 $x += 2;
a0d0e21e
LW
1154
1155is equivalent to
1156
db691027 1157 $x = $x + 2;
a0d0e21e
LW
1158
1159although without duplicating any side effects that dereferencing the lvalue
ba7f043c 1160might trigger, such as from C<tie()>. Other assignment operators work similarly.
54310121 1161The following are recognized:
a0d0e21e 1162
fb7054ba
FC
1163 **= += *= &= &.= <<= &&=
1164 -= /= |= |.= >>= ||=
1165 .= %= ^= ^.= //=
9f10b797 1166 x=
a0d0e21e 1167
19799a22 1168Although these are grouped by family, they all have the precedence
82848c10
FC
1169of assignment. These combined assignment operators can only operate on
1170scalars, whereas the ordinary assignment operator can assign to arrays,
1171hashes, lists and even references. (See L<"Context"|perldata/Context>
1172and L<perldata/List value constructors>, and L<perlref/Assigning to
1173References>.)
a0d0e21e 1174
b350dd2f
GS
1175Unlike in C, the scalar assignment operator produces a valid lvalue.
1176Modifying an assignment is equivalent to doing the assignment and
1177then modifying the variable that was assigned to. This is useful
1178for modifying a copy of something, like this:
a0d0e21e 1179
1ca345ed
TC
1180 ($tmp = $global) =~ tr/13579/24680/;
1181
1182Although as of 5.14, that can be also be accomplished this way:
1183
1184 use v5.14;
1185 $tmp = ($global =~ tr/13579/24680/r);
a0d0e21e
LW
1186
1187Likewise,
1188
db691027 1189 ($x += 2) *= 3;
a0d0e21e
LW
1190
1191is equivalent to
1192
db691027
SF
1193 $x += 2;
1194 $x *= 3;
a0d0e21e 1195
b350dd2f
GS
1196Similarly, a list assignment in list context produces the list of
1197lvalues assigned to, and a list assignment in scalar context returns
1198the number of elements produced by the expression on the right hand
1199side of the assignment.
1200
ba7f043c 1201The three dotted bitwise assignment operators (C<&.=> C<|.=> C<^.=>) are new in
fb7054ba
FC
1202Perl 5.22 and experimental. See L</Bitwise String Operators>.
1203
748a9306 1204=head2 Comma Operator
d74e8afc 1205X<comma> X<operator, comma> X<,>
a0d0e21e 1206
ba7f043c 1207Binary C<","> is the comma operator. In scalar context it evaluates
a0d0e21e
LW
1208its left argument, throws that value away, then evaluates its right
1209argument and returns that value. This is just like C's comma operator.
1210
5a964f20 1211In list context, it's just the list argument separator, and inserts
ed5c6d31
PJ
1212both its arguments into the list. These arguments are also evaluated
1213from left to right.
a0d0e21e 1214
ba7f043c
KW
1215The C<< => >> operator (sometimes pronounced "fat comma") is a synonym
1216for the comma except that it causes a
4e1988c6 1217word on its left to be interpreted as a string if it begins with a letter
344f2c40
IG
1218or underscore and is composed only of letters, digits and underscores.
1219This includes operands that might otherwise be interpreted as operators,
46f8a5ea 1220constants, single number v-strings or function calls. If in doubt about
c543c01b 1221this behavior, the left operand can be quoted explicitly.
344f2c40
IG
1222
1223Otherwise, the C<< => >> operator behaves exactly as the comma operator
1224or list argument separator, according to context.
1225
1226For example:
a44e5664
MS
1227
1228 use constant FOO => "something";
1229
1230 my %h = ( FOO => 23 );
1231
1232is equivalent to:
1233
1234 my %h = ("FOO", 23);
1235
1236It is I<NOT>:
1237
1238 my %h = ("something", 23);
1239
719b43e8
RGS
1240The C<< => >> operator is helpful in documenting the correspondence
1241between keys and values in hashes, and other paired elements in lists.
748a9306 1242
a12b8f3c
FC
1243 %hash = ( $key => $value );
1244 login( $username => $password );
a44e5664 1245
4e1988c6
FC
1246The special quoting behavior ignores precedence, and hence may apply to
1247I<part> of the left operand:
1248
1249 print time.shift => "bbb";
1250
ba7f043c 1251That example prints something like C<"1314363215shiftbbb">, because the
4e1988c6
FC
1252C<< => >> implicitly quotes the C<shift> immediately on its left, ignoring
1253the fact that C<time.shift> is the entire left operand.
1254
a0d0e21e 1255=head2 List Operators (Rightward)
d74e8afc 1256X<operator, list, rightward> X<list operator>
a0d0e21e 1257
c543c01b 1258On the right side of a list operator, the comma has very low precedence,
a0d0e21e
LW
1259such that it controls all comma-separated expressions found there.
1260The only operators with lower precedence are the logical operators
ba7f043c 1261C<"and">, C<"or">, and C<"not">, which may be used to evaluate calls to list
1ca345ed
TC
1262operators without the need for parentheses:
1263
1264 open HANDLE, "< :utf8", "filename" or die "Can't open: $!\n";
1265
1266However, some people find that code harder to read than writing
1267it with parentheses:
1268
1269 open(HANDLE, "< :utf8", "filename") or die "Can't open: $!\n";
1270
ba7f043c 1271in which case you might as well just use the more customary C<"||"> operator:
a0d0e21e 1272
1ca345ed 1273 open(HANDLE, "< :utf8", "filename") || die "Can't open: $!\n";
a0d0e21e 1274
a95b3d6a 1275See also discussion of list operators in L</Terms and List Operators (Leftward)>.
a0d0e21e
LW
1276
1277=head2 Logical Not
d74e8afc 1278X<operator, logical, not> X<not>
a0d0e21e 1279
ba7f043c
KW
1280Unary C<"not"> returns the logical negation of the expression to its right.
1281It's the equivalent of C<"!"> except for the very low precedence.
a0d0e21e
LW
1282
1283=head2 Logical And
d74e8afc 1284X<operator, logical, and> X<and>
a0d0e21e 1285
ba7f043c 1286Binary C<"and"> returns the logical conjunction of the two surrounding
c543c01b
TC
1287expressions. It's equivalent to C<&&> except for the very low
1288precedence. This means that it short-circuits: the right
a0d0e21e
LW
1289expression is evaluated only if the left expression is true.
1290
59ab9d6e 1291=head2 Logical or and Exclusive Or
f23102e2 1292X<operator, logical, or> X<operator, logical, xor>
59ab9d6e 1293X<operator, logical, exclusive or>
f23102e2 1294X<or> X<xor>
a0d0e21e 1295
ba7f043c 1296Binary C<"or"> returns the logical disjunction of the two surrounding
c543c01b
TC
1297expressions. It's equivalent to C<||> except for the very low precedence.
1298This makes it useful for control flow:
5a964f20
TC
1299
1300 print FH $data or die "Can't write to FH: $!";
1301
c543c01b
TC
1302This means that it short-circuits: the right expression is evaluated
1303only if the left expression is false. Due to its precedence, you must
1304be careful to avoid using it as replacement for the C<||> operator.
1305It usually works out better for flow control than in assignments:
5a964f20 1306
db691027
SF
1307 $x = $y or $z; # bug: this is wrong
1308 ($x = $y) or $z; # really means this
1309 $x = $y || $z; # better written this way
5a964f20 1310
19799a22 1311However, when it's a list-context assignment and you're trying to use
ba7f043c 1312C<||> for control flow, you probably need C<"or"> so that the assignment
5a964f20
TC
1313takes higher precedence.
1314
1315 @info = stat($file) || die; # oops, scalar sense of stat!
1316 @info = stat($file) or die; # better, now @info gets its due
1317
c963b151
BD
1318Then again, you could always use parentheses.
1319
ba7f043c 1320Binary C<"xor"> returns the exclusive-OR of the two surrounding expressions.
c543c01b 1321It cannot short-circuit (of course).
a0d0e21e 1322
59ab9d6e
MB
1323There is no low precedence operator for defined-OR.
1324
a0d0e21e 1325=head2 C Operators Missing From Perl
d74e8afc
ITB
1326X<operator, missing from perl> X<&> X<*>
1327X<typecasting> X<(TYPE)>
a0d0e21e
LW
1328
1329Here is what C has that Perl doesn't:
1330
1331=over 8
1332
1333=item unary &
1334
ba7f043c 1335Address-of operator. (But see the C<"\"> operator for taking a reference.)
a0d0e21e
LW
1336
1337=item unary *
1338
46f8a5ea 1339Dereference-address operator. (Perl's prefix dereferencing
ba7f043c 1340operators are typed: C<$>, C<@>, C<%>, and C<&>.)
a0d0e21e
LW
1341
1342=item (TYPE)
1343
19799a22 1344Type-casting operator.
a0d0e21e
LW
1345
1346=back
1347
5f05dabc 1348=head2 Quote and Quote-like Operators
89d205f2 1349X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
d74e8afc
ITB
1350X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
1351X<escape sequence> X<escape>
1352
a0d0e21e
LW
1353While we usually think of quotes as literal values, in Perl they
1354function as operators, providing various kinds of interpolating and
1355pattern matching capabilities. Perl provides customary quote characters
1356for these behaviors, but also provides a way for you to choose your
1357quote character for any of them. In the following table, a C<{}> represents
9f10b797 1358any pair of delimiters you choose.
a0d0e21e 1359
2c268ad5
TP
1360 Customary Generic Meaning Interpolates
1361 '' q{} Literal no
1362 "" qq{} Literal yes
af9219ee 1363 `` qx{} Command yes*
2c268ad5 1364 qw{} Word list no
af9219ee
MG
1365 // m{} Pattern match yes*
1366 qr{} Pattern yes*
1367 s{}{} Substitution yes*
2c268ad5 1368 tr{}{} Transliteration no (but see below)
c543c01b 1369 y{}{} Transliteration no (but see below)
7e3b091d 1370 <<EOF here-doc yes*
a0d0e21e 1371
af9219ee
MG
1372 * unless the delimiter is ''.
1373
87275199 1374Non-bracketing delimiters use the same character fore and aft, but the four
c543c01b 1375sorts of ASCII brackets (round, angle, square, curly) all nest, which means
9f10b797 1376that
87275199 1377
c543c01b 1378 q{foo{bar}baz}
35f2feb0 1379
9f10b797 1380is the same as
87275199 1381
c543c01b 1382 'foo{bar}baz'
87275199
GS
1383
1384Note, however, that this does not always work for quoting Perl code:
1385
db691027 1386 $s = q{ if($x eq "}") ... }; # WRONG
87275199 1387
ba7f043c 1388is a syntax error. The C<L<Text::Balanced>> module (standard as of v5.8,
c543c01b 1389and from CPAN before then) is able to do this properly.
87275199 1390
19799a22 1391There can be whitespace between the operator and the quoting
fb73857a 1392characters, except when C<#> is being used as the quoting character.
ba7f043c 1393C<q#foo#> is parsed as the string C<foo>, while S<C<q #foo#>> is the
19799a22
GS
1394operator C<q> followed by a comment. Its argument will be taken
1395from the next line. This allows you to write:
fb73857a
PP
1396
1397 s {foo} # Replace foo
1398 {bar} # with bar.
1399
c543c01b
TC
1400The following escape sequences are available in constructs that interpolate,
1401and in transliterations:
5691ca5f 1402X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N> X<\N{}>
04341565 1403X<\o{}>
5691ca5f 1404
2c4c1ff2
KW
1405 Sequence Note Description
1406 \t tab (HT, TAB)
1407 \n newline (NL)
1408 \r return (CR)
1409 \f form feed (FF)
1410 \b backspace (BS)
1411 \a alarm (bell) (BEL)
1412 \e escape (ESC)
c543c01b 1413 \x{263A} [1,8] hex char (example: SMILEY)
2c4c1ff2 1414 \x1b [2,8] restricted range hex char (example: ESC)
fb121860 1415 \N{name} [3] named Unicode character or character sequence
2c4c1ff2
KW
1416 \N{U+263D} [4,8] Unicode character (example: FIRST QUARTER MOON)
1417 \c[ [5] control char (example: chr(27))
1418 \o{23072} [6,8] octal char (example: SMILEY)
1419 \033 [7,8] restricted range octal char (example: ESC)
5691ca5f
KW
1420
1421=over 4
1422
1423=item [1]
1424
2c4c1ff2
KW
1425The result is the character specified by the hexadecimal number between
1426the braces. See L</[8]> below for details on which character.
96448467 1427
46f8a5ea 1428Only hexadecimal digits are valid between the braces. If an invalid
96448467
DG
1429character is encountered, a warning will be issued and the invalid
1430character and all subsequent characters (valid or invalid) within the
1431braces will be discarded.
