This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlop: Note that negative rep now warns
[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
ML
8In Perl, the operator determines what operation is performed,
9independent of the type of the operands. For example C<$a + $b>
10is always a numeric addition, and if C<$a> or C<$b> do not contain
11numbers, an attempt is made to convert them to numbers first.
12
13This is in contrast to many other dynamic languages, where the
14operation is determined by the type of the first argument. It also
15means that Perl has two versions of some operators, one for numeric
16and one for string comparison. For example C<$a == $b> compares
17two numbers for equality, and C<$a eq $b> compares two strings.
18
19There are a few exceptions though: C<x> can be either string
20repetition or list repetition, depending on the type of the left
21operand, and C<&>, C<|> and C<^> can be either string or numeric bit
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
31others. For example, in C<2 + 4 * 5>, the multiplication has higher
32precedence so C<4 * 5> is evaluated first yielding C<2 + 20 ==
3322> and not C<6 * 5 == 30>.
34
35I<Operator associativity> defines what happens if a sequence of the
36same operators is used one after another: whether the evaluator will
37evaluate the left operations first or the right. For example, in C<8
38- 4 - 2>, subtraction is left associative so Perl evaluates the
39expression left to right. C<8 - 4> is evaluated first making the
40expression C<4 - 2 == 2> and not 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
88If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
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
19799a22
GS
101the commas on the right of the sort are evaluated before the sort,
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
122the result of C<$foo & 255>). Then one is added to the return value
123of C<print> (usually 1). The result is something like this:
124
125 1 + 1, "\n"; # Obviously not what you meant.
126
127To do what you meant properly, you must write:
128
129 print(($foo & 255) + 1, "\n");
130
131See L<Named Unary Operators> for more discussion of this.
a0d0e21e
LW
132
133Also parsed as terms are the C<do {}> and 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
RS
156The dereferencing cases (as opposed to method-calling cases) are
157somewhat extended by the experimental C<postderef> feature. For the
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
d042e63d
MS
163"++" and "--" work as in C. That is, if placed before a variable,
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
89d205f2 173incremented or decremented. You just know it will be done sometime
b033823e 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
19799a22 205Binary "**" is the exponentiation operator. It binds even more
cb1a09d0
AD
206tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
207implemented using C's pow(3) function, which actually works on doubles
208internally.)
a0d0e21e
LW
209
210=head2 Symbolic Unary Operators
d74e8afc 211X<unary operator> X<operator, unary>
a0d0e21e 212
1ca345ed 213Unary "!" performs logical negation, that is, "not". See also C<not> for a lower
a0d0e21e 214precedence version of this.
d74e8afc 215X<!>
a0d0e21e 216
da2f94c5
FC
217Unary "-" performs arithmetic negation if the operand is numeric,
218including any string that looks like a number. If the operand is
219an identifier, a string consisting of a minus sign concatenated
220with the identifier is returned. Otherwise, if the string starts
221with a plus or minus, a string starting with the opposite sign is
222returned. One effect of these rules is that -bareword is equivalent
8705167b 223to the string "-bareword". If, however, the string begins with a
353c6505 224non-alphabetic character (excluding "+" or "-"), Perl will attempt to convert
06705523
SP
225the string to a numeric and the arithmetic negation is performed. If the
226string cannot be cleanly converted to a numeric, Perl will give the warning
227B<Argument "the string" isn't numeric in negation (-) at ...>.
d74e8afc 228X<-> X<negation, arithmetic>
a0d0e21e 229
1ca345ed 230Unary "~" performs bitwise negation, that is, 1's complement. For
972b05a9
JH
231example, C<0666 & ~027> is 0640. (See also L<Integer Arithmetic> and
232L<Bitwise String Operators>.) Note that the width of the result is
233platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
234bits wide on a 64-bit platform, so if you are expecting a certain bit
f113cf86 235width, remember to use the "&" operator to mask off the excess bits.
d74e8afc 236X<~> X<negation, binary>
a0d0e21e 237
f113cf86
TC
238When complementing strings, if all characters have ordinal values under
239256, then their complements will, also. But if they do not, all
240characters will be in either 32- or 64-bit complements, depending on your
241architecture. So for example, C<~"\x{3B1}"> is C<"\x{FFFF_FC4E}"> on
24232-bit machines and C<"\x{FFFF_FFFF_FFFF_FC4E}"> on 64-bit machines.
243
a0d0e21e
LW
244Unary "+" has no effect whatsoever, even on strings. It is useful
245syntactically for separating a function name from a parenthesized expression
246that would otherwise be interpreted as the complete list of function
5ba421f6 247arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
d74e8afc 248X<+>
a0d0e21e 249
19799a22
GS
250Unary "\" creates a reference to whatever follows it. See L<perlreftut>
251and L<perlref>. Do not confuse this behavior with the behavior of
252backslash within a string, although both forms do convey the notion
253of protecting the next thing from interpolation.
d74e8afc 254X<\> X<reference> X<backslash>
a0d0e21e
LW
255
256=head2 Binding Operators
d74e8afc 257X<binding> X<operator, binding> X<=~> X<!~>
a0d0e21e 258
c07a80fd 259Binary "=~" binds a scalar expression to a pattern match. Certain operations
cb1a09d0
AD
260search or modify the string $_ by default. This operator makes that kind
261of operation work on some other string. The right argument is a search
2c268ad5
TP
262pattern, substitution, or transliteration. The left argument is what is
263supposed to be searched, substituted, or transliterated instead of the default
f8bab1e9 264$_. When used in scalar context, the return value generally indicates the
8ff32507
FC
265success of the operation. The exceptions are substitution (s///)
266and transliteration (y///) with the C</r> (non-destructive) option,
267which cause the B<r>eturn value to be the result of the substitution.
268Behavior in list context depends on the particular operator.
000c65fc
DG
269See L</"Regexp Quote-Like Operators"> for details and L<perlretut> for
270examples using these operators.
f8bab1e9
GS
271
272If the right argument is an expression rather than a search pattern,
2c268ad5 273substitution, or transliteration, it is interpreted as a search pattern at run
89d205f2
YO
274time. Note that this means that its contents will be interpolated twice, so
275
1ca345ed 276 '\\' =~ q'\\';
89d205f2
YO
277
278is not ok, as the regex engine will end up trying to compile the
279pattern C<\>, which it will consider a syntax error.
a0d0e21e
LW
280
281Binary "!~" is just like "=~" except the return value is negated in
282the logical sense.
283
8ff32507
FC
284Binary "!~" with a non-destructive substitution (s///r) or transliteration
285(y///r) is a syntax error.