1432
1433If there are no valid digits between the braces, the generated character is
1434the NULL character (C<\x{00}>). However, an explicit empty brace (C<\x{}>)
c543c01b 1435will not cause a warning (currently).
40687185
KW
1436
1437=item [2]
1438
2c4c1ff2
KW
1439The result is the character specified by the hexadecimal number in the range
14400x00 to 0xFF. See L</[8]> below for details on which character.
96448467
DG
1441
1442Only hexadecimal digits are valid following C<\x>. When C<\x> is followed
2c4c1ff2 1443by fewer than two valid digits, any valid digits will be zero-padded. This
ba7f043c 1444means that C<\x7> will be interpreted as C<\x07>, and a lone C<"\x"> will be
2c4c1ff2 1445interpreted as C<\x00>. Except at the end of a string, having fewer than
c543c01b 1446two valid digits will result in a warning. Note that although the warning
96448467
DG
1447says the illegal character is ignored, it is only ignored as part of the
1448escape and will still be used as the subsequent character in the string.
1449For example:
1450
1451 Original Result Warns?
1452 "\x7" "\x07" no
1453 "\x" "\x00" no
1454 "\x7q" "\x07q" yes
1455 "\xq" "\x00q" yes
1456
40687185
KW
1457=item [3]
1458
fb121860 1459The result is the Unicode character or character sequence given by I<name>.
2c4c1ff2 1460See L<charnames>.
40687185
KW
1461
1462=item [4]
1463
ba7f043c 1464S<C<\N{U+I<hexadecimal number>}>> means the Unicode character whose Unicode code
2c4c1ff2 1465point is I<hexadecimal number>.
40687185
KW
1466
1467=item [5]
1468
5691ca5f
KW
1469The character following C<\c> is mapped to some other character as shown in the
1470table:
1471
1472 Sequence Value
1473 \c@ chr(0)
1474 \cA chr(1)
1475 \ca chr(1)
1476 \cB chr(2)
1477 \cb chr(2)
1478 ...
1479 \cZ chr(26)
1480 \cz chr(26)
1481 \c[ chr(27)
ba7f043c 1482 # See below for chr(28)
5691ca5f
KW
1483 \c] chr(29)
1484 \c^ chr(30)
c3e9d7a9 1485 \c_ chr(31)
ba7f043c
KW
1486 \c? chr(127) # (on ASCII platforms; see below for link to
1487 # EBCDIC discussion)
5691ca5f 1488
d813941f 1489In other words, it's the character whose code point has had 64 xor'd with
c3e9d7a9
KW
1490its uppercase. C<\c?> is DELETE on ASCII platforms because
1491S<C<ord("?") ^ 64>> is 127, and
ba7f043c 1492C<\c@> is NULL because the ord of C<"@"> is 64, so xor'ing 64 itself produces 0.
d813941f 1493
ba7f043c 1494Also, C<\c\I<X>> yields S<C< chr(28) . "I<X>">> for any I<X>, but cannot come at the
5691ca5f
KW
1495end of a string, because the backslash would be parsed as escaping the end
1496quote.
1497
1498On ASCII platforms, the resulting characters from the list above are the
1499complete set of ASCII controls. This isn't the case on EBCDIC platforms; see
c3e9d7a9
KW
1500L<perlebcdic/OPERATOR DIFFERENCES> for a full discussion of the
1501differences between these for ASCII versus EBCDIC platforms.
5691ca5f 1502
c3e9d7a9 1503Use of any other character following the C<"c"> besides those listed above is
63a63d81
KW
1504discouraged, and as of Perl v5.20, the only characters actually allowed
1505are the printable ASCII ones, minus the left brace C<"{">. What happens
1506for any of the allowed other characters is that the value is derived by
1507xor'ing with the seventh bit, which is 64, and a warning raised if
1508enabled. Using the non-allowed characters generates a fatal error.
5691ca5f
KW
1509
1510To get platform independent controls, you can use C<\N{...}>.
1511
40687185
KW
1512=item [6]
1513
2c4c1ff2
KW
1514The result is the character specified by the octal number between the braces.
1515See L</[8]> below for details on which character.
04341565
DG
1516
1517If a character that isn't an octal digit is encountered, a warning is raised,
1518and the value is based on the octal digits before it, discarding it and all
1519following characters up to the closing brace. It is a fatal error if there are
1520no octal digits at all.
1521
1522=item [7]
1523
c543c01b 1524The result is the character specified by the three-digit octal number in the
2c4c1ff2
KW
1525range 000 to 777 (but best to not use above 077, see next paragraph). See
1526L</[8]> below for details on which character.
1527
1528Some contexts allow 2 or even 1 digit, but any usage without exactly
40687185 1529three digits, the first being a zero, may give unintended results. (For
5db3e519
FC
1530example, in a regular expression it may be confused with a backreference;
1531see L<perlrebackslash/Octal escapes>.) Starting in Perl 5.14, you may
c543c01b 1532use C<\o{}> instead, which avoids all these problems. Otherwise, it is best to
04341565
DG
1533use this construct only for ordinals C<\077> and below, remembering to pad to
1534the left with zeros to make three digits. For larger ordinals, either use
ba7f043c
KW
1535C<\o{}>, or convert to something else, such as to hex and use C<\N{U+}>
1536(which is portable between platforms with different character sets) or
1537C<\x{}> instead.
40687185 1538
2c4c1ff2
KW
1539=item [8]
1540
c543c01b 1541Several constructs above specify a character by a number. That number
2c4c1ff2 1542gives the character's position in the character set encoding (indexed from 0).
c543c01b 1543This is called synonymously its ordinal, code position, or code point. Perl
2c4c1ff2
KW
1544works on platforms that have a native encoding currently of either ASCII/Latin1
1545or EBCDIC, each of which allow specification of 256 characters. In general, if
1546the number is 255 (0xFF, 0377) or below, Perl interprets this in the platform's
1547native encoding. If the number is 256 (0x100, 0400) or above, Perl interprets
c543c01b 1548it as a Unicode code point and the result is the corresponding Unicode
2c4c1ff2
KW
1549character. For example C<\x{50}> and C<\o{120}> both are the number 80 in
1550decimal, which is less than 256, so the number is interpreted in the native
1551character set encoding. In ASCII the character in the 80th position (indexed
ba7f043c 1552from 0) is the letter C<"P">, and in EBCDIC it is the ampersand symbol C<"&">.
2c4c1ff2
KW
1553C<\x{100}> and C<\o{400}> are both 256 in decimal, so the number is interpreted
1554as a Unicode code point no matter what the native encoding is. The name of the
9fef6a0d 1555character in the 256th position (indexed by 0) in Unicode is
2c4c1ff2
KW
1556C<LATIN CAPITAL LETTER A WITH MACRON>.
1557
9fef6a0d 1558There are a couple of exceptions to the above rule. S<C<\N{U+I<hex number>}>> is
ba7f043c
KW
1559always interpreted as a Unicode code point, so that C<\N{U+0050}> is C<"P"> even
1560on EBCDIC platforms. And if C<S<L<use encoding|encoding>>> is in effect, the
2c4c1ff2
KW
1561number is considered to be in that encoding, and is translated from that into
1562the platform's native encoding if there is a corresponding native character;
1563otherwise to Unicode.
1564
5691ca5f 1565=back
4c77eaa2 1566
e526e8bb 1567B<NOTE>: Unlike C and other languages, Perl has no C<\v> escape sequence for
8b312c40 1568the vertical tab (VT, which is 11 in both ASCII and EBCDIC), but you may
ba7f043c 1569use C<\N{VT}>, C<\ck>, C<\N{U+0b}>, or C<\x0b>. (C<\v>
e526e8bb
KW
1570does have meaning in regular expression patterns in Perl, see L<perlre>.)
1571
1572The following escape sequences are available in constructs that interpolate,
904501ec 1573but not in transliterations.
628253b8 1574X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q> X<\F>
904501ec 1575
c543c01b
TC
1576 \l lowercase next character only
1577 \u titlecase (not uppercase!) next character only
e4d34742
EB
1578 \L lowercase all characters till \E or end of string
1579 \U uppercase all characters till \E or end of string
628253b8 1580 \F foldcase all characters till \E or end of string
736fe711
KW
1581 \Q quote (disable) pattern metacharacters till \E or
1582 end of string
7e31b643 1583 \E end either case modification or quoted section
c543c01b
TC
1584 (whichever was last seen)
1585
736fe711
KW
1586See L<perlfunc/quotemeta> for the exact definition of characters that
1587are quoted by C<\Q>.
1588
628253b8 1589C<\L>, C<\U>, C<\F>, and C<\Q> can stack, in which case you need one
c543c01b
TC
1590C<\E> for each. For example:
1591
9fef6a0d
KW
1592 say"This \Qquoting \ubusiness \Uhere isn't quite\E done yet,\E is it?";
1593 This quoting\ Business\ HERE\ ISN\'T\ QUITE\ done\ yet\, is it?
a0d0e21e 1594
ba7f043c
KW
1595If a S<C<use locale>> form that includes C<LC_CTYPE> is in effect (see
1596L<perllocale>), the case map used by C<\l>, C<\L>, C<\u>, and C<\U> is
1597taken from the current locale. If Unicode (for example, C<\N{}> or code
1598points of 0x100 or beyond) is being used, the case map used by C<\l>,
1599C<\L>, C<\u>, and C<\U> is as defined by Unicode. That means that
1600case-mapping a single character can sometimes produce a sequence of
1601several characters.
1602Under S<C<use locale>>, C<\F> produces the same results as C<\L>
31f05a37
KW
1603for all locales but a UTF-8 one, where it instead uses the Unicode
1604definition.
a034a98d 1605
5a964f20
TC
1606All systems use the virtual C<"\n"> to represent a line terminator,
1607called a "newline". There is no such thing as an unvarying, physical
19799a22 1608newline character. It is only an illusion that the operating system,
5a964f20
TC
1609device drivers, C libraries, and Perl all conspire to preserve. Not all
1610systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
c543c01b 1611on the ancient Macs (pre-MacOS X) of yesteryear, these used to be reversed,
ba7f043c 1612and on systems without a line terminator,
c543c01b 1613printing C<"\n"> might emit no actual data. In general, use C<"\n"> when
5a964f20
TC
1614you mean a "newline" for your system, but use the literal ASCII when you
1615need an exact character. For example, most networking protocols expect
2a380090 1616and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
1617and although they often accept just C<"\012">, they seldom tolerate just
1618C<"\015">. If you get in the habit of using C<"\n"> for networking,
1619you may be burned some day.
d74e8afc
ITB
1620X<newline> X<line terminator> X<eol> X<end of line>
1621X<\n> X<\r> X<\r\n>
5a964f20 1622
904501ec
MG
1623For constructs that do interpolate, variables beginning with "C<$>"
1624or "C<@>" are interpolated. Subscripted variables such as C<$a[3]> or
ad0f383a
A
1625C<< $href->{key}[0] >> are also interpolated, as are array and hash slices.
1626But method calls such as C<< $obj->meth >> are not.
af9219ee
MG
1627
1628Interpolating an array or slice interpolates the elements in order,
1629separated by the value of C<$">, so is equivalent to interpolating
ba7f043c 1630S<C<join $", @array>>. "Punctuation" arrays such as C<@*> are usually
c543c01b
TC
1631interpolated only if the name is enclosed in braces C<@{*}>, but the
1632arrays C<@_>, C<@+>, and C<@-> are interpolated even without braces.
af9219ee 1633
bc7b91c6
EB
1634For double-quoted strings, the quoting from C<\Q> is applied after
1635interpolation and escapes are processed.
1636
1637 "abc\Qfoo\tbar$s\Exyz"
1638
1639is equivalent to
1640
1641 "abc" . quotemeta("foo\tbar$s") . "xyz"
1642
1643For the pattern of regex operators (C<qr//>, C<m//> and C<s///>),
1644the quoting from C<\Q> is applied after interpolation is processed,
46f8a5ea
FC
1645but before escapes are processed. This allows the pattern to match
1646literally (except for C<$> and C<@>). For example, the following matches:
bc7b91c6
EB
1647
1648 '\s\t' =~ /\Q\s\t/
1649
1650Because C<$> or C<@> trigger interpolation, you'll need to use something
1651like C</\Quser\E\@\Qhost/> to match them literally.
1d2dff63 1652
a0d0e21e
LW
1653Patterns are subject to an additional level of interpretation as a
1654regular expression. This is done as a second pass, after variables are
1655interpolated, so that regular expressions may be incorporated into the
1656pattern from the variables. If this is not what you want, use C<\Q> to
1657interpolate a variable literally.
1658
19799a22
GS
1659Apart from the behavior described above, Perl does not expand
1660multiple levels of interpolation. In particular, contrary to the
1661expectations of shell programmers, back-quotes do I<NOT> interpolate
1662within double quotes, nor do single quotes impede evaluation of
1663variables when used within double quotes.
a0d0e21e 1664
5f05dabc 1665=head2 Regexp Quote-Like Operators
d74e8afc 1666X<operator, regexp>
cb1a09d0 1667
5f05dabc 1668Here are the quote-like operators that apply to pattern
cb1a09d0
AD
1669matching and related activities.
1670
a0d0e21e
LW
1671=over 8
1672
ba7f043c 1673=item C<qr/I<STRING>/msixpodualn>
01c6f5f4 1674X<qr> X</i> X</m> X</o> X</s> X</x> X</p>
a0d0e21e 1675
87e95b7f
YO
1676This operator quotes (and possibly compiles) its I<STRING> as a regular
1677expression. I<STRING> is interpolated the same way as I<PATTERN>
ba7f043c 1678in C<m/I<PATTERN>/>. If C<"'"> is used as the delimiter, no interpolation
87e95b7f 1679is done. Returns a Perl value which may be used instead of the
ba7f043c 1680corresponding C</I<STRING>/msixpodualn> expression. The returned value is a
46f8a5ea 1681normalized version of the original pattern. It magically differs from
1c8ee595
CO
1682a string containing the same characters: C<ref(qr/x/)> returns "Regexp";
1683however, dereferencing it is not well defined (you currently get the
1684normalized version of the original pattern, but this may change).
1685
a0d0e21e 1686
87e95b7f
YO
1687For example,
1688
1689 $rex = qr/my.STRING/is;
85dd5c8b 1690 print $rex; # prints (?si-xm:my.STRING)
87e95b7f
YO
1691 s/$rex/foo/;
1692
1693is equivalent to
1694
1695 s/my.STRING/foo/is;
1696
1697The result may be used as a subpattern in a match:
1698
1699 $re = qr/$pattern/;
7188ca43
KW
1700 $string =~ /foo${re}bar/; # can be interpolated in other
1701 # patterns
87e95b7f
YO
1702 $string =~ $re; # or used standalone
1703 $string =~ /$re/; # or this way
1704
ba7f043c
KW
1705Since Perl may compile the pattern at the moment of execution of the C<qr()>
1706operator, using C<qr()> may have speed advantages in some situations,
1707notably if the result of C<qr()> is used standalone:
87e95b7f
YO
1708
1709 sub match {
1710 my $patterns = shift;
1711 my @compiled = map qr/$_/i, @$patterns;
1712 grep {
1713 my $success = 0;
1714 foreach my $pat (@compiled) {
1715 $success = 1, last if /$pat/;
1716 }
1717 $success;
1718 } @_;
5a964f20
TC
1719 }
1720
87e95b7f 1721Precompilation of the pattern into an internal representation at
ba7f043c 1722the moment of C<qr()> avoids the need to recompile the pattern every
87e95b7f
YO
1723time a match C</$pat/> is attempted. (Perl has many other internal
1724optimizations, but none would be triggered in the above example if
ba7f043c 1725we did not use C<qr()> operator.)
87e95b7f 1726
765fa144 1727Options (specified by the following modifiers) are:
87e95b7f
YO
1728
1729 m Treat string as multiple lines.
1730 s Treat string as single line. (Make . match a newline)
1731 i Do case-insensitive pattern matching.
1732 x Use extended regular expressions.
1733 p When matching preserve a copy of the matched string so
7188ca43 1734 that ${^PREMATCH}, ${^MATCH}, ${^POSTMATCH} will be
ba7f043c
KW
1735 defined (ignored starting in v5.20) as these are always
1736 defined starting in that relese
87e95b7f 1737 o Compile pattern only once.
7188ca43 1738 a ASCII-restrict: Use ASCII for \d, \s, \w; specifying two
ba7f043c
KW
1739 a's further restricts things to that that no ASCII
1740 character will match a non-ASCII one under /i.
1741 l Use the current run-time locale's rules.
48cbae4f
SK
1742 u Use Unicode rules.
1743 d Use Unicode or native charset, as in 5.12 and earlier.
33be4c61 1744 n Non-capture mode. Don't let () fill in $1, $2, etc...
87e95b7f
YO
1745
1746If a precompiled pattern is embedded in a larger pattern then the effect
ba7f043c
KW
1747of C<"msixpluadn"> will be propagated appropriately. The effect that the
1748C</o> modifier has is not propagated, being restricted to those patterns
87e95b7f
YO
1749explicitly using it.
1750
b6fa137b 1751The last four modifiers listed above, added in Perl 5.14,
850b7ec9 1752control the character set rules, but C</a> is the only one you are likely
18509dec
KW
1753to want to specify explicitly; the other three are selected
1754automatically by various pragmas.
da392a17 1755
ba7f043c 1756See L<perlre> for additional information on valid syntax for I<STRING>, and
5e2aa8f5 1757for a detailed look at the semantics of regular expressions. In
1ca345ed
TC
1758particular, all modifiers except the largely obsolete C</o> are further
1759explained in L<perlre/Modifiers>. C</o> is described in the next section.
a0d0e21e 1760
ba7f043c 1761=item C<m/I<PATTERN>/msixpodualngc>
89d205f2
YO
1762X<m> X<operator, match>
1763X<regexp, options> X<regexp> X<regex, options> X<regex>
01c6f5f4 1764X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c>
a0d0e21e 1765
ba7f043c 1766=item C</I<PATTERN>/msixpodualngc>
a0d0e21e 1767
5a964f20 1768Searches a string for a pattern match, and in scalar context returns
19799a22 1769true if it succeeds, false if it fails. If no string is specified
ba7f043c 1770via the C<=~> or C<!~> operator, the C<$_> string is searched. (The
19799a22
GS
1771string specified with C<=~> need not be an lvalue--it may be the
1772result of an expression evaluation, but remember the C<=~> binds
006671a6 1773rather tightly.) See also L<perlre>.
a0d0e21e 1774
f6050459 1775Options are as described in C<qr//> above; in addition, the following match
01c6f5f4 1776process modifiers are available:
a0d0e21e 1777
950b09ed 1778 g Match globally, i.e., find all occurrences.
7188ca43
KW
1779 c Do not reset search position on a failed match when /g is
1780 in effect.
a0d0e21e 1781
ba7f043c 1782If C<"/"> is the delimiter then the initial C<m> is optional. With the C<m>
c543c01b 1783you can use any pair of non-whitespace (ASCII) characters
725a61d7 1784as delimiters. This is particularly useful for matching path names
ba7f043c 1785that contain C<"/">, to avoid LTS (leaning toothpick syndrome). If C<"?"> is
725a61d7 1786the delimiter, then a match-only-once rule applies,
ba7f043c
KW
1787described in C<m?I<PATTERN>?> below. If C<"'"> (single quote) is the delimiter,
1788no interpolation is performed on the I<PATTERN>.
1789When using a delimiter character valid in an identifier, whitespace is required
ed02a3bf 1790after the C<m>.
a0d0e21e 1791
ba7f043c 1792I<PATTERN> may contain variables, which will be interpolated
532c9e80 1793every time the pattern search is evaluated, except
1f247705
GS
1794for when the delimiter is a single quote. (Note that C<$(>, C<$)>, and
1795C<$|> are not interpolated because they look like end-of-string tests.)
532c9e80
KW
1796Perl will not recompile the pattern unless an interpolated
1797variable that it contains changes. You can force Perl to skip the
1798test and never recompile by adding a C</o> (which stands for "once")
1799after the trailing delimiter.
1800Once upon a time, Perl would recompile regular expressions
1801unnecessarily, and this modifier was useful to tell it not to do so, in the
5cc41653 1802interests of speed. But now, the only reasons to use C</o> are one of:
532c9e80
KW
1803
1804=over
1805
1806=item 1
1807
1808The variables are thousands of characters long and you know that they
1809don't change, and you need to wring out the last little bit of speed by
1810having Perl skip testing for that. (There is a maintenance penalty for
1811doing this, as mentioning C</o> constitutes a promise that you won't
18509dec 1812change the variables in the pattern. If you do change them, Perl won't
532c9e80
KW
1813even notice.)
1814
1815=item 2
1816
1817you want the pattern to use the initial values of the variables
1818regardless of whether they change or not. (But there are saner ways
1819of accomplishing this than using C</o>.)
1820
fa9b8686
DM
1821=item 3
1822
1823If the pattern contains embedded code, such as
1824
1825 use re 'eval';
1826 $code = 'foo(?{ $x })';
1827 /$code/
1828
1829then perl will recompile each time, even though the pattern string hasn't
1830changed, to ensure that the current value of C<$x> is seen each time.
1831Use C</o> if you want to avoid this.
1832
532c9e80 1833=back
a0d0e21e 1834
18509dec
KW
1835The bottom line is that using C</o> is almost never a good idea.
1836
ba7f043c 1837=item The empty pattern C<//>
e9d89077 1838
ba7f043c 1839If the I<PATTERN> evaluates to the empty string, the last
46f8a5ea 1840I<successfully> matched regular expression is used instead. In this
c543c01b 1841case, only the C<g> and C<c> flags on the empty pattern are honored;
46f8a5ea 1842the other flags are taken from the original pattern. If no match has
d65afb4b
HS
1843previously succeeded, this will (silently) act instead as a genuine
1844empty pattern (which will always match).
a0d0e21e 1845
89d205f2
YO
1846Note that it's possible to confuse Perl into thinking C<//> (the empty
1847regex) is really C<//> (the defined-or operator). Perl is usually pretty
1848good about this, but some pathological cases might trigger this, such as
ba7f043c
KW
1849C<$x///> (is that S<C<($x) / (//)>> or S<C<$x // />>?) and S<C<print $fh //>>
1850(S<C<print $fh(//>> or S<C<print($fh //>>?). In all of these examples, Perl
89d205f2
YO
1851will assume you meant defined-or. If you meant the empty regex, just
1852use parentheses or spaces to disambiguate, or even prefix the empty
c963b151
BD
1853regex with an C<m> (so C<//> becomes C<m//>).
1854
e9d89077
DN
1855=item Matching in list context
1856
19799a22 1857If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1858list consisting of the subexpressions matched by the parentheses in the
3ff8ecf9
BF
1859pattern, that is, (C<$1>, C<$2>, C<$3>...) (Note that here C<$1> etc. are
1860also set). When there are no parentheses in the pattern, the return
1861value is the list C<(1)> for success.
1862With or without parentheses, an empty list is returned upon failure.
a0d0e21e
LW
1863
1864Examples:
1865
7188ca43
KW
1866 open(TTY, "+</dev/tty")
1867 || die "can't access /dev/tty: $!";
c543c01b 1868
7188ca43 1869 <TTY> =~ /^y/i && foo(); # do foo if desired
a0d0e21e 1870
7188ca43 1871 if (/Version: *([0-9.]*)/) { $version = $1; }
a0d0e21e 1872
7188ca43 1873 next if m#^/usr/spool/uucp#;
a0d0e21e 1874
7188ca43
KW
1875 # poor man's grep
1876 $arg = shift;
1877 while (<>) {
1878 print if /$arg/o; # compile only once (no longer needed!)
1879 }
a0d0e21e 1880
7188ca43 1881 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
a0d0e21e 1882
ba7f043c
KW
1883This last example splits C<$foo> into the first two words and the
1884remainder of the line, and assigns those three fields to C<$F1>, C<$F2>, and
1885C<$Etc>. The conditional is true if any variables were assigned; that is,
c543c01b 1886if the pattern matched.
a0d0e21e 1887
19799a22 1888The C</g> modifier specifies global pattern matching--that is,
46f8a5ea
FC
1889matching as many times as possible within the string. How it behaves
1890depends on the context. In list context, it returns a list of the
19799a22 1891substrings matched by any capturing parentheses in the regular
46f8a5ea 1892expression. If there are no parentheses, it returns a list of all
19799a22
GS
1893the matched strings, as if there were parentheses around the whole
1894pattern.
a0d0e21e 1895
7e86de3e 1896In scalar context, each execution of C<m//g> finds the next match,
19799a22 1897returning true if it matches, and false if there is no further match.