4f4d7508 286
a0d0e21e 287=head2 Multiplicative Operators
d74e8afc 288X<operator, multiplicative>
a0d0e21e
LW
289
290Binary "*" multiplies two numbers.
d74e8afc 291X<*>
a0d0e21e
LW
292
293Binary "/" divides two numbers.
d74e8afc 294X</> X<slash>
a0d0e21e 295
f7918450
KW
296Binary "%" is the modulo operator, which computes the division
297remainder of its first argument with respect to its second argument.
298Given integer
54310121 299operands C<$a> and C<$b>: If C<$b> is positive, then C<$a % $b> is
f7918450 300C<$a> minus the largest multiple of C<$b> less than or equal to
54310121 301C<$a>. If C<$b> is negative, then C<$a % $b> is C<$a> minus the
1ca345ed 302smallest multiple of C<$b> that is not less than C<$a> (that is, the
89b4f0ad 303result will be less than or equal to zero). If the operands
4848a83b
ST
304C<$a> and C<$b> are floating point values and the absolute value of
305C<$b> (that is C<abs($b)>) is less than C<(UV_MAX + 1)>, only
306the integer portion of C<$a> and C<$b> will be used in the operation
307(Note: here C<UV_MAX> means the maximum of the unsigned integer type).
308If the absolute value of the right operand (C<abs($b)>) is greater than
309or equal to C<(UV_MAX + 1)>, "%" computes the floating-point remainder
310C<$r> in the equation C<($r = $a - $i*$b)> where C<$i> is a certain
f7918450 311integer that makes C<$r> have the same sign as the right operand
4848a83b
ST
312C<$b> (B<not> as the left operand C<$a> like C function C<fmod()>)
313and the absolute value less than that of C<$b>.
0412d526 314Note that when C<use integer> is in scope, "%" gives you direct access
f7918450 315to the modulo operator as implemented by your C compiler. This
55d729e4
GS
316operator is not as well defined for negative operands, but it will
317execute faster.
f7918450 318X<%> X<remainder> X<modulo> X<mod>
55d729e4 319
62d10b70
GS
320Binary "x" is the repetition operator. In scalar context or if the left
321operand is not enclosed in parentheses, it returns a string consisting
322of the left operand repeated the number of times specified by the right
323operand. In list context, if the left operand is enclosed in
3585017f 324parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
31201a8e
KW
325If the right operand is zero or negative (raising a warning on
326negative), it returns an empty string
3585017f 327or an empty list, depending on the context.
d74e8afc 328X<x>
a0d0e21e
LW
329
330 print '-' x 80; # print row of dashes
331
332 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
333
334 @ones = (1) x 80; # a list of 80 1's
335 @ones = (5) x @ones; # set all elements to 5
336
337
338=head2 Additive Operators
d74e8afc 339X<operator, additive>
a0d0e21e 340
1ca345ed 341Binary C<+> returns the sum of two numbers.
d74e8afc 342X<+>
a0d0e21e 343
1ca345ed 344Binary C<-> returns the difference of two numbers.
d74e8afc 345X<->
a0d0e21e 346
1ca345ed 347Binary C<.> concatenates two strings.
d74e8afc
ITB
348X<string, concatenation> X<concatenation>
349X<cat> X<concat> X<concatenate> X<.>
a0d0e21e
LW
350
351=head2 Shift Operators
d74e8afc
ITB
352X<shift operator> X<operator, shift> X<<< << >>>
353X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
354X<shl> X<shr> X<shift, right> X<shift, left>
a0d0e21e 355
1ca345ed 356Binary C<<< << >>> returns the value of its left argument shifted left by the
55497cff 357number of bits specified by the right argument. Arguments should be
982ce180 358integers. (See also L<Integer Arithmetic>.)
a0d0e21e 359
1ca345ed 360Binary C<<< >> >>> returns the value of its left argument shifted right by
55497cff 361the number of bits specified by the right argument. Arguments should
982ce180 362be integers. (See also L<Integer Arithmetic>.)
a0d0e21e 363
1ca345ed
TC
364Note that both C<<< << >>> and C<<< >> >>> in Perl are implemented directly using
365C<<< << >>> and C<<< >> >>> in C. If C<use integer> (see L<Integer Arithmetic>) is
b16cf6df
JH
366in force then signed C integers are used, else unsigned C integers are
367used. Either way, the implementation isn't going to generate results
368larger than the size of the integer type Perl was built with (32 bits
369or 64 bits).
370
371The result of overflowing the range of the integers is undefined
372because it is undefined also in C. In other words, using 32-bit
373integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
374of bits is also undefined.
375
1ca345ed
TC
376If you get tired of being subject to your platform's native integers,
377the C<use bigint> pragma neatly sidesteps the issue altogether:
378
379 print 20 << 20; # 20971520
380 print 20 << 40; # 5120 on 32-bit machines,
381 # 21990232555520 on 64-bit machines
382 use bigint;
383 print 20 << 100; # 25353012004564588029934064107520
384
a0d0e21e 385=head2 Named Unary Operators
d74e8afc 386X<operator, named unary>
a0d0e21e
LW
387
388The various named unary operators are treated as functions with one
568e6d8b 389argument, with optional parentheses.
a0d0e21e
LW
390
391If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
392is followed by a left parenthesis as the next token, the operator and
393arguments within parentheses are taken to be of highest precedence,
3981b0eb 394just like a normal function call. For example,
1ca345ed 395because named unary operators are higher precedence than C<||>:
a0d0e21e
LW
396
397 chdir $foo || die; # (chdir $foo) || die
398 chdir($foo) || die; # (chdir $foo) || die
399 chdir ($foo) || die; # (chdir $foo) || die
400 chdir +($foo) || die; # (chdir $foo) || die
401
3981b0eb 402but, because * is higher precedence than named operators:
a0d0e21e
LW
403
404 chdir $foo * 20; # chdir ($foo * 20)
405 chdir($foo) * 20; # (chdir $foo) * 20
406 chdir ($foo) * 20; # (chdir $foo) * 20
407 chdir +($foo) * 20; # chdir ($foo * 20)
408
409 rand 10 * 20; # rand (10 * 20)
410 rand(10) * 20; # (rand 10) * 20
411 rand (10) * 20; # (rand 10) * 20
412 rand +(10) * 20; # rand (10 * 20)
413
568e6d8b
RGS
414Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
415treated like named unary operators, but they don't follow this functional
416parenthesis rule. That means, for example, that C<-f($file).".bak"> is
417equivalent to C<-f "$file.bak">.
d74e8afc 418X<-X> X<filetest> X<operator, filetest>
568e6d8b 419
5ba421f6 420See also L<"Terms and List Operators (Leftward)">.
a0d0e21e
LW
421
422=head2 Relational Operators
d74e8afc 423X<relational operator> X<operator, relational>
a0d0e21e 424
1ca345ed
TC
425Perl operators that return true or false generally return values
426that can be safely used as numbers. For example, the relational
427operators in this section and the equality operators in the next
428one return C<1> for true and a special version of the defined empty
429string, C<"">, which counts as a zero but is exempt from warnings
430about improper numeric conversions, just as C<"0 but true"> is.