3dd93342 1898The position after the last match can be read or set using the C<pos()>
46f8a5ea 1899function; see L<perlfunc/pos>. A failed match normally resets the
7e86de3e 1900search position to the beginning of the string, but you can avoid that
46f8a5ea 1901by adding the C</c> modifier (for example, C<m//gc>). Modifying the target
7e86de3e 1902string also resets the search position.
c90c0ff4 1903
ba7f043c 1904=item C<\G I<assertion>>
e9d89077 1905
c90c0ff4 1906You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
3dd93342 1907zero-width assertion that matches the exact position where the
46f8a5ea 1908previous C<m//g>, if any, left off. Without the C</g> modifier, the
3dd93342 1909C<\G> assertion still anchors at C<pos()> as it was at the start of
1910the operation (see L<perlfunc/pos>), but the match is of course only
46f8a5ea 1911attempted once. Using C<\G> without C</g> on a target string that has
3dd93342 1912not previously had a C</g> match applied to it is the same as using
1913the C<\A> assertion to match the beginning of the string. Note also
1914that, currently, C<\G> is only properly supported when anchored at the
1915very beginning of the pattern.
c90c0ff4
PP
1916
1917Examples:
a0d0e21e
LW
1918
1919 # list context
1920 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1921
1922 # scalar context
c543c01b
TC
1923 local $/ = "";
1924 while ($paragraph = <>) {
1925 while ($paragraph =~ /\p{Ll}['")]*[.!?]+['")]*\s/g) {
19799a22 1926 $sentences++;
a0d0e21e
LW
1927 }
1928 }
c543c01b
TC
1929 say $sentences;
1930
1931Here's another way to check for sentences in a paragraph:
1932
7188ca43
KW
1933 my $sentence_rx = qr{
1934 (?: (?<= ^ ) | (?<= \s ) ) # after start-of-string or
1935 # whitespace
1936 \p{Lu} # capital letter
1937 .*? # a bunch of anything
1938 (?<= \S ) # that ends in non-
1939 # whitespace
1940 (?<! \b [DMS]r ) # but isn't a common abbr.
1941 (?<! \b Mrs )
1942 (?<! \b Sra )
1943 (?<! \b St )
1944 [.?!] # followed by a sentence
1945 # ender
1946 (?= $ | \s ) # in front of end-of-string
1947 # or whitespace
1948 }sx;
1949 local $/ = "";
1950 while (my $paragraph = <>) {
1951 say "NEW PARAGRAPH";
1952 my $count = 0;
1953 while ($paragraph =~ /($sentence_rx)/g) {
1954 printf "\tgot sentence %d: <%s>\n", ++$count, $1;
c543c01b 1955 }
7188ca43 1956 }
c543c01b
TC
1957
1958Here's how to use C<m//gc> with C<\G>:
a0d0e21e 1959
137443ea 1960 $_ = "ppooqppqq";
44a8e56a
PP
1961 while ($i++ < 2) {
1962 print "1: '";
c90c0ff4 1963 print $1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1964 print "2: '";
c90c0ff4 1965 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1966 print "3: '";
c90c0ff4 1967 print $1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1968 }
5d43e42d 1969 print "Final: '$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a
PP
1970
1971The last example should print:
1972
1973 1: 'oo', pos=4
137443ea 1974 2: 'q', pos=5
44a8e56a
PP
1975 3: 'pp', pos=7
1976 1: '', pos=7
137443ea
PP
1977 2: 'q', pos=8
1978 3: '', pos=8
5d43e42d
DC
1979 Final: 'q', pos=8
1980
1981Notice that the final match matched C<q> instead of C<p>, which a match
46f8a5ea
FC
1982without the C<\G> anchor would have done. Also note that the final match
1983did not update C<pos>. C<pos> is only updated on a C</g> match. If the
c543c01b
TC
1984final match did indeed match C<p>, it's a good bet that you're running a
1985very old (pre-5.6.0) version of Perl.
44a8e56a 1986
c90c0ff4 1987A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1988combine several regexps like this to process a string part-by-part,
c90c0ff4
PP
1989doing different actions depending on which regexp matched. Each
1990regexp tries to match where the previous one leaves off.
e7ea3e70 1991
3fe9a6f1 1992 $_ = <<'EOL';
7188ca43
KW
1993 $url = URI::URL->new( "http://example.com/" );
1994 die if $url eq "xXx";
3fe9a6f1 1995 EOL
c543c01b
TC
1996
1997 LOOP: {
950b09ed 1998 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
7188ca43
KW
1999 print(" lowercase"), redo LOOP
2000 if /\G\p{Ll}+\b[,.;]?\s*/gc;
2001 print(" UPPERCASE"), redo LOOP
2002 if /\G\p{Lu}+\b[,.;]?\s*/gc;
2003 print(" Capitalized"), redo LOOP
2004 if /\G\p{Lu}\p{Ll}+\b[,.;]?\s*/gc;
c543c01b 2005 print(" MiXeD"), redo LOOP if /\G\pL+\b[,.;]?\s*/gc;
7188ca43
KW
2006 print(" alphanumeric"), redo LOOP
2007 if /\G[\p{Alpha}\pN]+\b[,.;]?\s*/gc;
c543c01b 2008 print(" line-noise"), redo LOOP if /\G\W+/gc;
950b09ed 2009 print ". That's all!\n";
c543c01b 2010 }
e7ea3e70
IZ
2011
2012Here is the output (split into several lines):
2013
7188ca43
KW
2014 line-noise lowercase line-noise UPPERCASE line-noise UPPERCASE
2015 line-noise lowercase line-noise lowercase line-noise lowercase
2016 lowercase line-noise lowercase lowercase line-noise lowercase
2017 lowercase line-noise MiXeD line-noise. That's all!
44a8e56a 2018
ba7f043c 2019=item C<m?I<PATTERN>?msixpodualngc>
725a61d7 2020X<?> X<operator, match-once>
87e95b7f 2021
ba7f043c 2022=item C<?I<PATTERN>?msixpodualngc>
55d389e7 2023
ba7f043c
KW
2024This is just like the C<m/I<PATTERN>/> search, except that it matches
2025only once between calls to the C<reset()> operator. This is a useful
87e95b7f 2026optimization when you want to see only the first occurrence of
ceb131e8 2027something in each file of a set of files, for instance. Only C<m??>
87e95b7f
YO
2028patterns local to the current package are reset.
2029
2030 while (<>) {
ceb131e8 2031 if (m?^$?) {
87e95b7f
YO
2032 # blank line between header and body
2033 }
2034 } continue {
725a61d7 2035 reset if eof; # clear m?? status for next file
87e95b7f
YO
2036 }
2037
c543c01b
TC
2038Another example switched the first "latin1" encoding it finds
2039to "utf8" in a pod file:
2040
2041 s//utf8/ if m? ^ =encoding \h+ \K latin1 ?x;
2042
2043The match-once behavior is controlled by the match delimiter being
4932eeca 2044C<?>; with any other delimiter this is the normal C<m//> operator.
725a61d7 2045
ba7f043c 2046In the past, the leading C<m> in C<m?I<PATTERN>?> was optional, but omitting it
0381ecf1
MH
2047would produce a deprecation warning. As of v5.22.0, omitting it produces a
2048syntax error. If you encounter this construct in older code, you can just add
2049C<m>.
87e95b7f 2050
ba7f043c 2051=item C<s/I<PATTERN>/I<REPLACEMENT>/msixpodualngcer>
87e95b7f 2052X<substitute> X<substitution> X<replace> X<regexp, replace>
4f4d7508 2053X<regexp, substitute> X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c> X</e> X</r>
87e95b7f
YO
2054
2055Searches a string for a pattern, and if found, replaces that pattern
2056with the replacement text and returns the number of substitutions
2057made. Otherwise it returns false (specifically, the empty string).
2058
c543c01b 2059If the C</r> (non-destructive) option is used then it runs the
679563bb
KW
2060substitution on a copy of the string and instead of returning the
2061number of substitutions, it returns the copy whether or not a
c543c01b
TC
2062substitution occurred. The original string is never changed when
2063C</r> is used. The copy will always be a plain string, even if the
2064input is an object or a tied variable.
4f4d7508 2065
87e95b7f 2066If no string is specified via the C<=~> or C<!~> operator, the C<$_>
c543c01b
TC
2067variable is searched and modified. Unless the C</r> option is used,
2068the string specified must be a scalar variable, an array element, a
2069hash element, or an assignment to one of those; that is, some sort of
2070scalar lvalue.
87e95b7f
YO
2071
2072If the delimiter chosen is a single quote, no interpolation is
ba7f043c
KW
2073done on either the I<PATTERN> or the I<REPLACEMENT>. Otherwise, if the
2074I<PATTERN> contains a C<$> that looks like a variable rather than an
87e95b7f
YO
2075end-of-string test, the variable will be interpolated into the pattern
2076at run-time. If you want the pattern compiled only once the first time
2077the variable is interpolated, use the C</o> option. If the pattern
2078evaluates to the empty string, the last successfully executed regular
2079expression is used instead. See L<perlre> for further explanation on these.
87e95b7f 2080
ba7f043c 2081Options are as with C<m//> with the addition of the following replacement
87e95b7f
YO
2082specific options:
2083
2084 e Evaluate the right side as an expression.
7188ca43
KW
2085 ee Evaluate the right side as a string then eval the
2086 result.
2087 r Return substitution and leave the original string
2088 untouched.
87e95b7f 2089
ed02a3bf
DN
2090Any non-whitespace delimiter may replace the slashes. Add space after
2091the C<s> when using a character allowed in identifiers. If single quotes
2092are used, no interpretation is done on the replacement string (the C</e>
3ff8ecf9 2093modifier overrides this, however). Note that Perl treats backticks
ed02a3bf 2094as normal delimiters; the replacement text is not evaluated as a command.
ba7f043c 2095If the I<PATTERN> is delimited by bracketing quotes, the I<REPLACEMENT> has
1ca345ed 2096its own pair of quotes, which may or may not be bracketing quotes, for example,
87e95b7f
YO
2097C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
2098replacement portion to be treated as a full-fledged Perl expression
2099and evaluated right then and there. It is, however, syntax checked at
46f8a5ea 2100compile-time. A second C<e> modifier will cause the replacement portion
87e95b7f
YO
2101to be C<eval>ed before being run as a Perl expression.
2102
2103Examples:
2104
7188ca43 2105 s/\bgreen\b/mauve/g; # don't change wintergreen
87e95b7f
YO
2106
2107 $path =~ s|/usr/bin|/usr/local/bin|;
2108
2109 s/Login: $foo/Login: $bar/; # run-time pattern
2110
7188ca43
KW
2111 ($foo = $bar) =~ s/this/that/; # copy first, then
2112 # change
2113 ($foo = "$bar") =~ s/this/that/; # convert to string,
2114 # copy, then change
4f4d7508
DC
2115 $foo = $bar =~ s/this/that/r; # Same as above using /r
2116 $foo = $bar =~ s/this/that/r
7188ca43
KW
2117 =~ s/that/the other/r; # Chained substitutes
2118 # using /r
2119 @foo = map { s/this/that/r } @bar # /r is very useful in
2120 # maps
87e95b7f 2121
7188ca43 2122 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-cnt
87e95b7f
YO
2123
2124 $_ = 'abc123xyz';
2125 s/\d+/$&*2/e; # yields 'abc246xyz'
2126 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
2127 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
2128
2129 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
2130 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
2131 s/^=(\w+)/pod($1)/ge; # use function call
2132
4f4d7508 2133 $_ = 'abc123xyz';
db691027 2134 $x = s/abc/def/r; # $x is 'def123xyz' and
4f4d7508
DC
2135 # $_ remains 'abc123xyz'.
2136
87e95b7f
YO
2137 # expand variables in $_, but dynamics only, using
2138 # symbolic dereferencing
2139 s/\$(\w+)/${$1}/g;
2140
2141 # Add one to the value of any numbers in the string
2142 s/(\d+)/1 + $1/eg;
2143
c543c01b
TC
2144 # Titlecase words in the last 30 characters only
2145 substr($str, -30) =~ s/\b(\p{Alpha}+)\b/\u\L$1/g;
2146
87e95b7f
YO
2147 # This will expand any embedded scalar variable
2148 # (including lexicals) in $_ : First $1 is interpolated
2149 # to the variable name, and then evaluated
2150 s/(\$\w+)/$1/eeg;
2151
2152 # Delete (most) C comments.
2153 $program =~ s {
2154 /\* # Match the opening delimiter.
2155 .*? # Match a minimal number of characters.
2156 \*/ # Match the closing delimiter.
2157 } []gsx;
2158
7188ca43
KW
2159 s/^\s*(.*?)\s*$/$1/; # trim whitespace in $_,
2160 # expensively
87e95b7f 2161
7188ca43
KW
2162 for ($variable) { # trim whitespace in $variable,
2163 # cheap
87e95b7f
YO
2164 s/^\s+//;
2165 s/\s+$//;
2166 }
2167
2168 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
2169
ba7f043c
KW
2170Note the use of C<$> instead of C<\> in the last example. Unlike
2171B<sed>, we use the \<I<digit>> form only in the left hand side.
87e95b7f
YO
2172Anywhere else it's $<I<digit>>.
2173
2174Occasionally, you can't use just a C</g> to get all the changes
2175to occur that you might want. Here are two common cases:
2176
2177 # put commas in the right places in an integer
2178 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
2179
2180 # expand tabs to 8-column spacing
2181 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
2182
2183=back
2184
2185=head2 Quote-Like Operators
2186X<operator, quote-like>
2187
01c6f5f4
RGS
2188=over 4
2189
ba7f043c 2190=item C<q/I<STRING>/>
5d44bfff 2191X<q> X<quote, single> X<'> X<''>
a0d0e21e 2192
ba7f043c 2193=item C<'I<STRING>'>
a0d0e21e 2194
19799a22 2195A single-quoted, literal string. A backslash represents a backslash
68dc0745
PP
2196unless followed by the delimiter or another backslash, in which case
2197the delimiter or backslash is interpolated.
a0d0e21e
LW
2198
2199 $foo = q!I said, "You said, 'She said it.'"!;
2200 $bar = q('This is it.');
68dc0745 2201 $baz = '\n'; # a two-character string
a0d0e21e 2202
ba7f043c 2203=item C<qq/I<STRING>/>
d74e8afc 2204X<qq> X<quote, double> X<"> X<"">
a0d0e21e 2205
ba7f043c 2206=item "I<STRING>"
a0d0e21e
LW
2207
2208A double-quoted, interpolated string.
2209
2210 $_ .= qq
2211 (*** The previous line contains the naughty word "$1".\n)
19799a22 2212 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 2213 $baz = "\n"; # a one-character string
a0d0e21e 2214
ba7f043c 2215=item C<qx/I<STRING>/>
d74e8afc 2216X<qx> X<`> X<``> X<backtick>
a0d0e21e 2217
ba7f043c 2218=item C<`I<STRING>`>
a0d0e21e 2219
43dd4d21 2220A string which is (possibly) interpolated and then executed as a
f703fc96 2221system command with F</bin/sh> or its equivalent. Shell wildcards,
43dd4d21
JH
2222pipes, and redirections will be honored. The collected standard
2223output of the command is returned; standard error is unaffected. In
2224scalar context, it comes back as a single (potentially multi-line)
ba7f043c
KW
2225string, or C<undef> if the command failed. In list context, returns a
2226list of lines (however you've defined lines with C<$/> or
2227C<$INPUT_RECORD_SEPARATOR>), or an empty list if the command failed.
5a964f20
TC
2228
2229Because backticks do not affect standard error, use shell file descriptor
2230syntax (assuming the shell supports this) if you care to address this.
2231To capture a command's STDERR and STDOUT together:
a0d0e21e 2232
5a964f20
TC
2233 $output = `cmd 2>&1`;
2234
2235To capture a command's STDOUT but discard its STDERR:
2236
2237 $output = `cmd 2>/dev/null`;
2238
2239To capture a command's STDERR but discard its STDOUT (ordering is
2240important here):
2241
2242 $output = `cmd 2>&1 1>/dev/null`;
2243
2244To exchange a command's STDOUT and STDERR in order to capture the STDERR
2245but leave its STDOUT to come out the old STDERR:
2246
2247 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
2248
2249To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
2250to redirect them separately to files, and then read from those files
2251when the program is done:
5a964f20 2252
2359510d 2253 system("program args 1>program.stdout 2>program.stderr");
5a964f20 2254
30398227
SP
2255The STDIN filehandle used by the command is inherited from Perl's STDIN.
2256For example:
2257
c543c01b
TC
2258 open(SPLAT, "stuff") || die "can't open stuff: $!";
2259 open(STDIN, "<&SPLAT") || die "can't dupe SPLAT: $!";
40bbb707 2260 print STDOUT `sort`;
30398227 2261
40bbb707 2262will print the sorted contents of the file named F<"stuff">.
30398227 2263
5a964f20
TC
2264Using single-quote as a delimiter protects the command from Perl's
2265double-quote interpolation, passing it on to the shell instead:
2266
2267 $perl_info = qx(ps $$); # that's Perl's $$
2268 $shell_info = qx'ps $$'; # that's the new shell's $$
2269
19799a22 2270How that string gets evaluated is entirely subject to the command
5a964f20
TC
2271interpreter on your system. On most platforms, you will have to protect
2272shell metacharacters if you want them treated literally. This is in
2273practice difficult to do, as it's unclear how to escape which characters.
ba7f043c 2274See L<perlsec> for a clean and safe example of a manual C<fork()> and C<exec()>
5a964f20 2275to emulate backticks safely.
a0d0e21e 2276
bb32b41a
GS
2277On some platforms (notably DOS-like ones), the shell may not be
2278capable of dealing with multiline commands, so putting newlines in
2279the string may not get you what you want. You may be able to evaluate
2280multiple commands in a single line by separating them with the command
1ca345ed
TC
2281separator character, if your shell supports that (for example, C<;> on
2282many Unix shells and C<&> on the Windows NT C<cmd> shell).
bb32b41a 2283
3ff8ecf9 2284Perl will attempt to flush all files opened for
0f897271
GS
2285output before starting the child process, but this may not be supported
2286on some platforms (see L<perlport>). To be safe, you may need to set
ba7f043c
KW
2287C<$|> (C<$AUTOFLUSH> in C<L<English>>) or call the C<autoflush()> method of
2288C<L<IO::Handle>> on any open handles.
0f897271 2289
bb32b41a
GS
2290Beware that some command shells may place restrictions on the length
2291of the command line. You must ensure your strings don't exceed this
2292limit after any necessary interpolations. See the platform-specific
2293release notes for more details about your particular environment.
2294
5a964f20
TC
2295Using this operator can lead to programs that are difficult to port,
2296because the shell commands called vary between systems, and may in
2297fact not be present at all. As one example, the C<type> command under
2298the POSIX shell is very different from the C<type> command under DOS.
2299That doesn't mean you should go out of your way to avoid backticks
2300when they're the right way to get something done. Perl was made to be
2301a glue language, and one of the things it glues together is commands.
2302Just understand what you're getting yourself into.
bb32b41a 2303
7cf4dd3e
DB
2304Like C<system>, backticks put the child process exit code in C<$?>.
2305If you'd like to manually inspect failure, you can check all possible
2306failure modes by inspecting C<$?> like this:
2307
2308 if ($? == -1) {
2309 print "failed to execute: $!\n";
2310 }
2311 elsif ($? & 127) {
2312 printf "child died with signal %d, %s coredump\n",
2313 ($? & 127), ($? & 128) ? 'with' : 'without';
2314 }
2315 else {
2316 printf "child exited with value %d\n", $? >> 8;
2317 }
2318
fe43a9cc
TC
2319Use the L<open> pragma to control the I/O layers used when reading the
2320output of the command, for example:
2321
2322 use open IN => ":encoding(UTF-8)";
2323 my $x = `cmd-producing-utf-8`;
2324
da87341d 2325See L</"I/O Operators"> for more discussion.
a0d0e21e 2326
ba7f043c 2327=item C<qw/I<STRING>/>
d74e8afc 2328X<qw> X<quote, list> X<quote, words>
945c54fd 2329
ba7f043c 2330Evaluates to a list of the words extracted out of I<STRING>, using embedded
945c54fd
JH
2331whitespace as the word delimiters. It can be understood as being roughly
2332equivalent to:
2333
c543c01b 2334 split(" ", q/STRING/);
945c54fd 2335
efb1e162
CW
2336the differences being that it generates a real list at compile time, and
2337in scalar context it returns the last element in the list. So
945c54fd
JH
2338this expression:
2339
2340 qw(foo bar baz)
2341
2342is semantically equivalent to the list:
2343
c543c01b 2344 "foo", "bar", "baz"
945c54fd
JH
2345
2346Some frequently seen examples:
2347
2348 use POSIX qw( setlocale localeconv )
2349 @EXPORT = qw( foo bar baz );
2350
ba7f043c 2351A common mistake is to try to separate the words with commas or to
945c54fd 2352put comments into a multi-line C<qw>-string. For this reason, the
ba7f043c
KW
2353S<C<use warnings>> pragma and the B<-w> switch (that is, the C<$^W> variable)
2354produces warnings if the I<STRING> contains the C<","> or the C<"#"> character.
945c54fd 2355
ba7f043c 2356=item C<tr/I<SEARCHLIST>/I<REPLACEMENTLIST>/cdsr>
d74e8afc 2357X<tr> X<y> X<transliterate> X</c> X</d> X</s>
a0d0e21e 2358
ba7f043c 2359=item C<y/I<SEARCHLIST>/I<REPLACEMENTLIST>/cdsr>
a0d0e21e 2360
2c268ad5 2361Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
2362with the corresponding character in the replacement list. It returns
2363the number of characters replaced or deleted. If no string is
ba7f043c 2364specified via the C<=~> or C<!~> operator, the C<$_> string is transliterated.
c543c01b
TC
2365
2366If the C</r> (non-destructive) option is present, a new copy of the string
2367is made and its characters transliterated, and this copy is returned no
2368matter whether it was modified or not: the original string is always
2369left unchanged. The new copy is always a plain string, even if the input
2370string is an object or a tied variable.
8ada0baa 2371
c543c01b
TC
2372Unless the C</r> option is used, the string specified with C<=~> must be a
2373scalar variable, an array element, a hash element, or an assignment to one
2374of those; in other words, an lvalue.
8ff32507 2375
89d205f2 2376A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
2c268ad5 2377does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121 2378For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
ba7f043c 2379I<SEARCHLIST> is delimited by bracketing quotes, the I<REPLACEMENTLIST> has
c543c01b
TC
2380its own pair of quotes, which may or may not be bracketing quotes;
2381for example, C<tr[aeiouy][yuoiea]> or C<tr(+\-*/)/ABCD/>.
2382
ba7f043c
KW
2383Characters may be literals or any of the escape sequences accepted in
2384double-quoted strings. But there is no interpolation, so C<"$"> and
2385C<"@"> are treated as literals. A hyphen at the beginning or end, or
2386preceded by a backslash is considered a literal. Escape sequence
2387details are in L<the table near the beginning of this section|/Quote and
f4240379 2388Quote-like Operators>.
ba7f043c 2389
c543c01b 2390Note that C<tr> does B<not> do regular expression character classes such as
ba7f043c 2391C<\d> or C<\pL>. The C<tr> operator is not equivalent to the C<L<tr(1)>>
c543c01b
TC
2392utility. If you want to map strings between lower/upper cases, see
2393L<perlfunc/lc> and L<perlfunc/uc>, and in general consider using the C<s>
2394operator if you need regular expressions. The C<\U>, C<\u>, C<\L>, and
2395C<\l> string-interpolation escapes on the right side of a substitution
2396operator will perform correct case-mappings, but C<tr[a-z][A-Z]> will not
2397(except sometimes on legacy 7-bit data).
cc255d5f 2398
f4240379
KW
2399Most ranges are unportable between character sets, but certain ones
2400signal Perl to do special handling to make them portable. There are two
2401classes of portable ranges. The first are any subsets of the ranges
2402C<A-Z>, C<a-z>, and C<0-9>, when expressed as literal characters.
2403
2404 tr/h-k/H-K/
2405
2406capitalizes the letters C<"h">, C<"i">, C<"j">, and C<"k"> and nothing
2407else, no matter what the platform's character set is. In contrast, all
2408of
2409
2410 tr/\x68-\x6B/\x48-\x4B/
2411 tr/h-\x6B/H-\x4B/
2412 tr/\x68-k/\x48-K/
2413
2414do the same capitalizations as the previous example when run on ASCII
2415platforms, but something completely different on EBCDIC ones.
2416
2417The second class of portable ranges is invoked when one or both of the
2418range's end points are expressed as C<\N{...}>
2419
2420 $string =~ tr/\N{U+20}-\N{U+7E}//d;
2421
2422removes from C<$string> all the platform's characters which are
2423equivalent to any of Unicode U+0020, U+0021, ... U+007D, U+007E. This
2424is a portable range, and has the same effect on every platform it is
2425run on. It turns out that in this example, these are the ASCII
2426printable characters. So after this is run, C<$string> has only
2427controls and characters which have no ASCII equivalents.
2428
2429But, even for portable ranges, it is not generally obvious what is
2430included without having to look things up. A sound principle is to use
2431only ranges that begin from and end at either ASCII alphabetics of equal
2432case (C<b-e>, C<b-E>), or digits (C<1-4>). Anything else is unclear
2433(and unportable unless C<\N{...}> is used). If in doubt, spell out the
8ada0baa
JH
2434character sets in full.