431
35f2feb0 432Binary "<" returns true if the left argument is numerically less than
a0d0e21e 433the right argument.
d74e8afc 434X<< < >>
a0d0e21e 435
35f2feb0 436Binary ">" returns true if the left argument is numerically greater
a0d0e21e 437than the right argument.
d74e8afc 438X<< > >>
a0d0e21e 439
35f2feb0 440Binary "<=" returns true if the left argument is numerically less than
a0d0e21e 441or equal to the right argument.
d74e8afc 442X<< <= >>
a0d0e21e 443
35f2feb0 444Binary ">=" returns true if the left argument is numerically greater
a0d0e21e 445than or equal to the right argument.
d74e8afc 446X<< >= >>
a0d0e21e
LW
447
448Binary "lt" returns true if the left argument is stringwise less than
449the right argument.
d74e8afc 450X<< lt >>
a0d0e21e
LW
451
452Binary "gt" returns true if the left argument is stringwise greater
453than the right argument.
d74e8afc 454X<< gt >>
a0d0e21e
LW
455
456Binary "le" returns true if the left argument is stringwise less than
457or equal to the right argument.
d74e8afc 458X<< le >>
a0d0e21e
LW
459
460Binary "ge" returns true if the left argument is stringwise greater
461than or equal to the right argument.
d74e8afc 462X<< ge >>
a0d0e21e
LW
463
464=head2 Equality Operators
d74e8afc 465X<equality> X<equal> X<equals> X<operator, equality>
a0d0e21e
LW
466
467Binary "==" returns true if the left argument is numerically equal to
468the right argument.
d74e8afc 469X<==>
a0d0e21e
LW
470
471Binary "!=" returns true if the left argument is numerically not equal
472to the right argument.
d74e8afc 473X<!=>
a0d0e21e 474
35f2feb0 475Binary "<=>" returns -1, 0, or 1 depending on whether the left
6ee5d4e7 476argument is numerically less than, equal to, or greater than the right
d4ad863d 477argument. If your platform supports NaNs (not-a-numbers) as numeric
7d3a9d88
NC
478values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
479"<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
480returns true, as does NaN != anything else. If your platform doesn't
481support NaNs then NaN is just a string with numeric value 0.
d74e8afc 482X<< <=> >> X<spaceship>
7d3a9d88 483
1ca345ed
TC
484 $ perl -le '$a = "NaN"; print "No NaN support here" if $a == $a'
485 $ perl -le '$a = "NaN"; print "NaN support here" if $a != $a'
486
487(Note that the L<bigint>, L<bigrat>, and L<bignum> pragmas all
488support "NaN".)
a0d0e21e
LW
489
490Binary "eq" returns true if the left argument is stringwise equal to
491the right argument.
d74e8afc 492X<eq>
a0d0e21e
LW
493
494Binary "ne" returns true if the left argument is stringwise not equal
495to the right argument.
d74e8afc 496X<ne>
a0d0e21e 497
d4ad863d
JH
498Binary "cmp" returns -1, 0, or 1 depending on whether the left
499argument is stringwise less than, equal to, or greater than the right
500argument.
d74e8afc 501X<cmp>
a0d0e21e 502
1ca345ed
TC
503Binary "~~" does a smartmatch between its arguments. Smart matching
504is described in the next section.
0d863452
RH
505X<~~>
506
a034a98d 507"lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
66cbab2c
KW
508by the current locale if a legacy C<use locale> (but not
509C<use locale ':not_characters'>) is in effect. See
1ca345ed
TC
510L<perllocale>. Do not mix these with Unicode, only with legacy binary
511encodings. The standard L<Unicode::Collate> and
512L<Unicode::Collate::Locale> modules offer much more powerful solutions to
513collation issues.
514
515=head2 Smartmatch Operator
516
517First available in Perl 5.10.1 (the 5.10.0 version behaved differently),
518binary C<~~> does a "smartmatch" between its arguments. This is mostly
519used implicitly in the C<when> construct described in L<perlsyn>, although
520not all C<when> clauses call the smartmatch operator. Unique among all of
cc08d69f
RS
521Perl's operators, the smartmatch operator can recurse. The smartmatch
522operator is L<experimental|perlpolicy/experimental> and its behavior is
523subject to change.
1ca345ed
TC
524
525It is also unique in that all other Perl operators impose a context
526(usually string or numeric context) on their operands, autoconverting
527those operands to those imposed contexts. In contrast, smartmatch
528I<infers> contexts from the actual types of its operands and uses that
529type information to select a suitable comparison mechanism.
530
531The C<~~> operator compares its operands "polymorphically", determining how
532to compare them according to their actual types (numeric, string, array,
533hash, etc.) Like the equality operators with which it shares the same
534precedence, C<~~> returns 1 for true and C<""> for false. It is often best
535read aloud as "in", "inside of", or "is contained in", because the left
536operand is often looked for I<inside> the right operand. That makes the
40bec8a5 537order of the operands to the smartmatch operand often opposite that of
1ca345ed
TC
538the regular match operator. In other words, the "smaller" thing is usually
539placed in the left operand and the larger one in the right.
540
541The behavior of a smartmatch depends on what type of things its arguments
542are, as determined by the following table. The first row of the table
543whose types apply determines the smartmatch behavior. Because what
544actually happens is mostly determined by the type of the second operand,
545the table is sorted on the right operand instead of on the left.
546
547 Left Right Description and pseudocode
548 ===============================================================
549 Any undef check whether Any is undefined
550 like: !defined Any
551
552 Any Object invoke ~~ overloading on Object, or die
553
554 Right operand is an ARRAY:
555
556 Left Right Description and pseudocode
557 ===============================================================
558 ARRAY1 ARRAY2 recurse on paired elements of ARRAY1 and ARRAY2[2]
559 like: (ARRAY1[0] ~~ ARRAY2[0])
560 && (ARRAY1[1] ~~ ARRAY2[1]) && ...