2435
a0d0e21e
LW
2436Options:
2437
2438 c Complement the SEARCHLIST.
2439 d Delete found but unreplaced characters.
2440 s Squash duplicate replaced characters.
8ff32507
FC
2441 r Return the modified string and leave the original string
2442 untouched.
a0d0e21e 2443
ba7f043c 2444If the C</c> modifier is specified, the I<SEARCHLIST> character set
19799a22 2445is complemented. If the C</d> modifier is specified, any characters
ba7f043c 2446specified by I<SEARCHLIST> not found in I<REPLACEMENTLIST> are deleted.
19799a22 2447(Note that this is slightly more flexible than the behavior of some
ba7f043c 2448B<tr> programs, which delete anything they find in the I<SEARCHLIST>,
46f8a5ea 2449period.) If the C</s> modifier is specified, sequences of characters
19799a22
GS
2450that were transliterated to the same character are squashed down
2451to a single instance of the character.
a0d0e21e 2452
ba7f043c
KW
2453If the C</d> modifier is used, the I<REPLACEMENTLIST> is always interpreted
2454exactly as specified. Otherwise, if the I<REPLACEMENTLIST> is shorter
2455than the I<SEARCHLIST>, the final character is replicated till it is long
2456enough. If the I<REPLACEMENTLIST> is empty, the I<SEARCHLIST> is replicated.
a0d0e21e
LW
2457This latter is useful for counting characters in a class or for
2458squashing character sequences in a class.
2459
2460Examples:
2461
c543c01b 2462 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case ASCII
a0d0e21e
LW
2463
2464 $cnt = tr/*/*/; # count the stars in $_
2465
2466 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
2467
2468 $cnt = tr/0-9//; # count the digits in $_
2469
2470 tr/a-zA-Z//s; # bookkeeper -> bokeper
2471
2472 ($HOST = $host) =~ tr/a-z/A-Z/;
c543c01b 2473 $HOST = $host =~ tr/a-z/A-Z/r; # same thing
8ff32507 2474
c543c01b 2475 $HOST = $host =~ tr/a-z/A-Z/r # chained with s///r
8ff32507 2476 =~ s/:/ -p/r;
a0d0e21e
LW
2477
2478 tr/a-zA-Z/ /cs; # change non-alphas to single space
2479
8ff32507
FC
2480 @stripped = map tr/a-zA-Z/ /csr, @original;
2481 # /r with map
2482
a0d0e21e 2483 tr [\200-\377]
c543c01b 2484 [\000-\177]; # wickedly delete 8th bit
a0d0e21e 2485
19799a22
GS
2486If multiple transliterations are given for a character, only the
2487first one is used:
748a9306
LW
2488
2489 tr/AAA/XYZ/
2490
2c268ad5 2491will transliterate any A to X.
748a9306 2492
19799a22 2493Because the transliteration table is built at compile time, neither
ba7f043c 2494the I<SEARCHLIST> nor the I<REPLACEMENTLIST> are subjected to double quote
19799a22 2495interpolation. That means that if you want to use variables, you
ba7f043c 2496must use an C<eval()>:
a0d0e21e
LW
2497
2498 eval "tr/$oldlist/$newlist/";
2499 die $@ if $@;
2500
2501 eval "tr/$oldlist/$newlist/, 1" or die $@;
2502
ba7f043c 2503=item C<< <<I<EOF> >>
d74e8afc 2504X<here-doc> X<heredoc> X<here-document> X<<< << >>>
7e3b091d
DA
2505
2506A line-oriented form of quoting is based on the shell "here-document"
2507syntax. Following a C<< << >> you specify a string to terminate
2508the quoted material, and all lines following the current line down to
89d205f2
YO
2509the terminating string are the value of the item.
2510
2511The terminating string may be either an identifier (a word), or some
2512quoted text. An unquoted identifier works like double quotes.
2513There may not be a space between the C<< << >> and the identifier,
2514unless the identifier is explicitly quoted. (If you put a space it
2515will be treated as a null identifier, which is valid, and matches the
2516first empty line.) The terminating string must appear by itself
2517(unquoted and with no surrounding whitespace) on the terminating line.
2518
2519If the terminating string is quoted, the type of quotes used determine
2520the treatment of the text.
2521
2522=over 4
2523
2524=item Double Quotes
2525
2526Double quotes indicate that the text will be interpolated using exactly
2527the same rules as normal double quoted strings.
7e3b091d
DA
2528
2529 print <<EOF;
2530 The price is $Price.
2531 EOF
2532
2533 print << "EOF"; # same as above
2534 The price is $Price.
2535 EOF
2536
89d205f2
YO
2537
2538=item Single Quotes
2539
2540Single quotes indicate the text is to be treated literally with no
46f8a5ea 2541interpolation of its content. This is similar to single quoted
89d205f2
YO
2542strings except that backslashes have no special meaning, with C<\\>
2543being treated as two backslashes and not one as they would in every
2544other quoting construct.
2545
c543c01b
TC
2546Just as in the shell, a backslashed bareword following the C<<< << >>>
2547means the same thing as a single-quoted string does:
2548
2549 $cost = <<'VISTA'; # hasta la ...
2550 That'll be $10 please, ma'am.
2551 VISTA
2552
2553 $cost = <<\VISTA; # Same thing!
2554 That'll be $10 please, ma'am.
2555 VISTA
2556
89d205f2
YO
2557This is the only form of quoting in perl where there is no need
2558to worry about escaping content, something that code generators
2559can and do make good use of.
2560
2561=item Backticks
2562
2563The content of the here doc is treated just as it would be if the
46f8a5ea 2564string were embedded in backticks. Thus the content is interpolated
89d205f2
YO
2565as though it were double quoted and then executed via the shell, with
2566the results of the execution returned.
2567
2568 print << `EOC`; # execute command and get results
7e3b091d 2569 echo hi there
7e3b091d
DA
2570 EOC
2571
89d205f2
YO
2572=back
2573
2574It is possible to stack multiple here-docs in a row:
2575
7e3b091d
DA
2576 print <<"foo", <<"bar"; # you can stack them
2577 I said foo.
2578 foo
2579 I said bar.
2580 bar
2581
2582 myfunc(<< "THIS", 23, <<'THAT');
2583 Here's a line
2584 or two.
2585 THIS
2586 and here's another.
2587 THAT
2588
2589Just don't forget that you have to put a semicolon on the end
2590to finish the statement, as Perl doesn't know you're not going to
2591try to do this:
2592
2593 print <<ABC
2594 179231
2595 ABC
2596 + 20;
2597
872d7e53
ST
2598If you want to remove the line terminator from your here-docs,
2599use C<chomp()>.
2600
2601 chomp($string = <<'END');
2602 This is a string.
2603 END
2604
2605If you want your here-docs to be indented with the rest of the code,
2606you'll need to remove leading whitespace from each line manually:
7e3b091d
DA
2607
2608 ($quote = <<'FINIS') =~ s/^\s+//gm;
89d205f2 2609 The Road goes ever on and on,
7e3b091d
DA
2610 down from the door where it began.
2611 FINIS
2612
2613If you use a here-doc within a delimited construct, such as in C<s///eg>,
1bf48760
FC
2614the quoted material must still come on the line following the
2615C<<< <<FOO >>> marker, which means it may be inside the delimited
2616construct:
7e3b091d
DA
2617
2618 s/this/<<E . 'that'
2619 the other
2620 E
2621 . 'more '/eg;
2622
1bf48760
FC
2623It works this way as of Perl 5.18. Historically, it was inconsistent, and
2624you would have to write
7e3b091d 2625
89d205f2
YO
2626 s/this/<<E . 'that'
2627 . 'more '/eg;
2628 the other
2629 E
7e3b091d 2630
1bf48760
FC
2631outside of string evals.
2632
c543c01b 2633Additionally, quoting rules for the end-of-string identifier are
46f8a5ea 2634unrelated to Perl's quoting rules. C<q()>, C<qq()>, and the like are not
89d205f2
YO
2635supported in place of C<''> and C<"">, and the only interpolation is for
2636backslashing the quoting character:
7e3b091d
DA
2637
2638 print << "abc\"def";
2639 testing...
2640 abc"def
2641
2642Finally, quoted strings cannot span multiple lines. The general rule is
2643that the identifier must be a string literal. Stick with that, and you
2644should be safe.
2645
a0d0e21e
LW
2646=back
2647
75e14d17 2648=head2 Gory details of parsing quoted constructs
d74e8afc 2649X<quote, gory details>
75e14d17 2650
19799a22
GS
2651When presented with something that might have several different
2652interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
2653principle to pick the most probable interpretation. This strategy
2654is so successful that Perl programmers often do not suspect the
2655ambivalence of what they write. But from time to time, Perl's
2656notions differ substantially from what the author honestly meant.
2657
2658This section hopes to clarify how Perl handles quoted constructs.
2659Although the most common reason to learn this is to unravel labyrinthine
2660regular expressions, because the initial steps of parsing are the
2661same for all quoting operators, they are all discussed together.
2662
2663The most important Perl parsing rule is the first one discussed
2664below: when processing a quoted construct, Perl first finds the end
2665of that construct, then interprets its contents. If you understand
2666this rule, you may skip the rest of this section on the first
2667reading. The other rules are likely to contradict the user's
2668expectations much less frequently than this first one.
2669
2670Some passes discussed below are performed concurrently, but because
2671their results are the same, we consider them individually. For different
2672quoting constructs, Perl performs different numbers of passes, from
6deea57f 2673one to four, but these passes are always performed in the same order.
75e14d17 2674
13a2d996 2675=over 4
75e14d17
IZ
2676
2677=item Finding the end
2678
ba7f043c
KW
2679The first pass is finding the end of the quoted construct. This results
2680in saving to a safe location a copy of the text (between the starting
2681and ending delimiters), normalized as necessary to avoid needing to know
2682what the original delimiters were.
6deea57f
ST
2683
2684If the construct is a here-doc, the ending delimiter is a line
46f8a5ea 2685that has a terminating string as the content. Therefore C<<<EOF> is
6deea57f
ST
2686terminated by C<EOF> immediately followed by C<"\n"> and starting
2687from the first column of the terminating line.
2688When searching for the terminating line of a here-doc, nothing
46f8a5ea 2689is skipped. In other words, lines after the here-doc syntax
6deea57f
ST
2690are compared with the terminating string line by line.
2691
2692For the constructs except here-docs, single characters are used as starting
46f8a5ea 2693and ending delimiters. If the starting delimiter is an opening punctuation
6deea57f
ST
2694(that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
2695corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
2696If the starting delimiter is an unpaired character like C</> or a closing
ba7f043c 2697punctuation, the ending delimiter is the same as the starting delimiter.
6deea57f 2698Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
fc693347 2699both C<qq[]> and C<qq]]> constructs.
6deea57f
ST
2700
2701When searching for single-character delimiters, escaped delimiters
1ca345ed 2702and C<\\> are skipped. For example, while searching for terminating C</>,
6deea57f
ST
2703combinations of C<\\> and C<\/> are skipped. If the delimiters are
2704bracketing, nested pairs are also skipped. For example, while searching
ba7f043c 2705for a closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
6deea57f
ST
2706and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
2707However, when backslashes are used as the delimiters (like C<qq\\> and
2708C<tr\\\>), nothing is skipped.
32581033 2709During the search for the end, backslashes that escape delimiters or
7188ca43 2710other backslashes are removed (exactly speaking, they are not copied to the
32581033 2711safe location).
75e14d17 2712
19799a22
GS
2713For constructs with three-part delimiters (C<s///>, C<y///>, and
2714C<tr///>), the search is repeated once more.
fc693347 2715If the first delimiter is not an opening punctuation, the three delimiters must
d74605e5
FC
2716be the same, such as C<s!!!> and C<tr)))>,
2717in which case the second delimiter
6deea57f 2718terminates the left part and starts the right part at once.
b6538e4f 2719If the left part is delimited by bracketing punctuation (that is C<()>,
6deea57f 2720C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
b6538e4f 2721delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespace
ba7f043c 2722and comments are allowed between the two parts, although the comment must follow
b6538e4f
TC
2723at least one whitespace character; otherwise a character expected as the
2724start of the comment may be regarded as the starting delimiter of the right part.
75e14d17 2725
19799a22
GS
2726During this search no attention is paid to the semantics of the construct.
2727Thus:
75e14d17
IZ
2728
2729 "$hash{"$foo/$bar"}"
2730
2a94b7ce 2731or:
75e14d17 2732
89d205f2 2733 m/
2a94b7ce 2734 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
2735 /x
2736
19799a22
GS
2737do not form legal quoted expressions. The quoted part ends on the
2738first C<"> and C</>, and the rest happens to be a syntax error.
2739Because the slash that terminated C<m//> was followed by a C<SPACE>,
2740the example above is not C<m//x>, but rather C<m//> with no C</x>
2741modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 2742
89d205f2 2743Also no attention is paid to C<\c\> (multichar control char syntax) during
46f8a5ea 2744this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
89d205f2 2745of C<\/>, and the following C</> is not recognized as a delimiter.
0d594e51
ST
2746Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
2747
75e14d17 2748=item Interpolation
d74e8afc 2749X<interpolation>
75e14d17 2750
19799a22 2751The next step is interpolation in the text obtained, which is now
89d205f2 2752delimiter-independent. There are multiple cases.
75e14d17 2753
13a2d996 2754=over 4
75e14d17 2755
89d205f2 2756=item C<<<'EOF'>
75e14d17
IZ
2757
2758No interpolation is performed.
6deea57f
ST
2759Note that the combination C<\\> is left intact, since escaped delimiters
2760are not available for here-docs.
75e14d17 2761
6deea57f 2762=item C<m''>, the pattern of C<s'''>
89d205f2 2763
6deea57f
ST
2764No interpolation is performed at this stage.
2765Any backslashed sequences including C<\\> are treated at the stage
2766to L</"parsing regular expressions">.
89d205f2 2767
6deea57f 2768=item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
75e14d17 2769
89d205f2 2770The only interpolation is removal of C<\> from pairs of C<\\>.
ba7f043c 2771Therefore C<"-"> in C<tr'''> and C<y'''> is treated literally
6deea57f
ST
2772as a hyphen and no character range is available.
2773C<\1> in the replacement of C<s'''> does not work as C<$1>.
89d205f2
YO
2774
2775=item C<tr///>, C<y///>
2776
6deea57f
ST
2777No variable interpolation occurs. String modifying combinations for
2778case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
2779The other escape sequences such as C<\200> and C<\t> and backslashed
2780characters such as C<\\> and C<\-> are converted to appropriate literals.
ba7f043c
KW
2781The character C<"-"> is treated specially and therefore C<\-> is treated
2782as a literal C<"-">.
75e14d17 2783
89d205f2 2784=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
75e14d17 2785
628253b8 2786C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> (possibly paired with C<\E>) are
19799a22 2787converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
ba7f043c 2788is converted to S<C<$foo . (quotemeta("baz" . $bar))>> internally.
6deea57f
ST
2789The other escape sequences such as C<\200> and C<\t> and backslashed
2790characters such as C<\\> and C<\-> are replaced with appropriate
2791expansions.
2a94b7ce 2792
19799a22
GS
2793Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
2794is interpolated in the usual way. Something like C<"\Q\\E"> has
48cbae4f 2795no C<\E> inside. Instead, it has C<\Q>, C<\\>, and C<E>, so the
19799a22
GS
2796result is the same as for C<"\\\\E">. As a general rule, backslashes
2797between C<\Q> and C<\E> may lead to counterintuitive results. So,
2798C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
2799as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
2800
2801 $str = '\t';
2802 return "\Q$str";
2803
2804may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
2805
19799a22 2806Interpolated scalars and arrays are converted internally to the C<join> and
ba7f043c 2807C<"."> catenation operations. Thus, S<C<"$foo XXX '@arr'">> becomes:
75e14d17 2808
19799a22 2809 $foo . " XXX '" . (join $", @arr) . "'";
75e14d17 2810
19799a22 2811All operations above are performed simultaneously, left to right.
75e14d17 2812
ba7f043c 2813Because the result of S<C<"\Q I<STRING> \E">> has all metacharacters
19799a22 2814quoted, there is no way to insert a literal C<$> or C<@> inside a
ba7f043c 2815C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to become
19799a22
GS
2816C<"\\\$">; if not, it is interpreted as the start of an interpolated
2817scalar.
75e14d17 2818
19799a22 2819Note also that the interpolation code needs to make a decision on
89d205f2 2820where the interpolated scalar ends. For instance, whether
ba7f043c 2821S<C<< "a $x -> {c}" >>> really means:
75e14d17 2822
db691027 2823 "a " . $x . " -> {c}";
75e14d17 2824
2a94b7ce 2825or:
75e14d17 2826
db691027 2827 "a " . $x -> {c};
75e14d17 2828
19799a22
GS
2829Most of the time, the longest possible text that does not include
2830spaces between components and which contains matching braces or
2831brackets. because the outcome may be determined by voting based
2832on heuristic estimators, the result is not strictly predictable.
2833Fortunately, it's usually correct for ambiguous cases.
75e14d17 2834
6deea57f 2835=item the replacement of C<s///>
75e14d17 2836
628253b8 2837Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> and interpolation
6deea57f
ST
2838happens as with C<qq//> constructs.
2839
2840It is at this step that C<\1> is begrudgingly converted to C<$1> in
2841the replacement text of C<s///>, in order to correct the incorrigible
2842I<sed> hackers who haven't picked up the saner idiom yet. A warning
ba7f043c 2843is emitted if the S<C<use warnings>> pragma or the B<-w> command-line flag
6deea57f
ST
2844(that is, the C<$^W> variable) was set.
2845
2846=item C<RE> in C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
2847
628253b8 2848Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F>, C<\E>,
cc74c5bd
ST
2849and interpolation happens (almost) as with C<qq//> constructs.
2850
5d03b57c
KW
2851Processing of C<\N{...}> is also done here, and compiled into an intermediate
2852form for the regex compiler. (This is because, as mentioned below, the regex
2853compilation may be done at execution time, and C<\N{...}> is a compile-time
2854construct.)
2855
cc74c5bd
ST
2856However any other combinations of C<\> followed by a character
2857are not substituted but only skipped, in order to parse them
2858as regular expressions at the following step.
6deea57f 2859As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
1749ea0d 2860treated as an array symbol (for example C<@foo>),
6deea57f 2861even though the same text in C<qq//> gives interpolation of C<\c@>.
6deea57f 2862
e128ab2c
DM
2863Code blocks such as C<(?{BLOCK})> are handled by temporarily passing control
2864back to the perl parser, in a similar way that an interpolated array
2865subscript expression such as C<"foo$array[1+f("[xyz")]bar"> would be.
2866
ba7f043c
KW
2867Moreover, inside C<(?{BLOCK})>, S<C<(?# comment )>>, and
2868a C<#>-comment in a C</x>-regular expression, no processing is
19799a22 2869performed whatsoever. This is the first step at which the presence
ba7f043c 2870of the C</x> modifier is relevant.
19799a22 2871
1749ea0d
ST
2872Interpolation in patterns has several quirks: C<$|>, C<$(>, C<$)>, C<@+>
2873and C<@-> are not interpolated, and constructs C<$var[SOMETHING]> are
2874voted (by several different estimators) to be either an array element
2875or C<$var> followed by an RE alternative. This is where the notation
19799a22
GS
2876C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
2877array element C<-9>, not as a regular expression from the variable
2878C<$arr> followed by a digit, which would be the interpretation of
2879C</$arr[0-9]/>. Since voting among different estimators may occur,
2880the result is not predictable.
2881
19799a22
GS
2882The lack of processing of C<\\> creates specific restrictions on
2883the post-processed text. If the delimiter is C</>, one cannot get
2884the combination C<\/> into the result of this step. C</> will
2885finish the regular expression, C<\/> will be stripped to C</> on
2886the previous step, and C<\\/> will be left as is. Because C</> is
2887equivalent to C<\/> inside a regular expression, this does not
2888matter unless the delimiter happens to be character special to the
2889RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
2890alphanumeric char, as in:
2a94b7ce
IZ
2891
2892 m m ^ a \s* b mmx;
2893
19799a22 2894In the RE above, which is intentionally obfuscated for illustration, the
6deea57f 2895delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
ba7f043c 2896RE is the same as for S<C<m/ ^ a \s* b /mx>>. There's more than one
19799a22
GS
2897reason you're encouraged to restrict your delimiters to non-alphanumeric,
2898non-whitespace choices.
75e14d17
IZ
2899
2900=back
2901
19799a22 2902This step is the last one for all constructs except regular expressions,
75e14d17
IZ
2903which are processed further.
2904
6deea57f
ST
2905=item parsing regular expressions
2906X<regexp, parse>
75e14d17 2907
19799a22 2908Previous steps were performed during the compilation of Perl code,
ac036724 2909but this one happens at run time, although it may be optimized to
19799a22 2910be calculated at compile time if appropriate. After preprocessing
6deea57f 2911described above, and possibly after evaluation if concatenation,
19799a22
GS
2912joining, casing translation, or metaquoting are involved, the
2913resulting I<string> is passed to the RE engine for compilation.
2914
2915Whatever happens in the RE engine might be better discussed in L<perlre>,
2916but for the sake of continuity, we shall do so here.
2917
ba7f043c 2918This is another step where the presence of the C</x> modifier is
19799a22 2919relevant. The RE engine scans the string from left to right and
ba7f043c 2920converts it into a finite automaton.
19799a22
GS
2921
2922Backslashed characters are either replaced with corresponding
2923literal strings (as with C<\{>), or else they generate special nodes
2924in the finite automaton (as with C<\b>). Characters special to the
2925RE engine (such as C<|>) generate corresponding nodes or groups of
2926nodes. C<(?#...)> comments are ignored. All the rest is either
2927converted to literal strings to match, or else is ignored (as is
ba7f043c 2928whitespace and C<#>-style comments if C</x> is present).
19799a22
GS
2929
2930Parsing of the bracketed character class construct, C<[...]>, is
2931rather different than the rule used for the rest of the pattern.
2932The terminator of this construct is found using the same rules as
2933for finding the terminator of a C<{}>-delimited construct, the only
2934exception being that C<]> immediately following C<[> is treated as
e128ab2c
DM
2935though preceded by a backslash.
2936
2937The terminator of runtime C<(?{...})> is found by temporarily switching
2938control to the perl parser, which should stop at the point where the
2939logically balancing terminating C<}> is found.
19799a22
GS
2940
2941It is possible to inspect both the string given to RE engine and the
2942resulting finite automaton. See the arguments C<debug>/C<debugcolor>
ba7f043c 2943in the S<C<use L<re>>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 2944switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
2945
2946=item Optimization of regular expressions
d74e8afc 2947X<regexp, optimization>
75e14d17 2948
7522fed5 2949This step is listed for completeness only. Since it does not change
75e14d17 2950semantics, details of this step are not documented and are subject
19799a22
GS
2951to change without notice. This step is performed over the finite
2952automaton that was generated during the previous pass.
2a94b7ce 2953
19799a22
GS
2954It is at this stage that C<split()> silently optimizes C</^/> to
2955mean C</^/m>.
75e14d17
IZ
2956
2957=back
2958
a0d0e21e 2959=head2 I/O Operators
d74e8afc 2960X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
80a96bfc 2961X<< <> >> X<< <<>> >> X<@ARGV>
a0d0e21e 2962
54310121 2963There are several I/O operators you should know about.
fbad3eb5 2964
7b8d334a 2965A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
2966double-quote interpolation. It is then interpreted as an external
2967command, and the output of that command is the value of the
e9c56f9b
JH
2968backtick string, like in a shell. In scalar context, a single string
2969consisting of all output is returned. In list context, a list of
2970values is returned, one per line of output. (You can set C<$/> to use
2971a different line terminator.) The command is executed each time the
2972pseudo-literal is evaluated. The status value of the command is
2973returned in C<$?> (see L<perlvar> for the interpretation of C<$?>).
2974Unlike in B<csh>, no translation is done on the return data--newlines
2975remain newlines. Unlike in any of the shells, single quotes do not
2976hide variable names in the command from interpretation. To pass a
2977literal dollar-sign through to the shell you need to hide it with a
2978backslash. The generalized form of backticks is C<qx//>. (Because
2979backticks always undergo shell expansion as well, see L<perlsec> for
2980security concerns.)
d74e8afc 2981X<qx> X<`> X<``> X<backtick> X<glob>
19799a22
GS
2982
2983In scalar context, evaluating a filehandle in angle brackets yields
2984the next line from that file (the newline, if any, included), or
2985C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
2986(sometimes known as file-slurp mode) and the file is empty, it
2987returns C<''> the first time, followed by C<undef> subsequently.