561 HASH ARRAY any ARRAY elements exist as HASH keys
562 like: grep { exists HASH->{$_} } ARRAY
563 Regexp ARRAY any ARRAY elements pattern match Regexp
564 like: grep { /Regexp/ } ARRAY
565 undef ARRAY undef in ARRAY
566 like: grep { !defined } ARRAY
40bec8a5 567 Any ARRAY smartmatch each ARRAY element[3]
1ca345ed
TC
568 like: grep { Any ~~ $_ } ARRAY
569
570 Right operand is a HASH:
571
572 Left Right Description and pseudocode
573 ===============================================================
574 HASH1 HASH2 all same keys in both HASHes
575 like: keys HASH1 ==
576 grep { exists HASH2->{$_} } keys HASH1
577 ARRAY HASH any ARRAY elements exist as HASH keys
578 like: grep { exists HASH->{$_} } ARRAY
579 Regexp HASH any HASH keys pattern match Regexp
580 like: grep { /Regexp/ } keys HASH
581 undef HASH always false (undef can't be a key)
582 like: 0 == 1
583 Any HASH HASH key existence
584 like: exists HASH->{Any}
585
586 Right operand is CODE:
f703fc96 587
1ca345ed
TC
588 Left Right Description and pseudocode
589 ===============================================================
590 ARRAY CODE sub returns true on all ARRAY elements[1]
591 like: !grep { !CODE->($_) } ARRAY
592 HASH CODE sub returns true on all HASH keys[1]
593 like: !grep { !CODE->($_) } keys HASH
594 Any CODE sub passed Any returns true
595 like: CODE->(Any)
596
597Right operand is a Regexp:
598
599 Left Right Description and pseudocode
600 ===============================================================
601 ARRAY Regexp any ARRAY elements match Regexp
602 like: grep { /Regexp/ } ARRAY
603 HASH Regexp any HASH keys match Regexp
604 like: grep { /Regexp/ } keys HASH
605 Any Regexp pattern match
606 like: Any =~ /Regexp/
607
608 Other:
609
610 Left Right Description and pseudocode
611 ===============================================================
612 Object Any invoke ~~ overloading on Object,
613 or fall back to...
614
615 Any Num numeric equality
616 like: Any == Num
617 Num nummy[4] numeric equality
618 like: Num == nummy
619 undef Any check whether undefined
620 like: !defined(Any)
621 Any Any string equality
622 like: Any eq Any
623
624
625Notes:
626
627=over
628
629=item 1.
630Empty hashes or arrays match.
631
632=item 2.
40bec8a5 633That is, each element smartmatches the element of the same index in the other array.[3]
1ca345ed
TC
634
635=item 3.
636If a circular reference is found, fall back to referential equality.
637
638=item 4.
639Either an actual number, or a string that looks like one.
640
641=back
642
643The smartmatch implicitly dereferences any non-blessed hash or array
644reference, so the C<I<HASH>> and C<I<ARRAY>> entries apply in those cases.
645For blessed references, the C<I<Object>> entries apply. Smartmatches
646involving hashes only consider hash keys, never hash values.
647
648The "like" code entry is not always an exact rendition. For example, the
40bec8a5 649smartmatch operator short-circuits whenever possible, but C<grep> does
1ca345ed
TC
650not. Also, C<grep> in scalar context returns the number of matches, but
651C<~~> returns only true or false.
652
653Unlike most operators, the smartmatch operator knows to treat C<undef>
654specially:
655
656 use v5.10.1;
657 @array = (1, 2, 3, undef, 4, 5);
658 say "some elements undefined" if undef ~~ @array;
659
660Each operand is considered in a modified scalar context, the modification
661being that array and hash variables are passed by reference to the
662operator, which implicitly dereferences them. Both elements
663of each pair are the same:
664
665 use v5.10.1;
666
667 my %hash = (red => 1, blue => 2, green => 3,
668 orange => 4, yellow => 5, purple => 6,
669 black => 7, grey => 8, white => 9);
670
671 my @array = qw(red blue green);
672
673 say "some array elements in hash keys" if @array ~~ %hash;
674 say "some array elements in hash keys" if \@array ~~ \%hash;
675
676 say "red in array" if "red" ~~ @array;
677 say "red in array" if "red" ~~ \@array;
678
679 say "some keys end in e" if /e$/ ~~ %hash;
680 say "some keys end in e" if /e$/ ~~ \%hash;
681
40bec8a5
TC
682Two arrays smartmatch if each element in the first array smartmatches
683(that is, is "in") the corresponding element in the second array,
684recursively.
1ca345ed
TC
685
686 use v5.10.1;
687 my @little = qw(red blue green);
688 my @bigger = ("red", "blue", [ "orange", "green" ] );
689 if (@little ~~ @bigger) { # true!
690 say "little is contained in bigger";
691 }
692
693Because the smartmatch operator recurses on nested arrays, this
694will still report that "red" is in the array.
695
696 use v5.10.1;
697 my @array = qw(red blue green);
698 my $nested_array = [[[[[[[ @array ]]]]]]];
699 say "red in array" if "red" ~~ $nested_array;
700
701If two arrays smartmatch each other, then they are deep
702copies of each others' values, as this example reports:
703
704 use v5.12.0;
705 my @a = (0, 1, 2, [3, [4, 5], 6], 7);
706 my @b = (0, 1, 2, [3, [4, 5], 6], 7);
707
708 if (@a ~~ @b && @b ~~ @a) {
709 say "a and b are deep copies of each other";
710 }
711 elsif (@a ~~ @b) {
712 say "a smartmatches in b";
713 }
714 elsif (@b ~~ @a) {
715 say "b smartmatches in a";
716 }
717 else {
718 say "a and b don't smartmatch each other at all";
719 }
720
721
722If you were to set C<$b[3] = 4>, then instead of reporting that "a and b
723are deep copies of each other", it now reports that "b smartmatches in a".
724That because the corresponding position in C<@a> contains an array that
725(eventually) has a 4 in it.
726
727Smartmatching one hash against another reports whether both contain the
728same keys, no more and no less. This could be used to see whether two
729records have the same field names, without caring what values those fields
730might have. For example:
731
732 use v5.10.1;
733 sub make_dogtag {
734 state $REQUIRED_FIELDS = { name=>1, rank=>1, serial_num=>1 };
735
736 my ($class, $init_fields) = @_;
737
738 die "Must supply (only) name, rank, and serial number"
739 unless $init_fields ~~ $REQUIRED_FIELDS;
740
741 ...
742 }
743
744or, if other non-required fields are allowed, use ARRAY ~~ HASH:
745
746 use v5.10.1;
747 sub make_dogtag {
748 state $REQUIRED_FIELDS = { name=>1, rank=>1, serial_num=>1 };
749
750 my ($class, $init_fields) = @_;
751
752 die "Must supply (at least) name, rank, and serial number"
753 unless [keys %{$init_fields}] ~~ $REQUIRED_FIELDS;
754
755 ...
756 }
757
758The smartmatch operator is most often used as the implicit operator of a
759C<when> clause. See the section on "Switch Statements" in L<perlsyn>.
760
761=head3 Smartmatching of Objects
762
40bec8a5
TC
763To avoid relying on an object's underlying representation, if the
764smartmatch's right operand is an object that doesn't overload C<~~>,
765it raises the exception "C<Smartmatching a non-overloaded object
766breaks encapsulation>". That's because one has no business digging
767around to see whether something is "in" an object. These are all
768illegal on objects without a C<~~> overload:
1ca345ed
TC
769
770 %hash ~~ $object
771 42 ~~ $object
772 "fred" ~~ $object
773
774However, you can change the way an object is smartmatched by overloading
775the C<~~> operator. This is allowed to extend the usual smartmatch semantics.
776For objects that do have an C<~~> overload, see L<overload>.
777
778Using an object as the left operand is allowed, although not very useful.
779Smartmatching rules take precedence over overloading, so even if the
780object in the left operand has smartmatch overloading, this will be
781ignored. A left operand that is a non-overloaded object falls back on a
782string or numeric comparison of whatever the C<ref> operator returns. That
783means that
784
785 $object ~~ X
786
787does I<not> invoke the overload method with C<I<X>> as an argument.
788Instead the above table is consulted as normal, and based on the type of
789C<I<X>>, overloading may or may not be invoked. For simple strings or
790numbers, in becomes equivalent to this:
791
792 $object ~~ $number ref($object) == $number
793 $object ~~ $string ref($object) eq $string
794
795For example, this reports that the handle smells IOish
796(but please don't really do this!):
797
798 use IO::Handle;
799 my $fh = IO::Handle->new();
800 if ($fh ~~ /\bIO\b/) {
801 say "handle smells IOish";
802 }
803
804That's because it treats C<$fh> as a string like
805C<"IO::Handle=GLOB(0x8039e0)">, then pattern matches against that.
a034a98d 806
a0d0e21e 807=head2 Bitwise And
d74e8afc 808X<operator, bitwise, and> X<bitwise and> X<&>
a0d0e21e 809
c791a246
KW
810Binary "&" returns its operands ANDed together bit by bit. Although no
811warning is currently raised, the result is not well defined when this operation
812is performed on operands that aren't either numbers (see
813L<Integer Arithmetic>) or bitstrings (see L<Bitwise String Operators>).
a0d0e21e 814
2cdc098b 815Note that "&" has lower priority than relational operators, so for example
1ca345ed 816the parentheses are essential in a test like
2cdc098b 817
1ca345ed 818 print "Even\n" if ($x & 1) == 0;
2cdc098b 819
a0d0e21e 820=head2 Bitwise Or and Exclusive Or
d74e8afc
ITB
821X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
822X<bitwise xor> X<^>
a0d0e21e 823
2cdc098b 824Binary "|" returns its operands ORed together bit by bit.
a0d0e21e 825
2cdc098b 826Binary "^" returns its operands XORed together bit by bit.
c791a246
KW
827
828Although no warning is currently raised, the results are not well
829defined when these operations are performed on operands that aren't either
830numbers (see L<Integer Arithmetic>) or bitstrings (see L<Bitwise String
831Operators>).
a0d0e21e 832
2cdc098b
MG
833Note that "|" and "^" have lower priority than relational operators, so
834for example the brackets are essential in a test like
835
1ca345ed 836 print "false\n" if (8 | 2) != 10;
2cdc098b 837
a0d0e21e 838=head2 C-style Logical And
d74e8afc 839X<&&> X<logical and> X<operator, logical, and>
a0d0e21e
LW
840
841Binary "&&" performs a short-circuit logical AND operation. That is,
842if the left operand is false, the right operand is not even evaluated.
843Scalar or list context propagates down to the right operand if it
844is evaluated.
845
846=head2 C-style Logical Or
d74e8afc 847X<||> X<operator, logical, or>
a0d0e21e
LW
848
849Binary "||" performs a short-circuit logical OR operation. That is,
850if the left operand is true, the right operand is not even evaluated.
851Scalar or list context propagates down to the right operand if it
852is evaluated.
853
26d9d83b 854=head2 Logical Defined-Or
d74e8afc 855X<//> X<operator, logical, defined-or>
c963b151
BD
856
857Although it has no direct equivalent in C, Perl's C<//> operator is related
89d205f2 858to its C-style or. In fact, it's exactly the same as C<||>, except that it
95bee9ba
A
859tests the left hand side's definedness instead of its truth. Thus,
860C<< EXPR1 // EXPR2 >> returns the value of C<< EXPR1 >> if it's defined,
861otherwise, the value of C<< EXPR2 >> is returned. (C<< EXPR1 >> is evaluated
862in scalar context, C<< EXPR2 >> in the context of C<< // >> itself). Usually,
863this is the same result as C<< defined(EXPR1) ? EXPR1 : EXPR2 >> (except that
864the ternary-operator form can be used as a lvalue, while C<< EXPR1 // EXPR2 >>
865cannot). This is very useful for
bdc7923b
RGS
866providing default values for variables. If you actually want to test if
867at least one of C<$a> and C<$b> is defined, use C<defined($a // $b)>.
c963b151 868
d042e63d
MS
869The C<||>, C<//> and C<&&> operators return the last value evaluated
870(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
871portable way to find out the home directory might be:
a0d0e21e 872
c543c01b
TC
873 $home = $ENV{HOME}
874 // $ENV{LOGDIR}
875 // (getpwuid($<))[7]
876 // die "You're homeless!\n";
a0d0e21e 877
5a964f20
TC
878In particular, this means that you shouldn't use this
879for selecting between two aggregates for assignment:
880
881 @a = @b || @c; # this is wrong
882 @a = scalar(@b) || @c; # really meant this
883 @a = @b ? @b : @c; # this works fine, though
884
1ca345ed 885As alternatives to C<&&> and C<||> when used for
f23102e2
RGS
886control flow, Perl provides the C<and> and C<or> operators (see below).
887The short-circuit behavior is identical. The precedence of "and"
c963b151 888and "or" is much lower, however, so that you can safely use them after a
5a964f20 889list operator without the need for parentheses:
a0d0e21e
LW
890
891 unlink "alpha", "beta", "gamma"
892 or gripe(), next LINE;
893
894With the C-style operators that would have been written like this:
895
896 unlink("alpha", "beta", "gamma")
897 || (gripe(), next LINE);
898
1ca345ed
TC
899It would be even more readable to write that this way:
900
901 unless(unlink("alpha", "beta", "gamma")) {
902 gripe();
903 next LINE;
904 }
905
eeb6a2c9 906Using "or" for assignment is unlikely to do what you want; see below.