2988
2989Ordinarily you must assign the returned value to a variable, but
2990there is one situation where an automatic assignment happens. If
2991and only if the input symbol is the only thing inside the conditional
2992of a C<while> statement (even if disguised as a C<for(;;)> loop),
ba7f043c 2993the value is automatically assigned to the global variable C<$_>,
19799a22
GS
2994destroying whatever was there previously. (This may seem like an
2995odd thing to you, but you'll use the construct in almost every Perl
ba7f043c
KW
2996script you write.) The C<$_> variable is not implicitly localized.
2997You'll have to put a S<C<local $_;>> before the loop if you want that
19799a22
GS
2998to happen.
2999
3000The following lines are equivalent:
a0d0e21e 3001
748a9306 3002 while (defined($_ = <STDIN>)) { print; }
7b8d334a 3003 while ($_ = <STDIN>) { print; }
a0d0e21e
LW
3004 while (<STDIN>) { print; }
3005 for (;<STDIN>;) { print; }
748a9306 3006 print while defined($_ = <STDIN>);
7b8d334a 3007 print while ($_ = <STDIN>);
a0d0e21e
LW
3008 print while <STDIN>;
3009
1ca345ed
TC
3010This also behaves similarly, but assigns to a lexical variable
3011instead of to C<$_>:
7b8d334a 3012
89d205f2 3013 while (my $line = <STDIN>) { print $line }
7b8d334a 3014
19799a22
GS
3015In these loop constructs, the assigned value (whether assignment
3016is automatic or explicit) is then tested to see whether it is
1ca345ed
TC
3017defined. The defined test avoids problems where the line has a string
3018value that would be treated as false by Perl; for example a "" or
ba7f043c 3019a C<"0"> with no trailing newline. If you really mean for such values
19799a22 3020to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
3021
3022 while (($_ = <STDIN>) ne '0') { ... }
3023 while (<STDIN>) { last unless $_; ... }
3024
ba7f043c 3025In other boolean contexts, C<< <I<FILEHANDLE>> >> without an
5ef4d93e 3026explicit C<defined> test or comparison elicits a warning if the
ba7f043c 3027S<C<use warnings>> pragma or the B<-w>
19799a22 3028command-line switch (the C<$^W> variable) is in effect.
7b8d334a 3029
5f05dabc 3030The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
3031filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
3032in packages, where they would be interpreted as local identifiers
3033rather than global.) Additional filehandles may be created with
ba7f043c 3034the C<open()> function, amongst others. See L<perlopentut> and
19799a22 3035L<perlfunc/open> for details on this.
d74e8afc 3036X<stdin> X<stdout> X<sterr>
a0d0e21e 3037
ba7f043c 3038If a C<< <I<FILEHANDLE>> >> is used in a context that is looking for
19799a22
GS
3039a list, a list comprising all input lines is returned, one line per
3040list element. It's easy to grow to a rather large data space this
3041way, so use with care.
a0d0e21e 3042
ba7f043c 3043C<< <I<FILEHANDLE>> >> may also be spelled C<readline(*I<FILEHANDLE>)>.
19799a22 3044See L<perlfunc/readline>.
fbad3eb5 3045
ba7f043c 3046The null filehandle C<< <> >> is special: it can be used to emulate the
1ca345ed
TC
3047behavior of B<sed> and B<awk>, and any other Unix filter program
3048that takes a list of filenames, doing the same to each line
ba7f043c 3049of input from all of them. Input from C<< <> >> comes either from
a0d0e21e 3050standard input, or from each file listed on the command line. Here's
ba7f043c
KW
3051how it works: the first time C<< <> >> is evaluated, the C<@ARGV> array is
3052checked, and if it is empty, C<$ARGV[0]> is set to C<"-">, which when opened
3053gives you standard input. The C<@ARGV> array is then processed as a list
a0d0e21e
LW
3054of filenames. The loop
3055
3056 while (<>) {
3057 ... # code for each line
3058 }
3059
3060is equivalent to the following Perl-like pseudo code:
3061
3e3baf6d 3062 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
3063 while ($ARGV = shift) {
3064 open(ARGV, $ARGV);
3065 while (<ARGV>) {
3066 ... # code for each line
3067 }
3068 }
3069
19799a22 3070except that it isn't so cumbersome to say, and will actually work.
ba7f043c
KW
3071It really does shift the C<@ARGV> array and put the current filename
3072into the C<$ARGV> variable. It also uses filehandle I<ARGV>
3073internally. C<< <> >> is just a synonym for C<< <ARGV> >>, which
19799a22 3074is magical. (The pseudo code above doesn't work because it treats
ba7f043c 3075C<< <ARGV> >> as non-magical.)
a0d0e21e 3076
48ab5743
ML
3077Since the null filehandle uses the two argument form of L<perlfunc/open>
3078it interprets special characters, so if you have a script like this:
3079
3080 while (<>) {
3081 print;
3082 }
3083
ba7f043c 3084and call it with S<C<perl dangerous.pl 'rm -rfv *|'>>, it actually opens a
48ab5743
ML
3085pipe, executes the C<rm> command and reads C<rm>'s output from that pipe.
3086If you want all items in C<@ARGV> to be interpreted as file names, you
1033ba6e
PM
3087can use the module C<ARGV::readonly> from CPAN, or use the double bracket:
3088
3089 while (<<>>) {
3090 print;
3091 }
3092
3093Using double angle brackets inside of a while causes the open to use the
3094three argument form (with the second argument being C<< < >>), so all
ba7f043c
KW
3095arguments in C<ARGV> are treated as literal filenames (including C<"-">).
3096(Note that for convenience, if you use C<< <<>> >> and if C<@ARGV> is
80a96bfc 3097empty, it will still read from the standard input.)
48ab5743 3098
ba7f043c 3099You can modify C<@ARGV> before the first C<< <> >> as long as the array ends up
a0d0e21e 3100containing the list of filenames you really want. Line numbers (C<$.>)
19799a22
GS
3101continue as though the input were one big happy file. See the example
3102in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20 3103
ba7f043c
KW
3104If you want to set C<@ARGV> to your own list of files, go right ahead.
3105This sets C<@ARGV> to all plain text files if no C<@ARGV> was given:
5a964f20
TC
3106
3107 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 3108
5a964f20
TC
3109You can even set them to pipe commands. For example, this automatically
3110filters compressed arguments through B<gzip>:
3111
3112 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
3113
3114If you want to pass switches into your script, you can use one of the
ba7f043c 3115C<Getopts> modules or put a loop on the front like this:
a0d0e21e
LW
3116
3117 while ($_ = $ARGV[0], /^-/) {
3118 shift;
3119 last if /^--$/;
3120 if (/^-D(.*)/) { $debug = $1 }
3121 if (/^-v/) { $verbose++ }
5a964f20 3122 # ... # other switches
a0d0e21e 3123 }
5a964f20 3124
a0d0e21e 3125 while (<>) {
5a964f20 3126 # ... # code for each line
a0d0e21e
LW
3127 }
3128
ba7f043c 3129The C<< <> >> symbol will return C<undef> for end-of-file only once.
89d205f2 3130If you call it again after this, it will assume you are processing another
ba7f043c 3131C<@ARGV> list, and if you haven't set C<@ARGV>, will read input from STDIN.
a0d0e21e 3132
1ca345ed 3133If what the angle brackets contain is a simple scalar variable (for example,
ba7f043c 3134C<$foo>), then that variable contains the name of the
19799a22
GS
3135filehandle to input from, or its typeglob, or a reference to the
3136same. For example:
cb1a09d0
AD
3137
3138 $fh = \*STDIN;
3139 $line = <$fh>;
a0d0e21e 3140
5a964f20
TC
3141If what's within the angle brackets is neither a filehandle nor a simple
3142scalar variable containing a filehandle name, typeglob, or typeglob
3143reference, it is interpreted as a filename pattern to be globbed, and
3144either a list of filenames or the next filename in the list is returned,
19799a22 3145depending on context. This distinction is determined on syntactic
ba7f043c
KW
3146grounds alone. That means C<< <$x> >> is always a C<readline()> from
3147an indirect handle, but C<< <$hash{key}> >> is always a C<glob()>.
3148That's because C<$x> is a simple scalar variable, but C<$hash{key}> is
ef191992
YST
3149not--it's a hash element. Even C<< <$x > >> (note the extra space)
3150is treated as C<glob("$x ")>, not C<readline($x)>.
5a964f20
TC
3151
3152One level of double-quote interpretation is done first, but you can't
35f2feb0 3153say C<< <$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
3154in the previous paragraph. (In older versions of Perl, programmers
3155would insert curly brackets to force interpretation as a filename glob:
35f2feb0 3156C<< <${foo}> >>. These days, it's considered cleaner to call the
5a964f20 3157internal function directly as C<glob($foo)>, which is probably the right
19799a22 3158way to have done it in the first place.) For example:
a0d0e21e
LW
3159
3160 while (<*.c>) {
3161 chmod 0644, $_;
3162 }
3163
3a4b19e4 3164is roughly equivalent to:
a0d0e21e
LW
3165
3166 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
3167 while (<FOO>) {
5b3eff12 3168 chomp;
a0d0e21e
LW
3169 chmod 0644, $_;
3170 }
3171
3a4b19e4 3172except that the globbing is actually done internally using the standard
ba7f043c 3173C<L<File::Glob>> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
3174
3175 chmod 0644, <*.c>;
3176
19799a22
GS
3177A (file)glob evaluates its (embedded) argument only when it is
3178starting a new list. All values must be read before it will start
3179over. In list context, this isn't important because you automatically
3180get them all anyway. However, in scalar context the operator returns
069e01df 3181the next value each time it's called, or C<undef> when the list has
19799a22
GS
3182run out. As with filehandle reads, an automatic C<defined> is
3183generated when the glob occurs in the test part of a C<while>,
1ca345ed
TC
3184because legal glob returns (for example,
3185a file called F<0>) would otherwise
19799a22
GS
3186terminate the loop. Again, C<undef> is returned only once. So if
3187you're expecting a single value from a glob, it is much better to
3188say
4633a7c4
LW
3189
3190 ($file) = <blurch*>;
3191
3192than
3193
3194 $file = <blurch*>;
3195
3196because the latter will alternate between returning a filename and
19799a22 3197returning false.
4633a7c4 3198
b159ebd3 3199If you're trying to do variable interpolation, it's definitely better
ba7f043c 3200to use the C<glob()> function, because the older notation can cause people
e37d713d 3201to become confused with the indirect filehandle notation.
4633a7c4
LW
3202
3203 @files = glob("$dir/*.[ch]");
3204 @files = glob($files[$i]);
3205
a0d0e21e 3206=head2 Constant Folding
d74e8afc 3207X<constant folding> X<folding>
a0d0e21e
LW
3208
3209Like C, Perl does a certain amount of expression evaluation at
19799a22 3210compile time whenever it determines that all arguments to an
a0d0e21e
LW
3211operator are static and have no side effects. In particular, string
3212concatenation happens at compile time between literals that don't do
19799a22 3213variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
3214compile time. You can say
3215
1ca345ed
TC
3216 'Now is the time for all'
3217 . "\n"
3218 . 'good men to come to.'
a0d0e21e 3219
54310121 3220and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
3221you say
3222
3223 foreach $file (@filenames) {
5a964f20 3224 if (-s $file > 5 + 100 * 2**16) { }
54310121 3225 }
a0d0e21e 3226
1ca345ed 3227the compiler precomputes the number which that expression
19799a22 3228represents so that the interpreter won't have to.
a0d0e21e 3229
fd1abbef 3230=head2 No-ops
d74e8afc 3231X<no-op> X<nop>
fd1abbef
DN
3232
3233Perl doesn't officially have a no-op operator, but the bare constants
1ca345ed 3234C<0> and C<1> are special-cased not to produce a warning in void
fd1abbef
DN
3235context, so you can for example safely do
3236
3237 1 while foo();
3238
2c268ad5 3239=head2 Bitwise String Operators
fb7054ba 3240X<operator, bitwise, string> X<&.> X<|.> X<^.> X<~.>
2c268ad5
TP
3241
3242Bitstrings of any size may be manipulated by the bitwise operators
3243(C<~ | & ^>).
3244
19799a22
GS
3245If the operands to a binary bitwise op are strings of different
3246sizes, B<|> and B<^> ops act as though the shorter operand had
3247additional zero bits on the right, while the B<&> op acts as though
3248the longer operand were truncated to the length of the shorter.
3249The granularity for such extension or truncation is one or more
3250bytes.
2c268ad5 3251
89d205f2 3252 # ASCII-based examples
2c268ad5
TP
3253 print "j p \n" ^ " a h"; # prints "JAPH\n"
3254 print "JA" | " ph\n"; # prints "japh\n"
3255 print "japh\nJunk" & '_____'; # prints "JAPH\n";
3256 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
3257
19799a22