5a964f20
TC
907
908=head2 Range Operators
d74e8afc 909X<operator, range> X<range> X<..> X<...>
a0d0e21e
LW
910
911Binary ".." is the range operator, which is really two different
fb53bbb2 912operators depending on the context. In list context, it returns a
54ae734e 913list of values counting (up by ones) from the left value to the right
2cdbc966 914value. If the left value is greater than the right value then it
fb53bbb2 915returns the empty list. The range operator is useful for writing
54ae734e 916C<foreach (1..10)> loops and for doing slice operations on arrays. In
2cdbc966
JD
917the current implementation, no temporary array is created when the
918range operator is used as the expression in C<foreach> loops, but older
919versions of Perl might burn a lot of memory when you write something
920like this:
a0d0e21e
LW
921
922 for (1 .. 1_000_000) {
923 # code
54310121 924 }
a0d0e21e 925
8f0f46f8 926The range operator also works on strings, using the magical
927auto-increment, see below.
54ae734e 928
5a964f20 929In scalar context, ".." returns a boolean value. The operator is
8f0f46f8 930bistable, like a flip-flop, and emulates the line-range (comma)
931operator of B<sed>, B<awk>, and various editors. Each ".." operator
932maintains its own boolean state, even across calls to a subroutine
933that contains it. It is false as long as its left operand is false.
a0d0e21e
LW
934Once the left operand is true, the range operator stays true until the
935right operand is true, I<AFTER> which the range operator becomes false
8f0f46f8 936again. It doesn't become false till the next time the range operator
937is evaluated. It can test the right operand and become false on the
938same evaluation it became true (as in B<awk>), but it still returns
939true once. If you don't want it to test the right operand until the
940next evaluation, as in B<sed>, just use three dots ("...") instead of
19799a22
GS
941two. In all other regards, "..." behaves just like ".." does.
942
943The right operand is not evaluated while the operator is in the
944"false" state, and the left operand is not evaluated while the
945operator is in the "true" state. The precedence is a little lower
946than || and &&. The value returned is either the empty string for
8f0f46f8 947false, or a sequence number (beginning with 1) for true. The sequence
948number is reset for each range encountered. The final sequence number
949in a range has the string "E0" appended to it, which doesn't affect
950its numeric value, but gives you something to search for if you want
951to exclude the endpoint. You can exclude the beginning point by
952waiting for the sequence number to be greater than 1.
df5f8116
CW
953
954If either operand of scalar ".." is a constant expression,
955that operand is considered true if it is equal (C<==>) to the current
956input line number (the C<$.> variable).
957
958To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
959but that is only an issue if you use a floating point expression; when
960implicitly using C<$.> as described in the previous paragraph, the
961comparison is C<int(EXPR) == int($.)> which is only an issue when C<$.>
962is set to a floating point value and you are not reading from a file.
963Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
964you want in scalar context because each of the operands are evaluated
965using their integer representation.
966
967Examples:
a0d0e21e
LW
968
969As a scalar operator:
970
df5f8116 971 if (101 .. 200) { print; } # print 2nd hundred lines, short for
950b09ed 972 # if ($. == 101 .. $. == 200) { print; }
9f10b797
RGS
973
974 next LINE if (1 .. /^$/); # skip header lines, short for
f343f960 975 # next LINE if ($. == 1 .. /^$/);
9f10b797
RGS
976 # (typically in a loop labeled LINE)
977
978 s/^/> / if (/^$/ .. eof()); # quote body
a0d0e21e 979
5a964f20
TC
980 # parse mail messages
981 while (<>) {
982 $in_header = 1 .. /^$/;
df5f8116
CW
983 $in_body = /^$/ .. eof;
984 if ($in_header) {
f343f960 985 # do something
df5f8116 986 } else { # in body
f343f960 987 # do something else
df5f8116 988 }
5a964f20 989 } continue {
df5f8116 990 close ARGV if eof; # reset $. each file
5a964f20
TC
991 }
992
acf31ca5
SF
993Here's a simple example to illustrate the difference between
994the two range operators:
995
996 @lines = (" - Foo",
997 "01 - Bar",
998 "1 - Baz",
999 " - Quux");
1000
9f10b797
RGS
1001 foreach (@lines) {
1002 if (/0/ .. /1/) {
acf31ca5
SF
1003 print "$_\n";
1004 }
1005 }
1006
9f10b797
RGS
1007This program will print only the line containing "Bar". If
1008the range operator is changed to C<...>, it will also print the
acf31ca5
SF
1009"Baz" line.
1010
1011And now some examples as a list operator:
a0d0e21e 1012
1ca345ed
TC
1013 for (101 .. 200) { print } # print $_ 100 times
1014 @foo = @foo[0 .. $#foo]; # an expensive no-op
1015 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
a0d0e21e 1016
5a964f20 1017The range operator (in list context) makes use of the magical
5f05dabc 1018auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
1019can say
1020
c543c01b 1021 @alphabet = ("A" .. "Z");
a0d0e21e 1022
54ae734e 1023to get all normal letters of the English alphabet, or
a0d0e21e 1024
c543c01b 1025 $hexdigit = (0 .. 9, "a" .. "f")[$num & 15];
a0d0e21e
LW
1026
1027to get a hexadecimal digit, or
1028
1ca345ed
TC
1029 @z2 = ("01" .. "31");
1030 print $z2[$mday];
a0d0e21e 1031
ea4f5703
YST
1032to get dates with leading zeros.
1033
1034If the final value specified is not in the sequence that the magical
1035increment would produce, the sequence goes until the next value would
1036be longer than the final value specified.
1037
1038If the initial value specified isn't part of a magical increment
c543c01b 1039sequence (that is, a non-empty string matching C</^[a-zA-Z]*[0-9]*\z/>),
ea4f5703
YST
1040only the initial value will be returned. So the following will only
1041return an alpha:
1042
c543c01b 1043 use charnames "greek";
ea4f5703
YST
1044 my @greek_small = ("\N{alpha}" .. "\N{omega}");
1045
c543c01b
TC
1046To get the 25 traditional lowercase Greek letters, including both sigmas,
1047you could use this instead:
ea4f5703 1048
c543c01b 1049 use charnames "greek";
1ca345ed
TC
1050 my @greek_small = map { chr } ( ord("\N{alpha}")
1051 ..
1052 ord("\N{omega}")
1053 );
c543c01b
TC
1054
1055However, because there are I<many> other lowercase Greek characters than
1056just those, to match lowercase Greek characters in a regular expression,
47c56cc8
KW
1057you could use the pattern C</(?:(?=\p{Greek})\p{Lower})+/> (or the
1058L<experimental feature|perlrecharclass/Extended Bracketed Character
1059Classes> C<S</(?[ \p{Greek} & \p{Lower} ])+/>>).
a0d0e21e 1060
df5f8116
CW
1061Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
1062return two elements in list context.
1063
1064 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
1065
a0d0e21e 1066=head2 Conditional Operator
d74e8afc 1067X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
a0d0e21e
LW
1068
1069Ternary "?:" is the conditional operator, just as in C. It works much
1070like an if-then-else. If the argument before the ? is true, the
1071argument before the : is returned, otherwise the argument after the :
cb1a09d0
AD
1072is returned. For example:
1073
54310121 1074 printf "I have %d dog%s.\n", $n,
c543c01b 1075 ($n == 1) ? "" : "s";
cb1a09d0
AD
1076
1077Scalar or list context propagates downward into the 2nd
54310121 1078or 3rd argument, whichever is selected.
cb1a09d0
AD
1079
1080 $a = $ok ? $b : $c; # get a scalar
1081 @a = $ok ? @b : @c; # get an array
1082 $a = $ok ? @b : @c; # oops, that's just a count!
1083
1084The operator may be assigned to if both the 2nd and 3rd arguments are
1085legal lvalues (meaning that you can assign to them):
a0d0e21e
LW
1086
1087 ($a_or_b ? $a : $b) = $c;
1088
5a964f20
TC
1089Because this operator produces an assignable result, using assignments
1090without parentheses will get you in trouble. For example, this:
1091
1092 $a % 2 ? $a += 10 : $a += 2
1093
1094Really means this:
1095
1096 (($a % 2) ? ($a += 10) : $a) += 2
1097
1098Rather than this:
1099
1100 ($a % 2) ? ($a += 10) : ($a += 2)
1101
19799a22
GS
1102That should probably be written more simply as:
1103
1104 $a += ($a % 2) ? 10 : 2;
1105
4633a7c4 1106=head2 Assignment Operators
d74e8afc 1107X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
5ac3b81c 1108X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
d74e8afc 1109X<%=> X<^=> X<x=>
a0d0e21e
LW
1110
1111"=" is the ordinary assignment operator.
1112
1113Assignment operators work as in C. That is,
1114
1115 $a += 2;
1116
1117is equivalent to
1118
1119 $a = $a + 2;
1120
1121although without duplicating any side effects that dereferencing the lvalue
54310121
PP
1122might trigger, such as from tie(). Other assignment operators work similarly.
1123The following are recognized:
a0d0e21e
LW
1124
1125 **= += *= &= <<= &&=
9f10b797
RGS
1126 -= /= |= >>= ||=
1127 .= %= ^= //=
1128 x=
a0d0e21e 1129
19799a22 1130Although these are grouped by family, they all have the precedence
a0d0e21e
LW
1131of assignment.
1132
b350dd2f
GS
1133Unlike in C, the scalar assignment operator produces a valid lvalue.
1134Modifying an assignment is equivalent to doing the assignment and
1135then modifying the variable that was assigned to. This is useful
1136for modifying a copy of something, like this:
a0d0e21e 1137
1ca345ed
TC
1138 ($tmp = $global) =~ tr/13579/24680/;
1139
1140Although as of 5.14, that can be also be accomplished this way:
1141
1142 use v5.14;
1143 $tmp = ($global =~ tr/13579/24680/r);
a0d0e21e
LW
1144
1145Likewise,
1146
1147 ($a += 2) *= 3;
1148
1149is equivalent to
1150
1151 $a += 2;
1152 $a *= 3;
1153
b350dd2f
GS
1154Similarly, a list assignment in list context produces the list of
1155lvalues assigned to, and a list assignment in scalar context returns
1156the number of elements produced by the expression on the right hand
1157side of the assignment.
1158
748a9306 1159=head2 Comma Operator
d74e8afc 1160X<comma> X<operator, comma> X<,>
a0d0e21e 1161
5a964f20 1162Binary "," is the comma operator. In scalar context it evaluates
a0d0e21e
LW
1163its left argument, throws that value away, then evaluates its right
1164argument and returns that value. This is just like C's comma operator.
1165
5a964f20 1166In list context, it's just the list argument separator, and inserts
ed5c6d31
PJ
1167both its arguments into the list. These arguments are also evaluated
1168from left to right.
a0d0e21e 1169
4e1988c6
FC
1170The C<< => >> operator is a synonym for the comma except that it causes a
1171word on its left to be interpreted as a string if it begins with a letter
344f2c40
IG
1172or underscore and is composed only of letters, digits and underscores.
1173This includes operands that might otherwise be interpreted as operators,
1174constants, single number v-strings or function calls. If in doubt about
c543c01b 1175this behavior, the left operand can be quoted explicitly.
344f2c40
IG
1176
1177Otherwise, the C<< => >> operator behaves exactly as the comma operator
1178or list argument separator, according to context.
1179
1180For example:
a44e5664
MS
1181
1182 use constant FOO => "something";
1183
1184 my %h = ( FOO => 23 );
1185
1186is equivalent to:
1187
1188 my %h = ("FOO", 23);
1189
1190It is I<NOT>:
1191
1192 my %h = ("something", 23);
1193
719b43e8
RGS
1194The C<< => >> operator is helpful in documenting the correspondence
1195between keys and values in hashes, and other paired elements in lists.
748a9306 1196
a12b8f3c
FC
1197 %hash = ( $key => $value );
1198 login( $username => $password );
a44e5664 1199
4e1988c6
FC
1200The special quoting behavior ignores precedence, and hence may apply to
1201I<part> of the left operand:
1202
1203 print time.shift => "bbb";
1204
1205That example prints something like "1314363215shiftbbb", because the
1206C<< => >> implicitly quotes the C<shift> immediately on its left, ignoring
1207the fact that C<time.shift> is the entire left operand.
1208
a0d0e21e 1209=head2 List Operators (Rightward)
d74e8afc 1210X<operator, list, rightward> X<list operator>
a0d0e21e 1211
c543c01b 1212On the right side of a list operator, the comma has very low precedence,
a0d0e21e
LW
1213such that it controls all comma-separated expressions found there.
1214The only operators with lower precedence are the logical operators
1215"and", "or", and "not", which may be used to evaluate calls to list
1ca345ed
TC
1216operators without the need for parentheses:
1217
1218 open HANDLE, "< :utf8", "filename" or die "Can't open: $!\n";
1219
1220However, some people find that code harder to read than writing
1221it with parentheses:
1222
1223 open(HANDLE, "< :utf8", "filename") or die "Can't open: $!\n";
1224
1225in which case you might as well just use the more customary "||" operator:
a0d0e21e 1226
1ca345ed 1227 open(HANDLE, "< :utf8", "filename") || die "Can't open: $!\n";
a0d0e21e 1228
5ba421f6 1229See also discussion of list operators in L<Terms and List Operators (Leftward)>.
a0d0e21e
LW
1230
1231=head2 Logical Not
d74e8afc 1232X<operator, logical, not> X<not>
a0d0e21e
LW
1233
1234Unary "not" returns the logical negation of the expression to its right.
1235It's the equivalent of "!" except for the very low precedence.
1236
1237=head2 Logical And
d74e8afc 1238X<operator, logical, and> X<and>
a0d0e21e
LW
1239
1240Binary "and" returns the logical conjunction of the two surrounding
c543c01b
TC
1241expressions. It's equivalent to C<&&> except for the very low
1242precedence. This means that it short-circuits: the right
a0d0e21e
LW
1243expression is evaluated only if the left expression is true.
1244
59ab9d6e 1245=head2 Logical or and Exclusive Or
f23102e2 1246X<operator, logical, or> X<operator, logical, xor>
59ab9d6e 1247X<operator, logical, exclusive or>
f23102e2 1248X<or> X<xor>
a0d0e21e
LW
1249
1250Binary "or" returns the logical disjunction of the two surrounding
c543c01b
TC
1251expressions. It's equivalent to C<||> except for the very low precedence.
1252This makes it useful for control flow:
5a964f20
TC
1253
1254 print FH $data or die "Can't write to FH: $!";
1255
c543c01b
TC
1256This means that it short-circuits: the right expression is evaluated
1257only if the left expression is false. Due to its precedence, you must
1258be careful to avoid using it as replacement for the C<||> operator.
1259It usually works out better for flow control than in assignments:
5a964f20
TC
1260
1261 $a = $b or $c; # bug: this is wrong
1262 ($a = $b) or $c; # really means this
1263 $a = $b || $c; # better written this way
1264
19799a22 1265However, when it's a list-context assignment and you're trying to use
c543c01b 1266C<||> for control flow, you probably need "or" so that the assignment
5a964f20
TC
1267takes higher precedence.
1268
1269 @info = stat($file) || die; # oops, scalar sense of stat!
1270 @info = stat($file) or die; # better, now @info gets its due
1271
c963b151
BD
1272Then again, you could always use parentheses.
1273
1ca345ed 1274Binary C<xor> returns the exclusive-OR of the two surrounding expressions.
c543c01b 1275It cannot short-circuit (of course).
a0d0e21e 1276
59ab9d6e
MB
1277There is no low precedence operator for defined-OR.
1278
a0d0e21e 1279=head2 C Operators Missing From Perl
d74e8afc
ITB
1280X<operator, missing from perl> X<&> X<*>
1281X<typecasting> X<(TYPE)>
a0d0e21e
LW
1282
1283Here is what C has that Perl doesn't:
1284
1285=over 8
1286
1287=item unary &
1288
1289Address-of operator. (But see the "\" operator for taking a reference.)
1290
1291=item unary *
1292
54310121 1293Dereference-address operator. (Perl's prefix dereferencing
a0d0e21e
LW
1294operators are typed: $, @, %, and &.)
1295
1296=item (TYPE)
1297
19799a22 1298Type-casting operator.
a0d0e21e
LW
1299
1300=back
1301
5f05dabc 1302=head2 Quote and Quote-like Operators
89d205f2 1303X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
d74e8afc
ITB
1304X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
1305X<escape sequence> X<escape>
1306
a0d0e21e
LW
1307While we usually think of quotes as literal values, in Perl they
1308function as operators, providing various kinds of interpolating and
1309pattern matching capabilities. Perl provides customary quote characters
1310for these behaviors, but also provides a way for you to choose your
1311quote character for any of them. In the following table, a C<{}> represents
9f10b797 1312any pair of delimiters you choose.
a0d0e21e 1313
2c268ad5
TP
1314 Customary Generic Meaning Interpolates
1315 '' q{} Literal no
1316 "" qq{} Literal yes
af9219ee 1317 `` qx{} Command yes*
2c268ad5 1318 qw{} Word list no
af9219ee
MG
1319 // m{} Pattern match yes*
1320 qr{} Pattern yes*
1321 s{}{} Substitution yes*
2c268ad5 1322 tr{}{} Transliteration no (but see below)
c543c01b 1323 y{}{} Transliteration no (but see below)
7e3b091d 1324 <<EOF here-doc yes*
a0d0e21e 1325
af9219ee
MG
1326 * unless the delimiter is ''.
1327
87275199 1328Non-bracketing delimiters use the same character fore and aft, but the four
c543c01b 1329sorts of ASCII brackets (round, angle, square, curly) all nest, which means
9f10b797 1330that
87275199 1331
c543c01b 1332 q{foo{bar}baz}
35f2feb0 1333
9f10b797 1334is the same as
87275199 1335
c543c01b 1336 'foo{bar}baz'
87275199
GS
1337
1338Note, however, that this does not always work for quoting Perl code:
1339
c543c01b 1340 $s = q{ if($a eq "}") ... }; # WRONG
87275199 1341
c543c01b
TC
1342is a syntax error. The C<Text::Balanced> module (standard as of v5.8,
1343and from CPAN before then) is able to do this properly.
87275199 1344
19799a22 1345There can be whitespace between the operator and the quoting
fb73857a 1346characters, except when C<#> is being used as the quoting character.
19799a22
GS
1347C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
1348operator C<q> followed by a comment. Its argument will be taken
1349from the next line. This allows you to write:
fb73857a
PP
1350
1351 s {foo} # Replace foo
1352 {bar} # with bar.
1353
c543c01b
TC
1354The following escape sequences are available in constructs that interpolate,
1355and in transliterations:
5691ca5f 1356X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N> X<\N{}>
04341565 1357X<\o{}>
5691ca5f 1358
2c4c1ff2
KW
1359 Sequence Note Description
1360 \t tab (HT, TAB)
1361 \n newline (NL)
1362 \r return (CR)
1363 \f form feed (FF)
1364 \b backspace (BS)
1365 \a alarm (bell) (BEL)
1366 \e escape (ESC)
c543c01b 1367 \x{263A} [1,8] hex char (example: SMILEY)
2c4c1ff2 1368 \x1b [2,8] restricted range hex char (example: ESC)
fb121860 1369 \N{name} [3] named Unicode character or character sequence
2c4c1ff2
KW
1370 \N{U+263D} [4,8] Unicode character (example: FIRST QUARTER MOON)
1371 \c[ [5] control char (example: chr(27))
1372 \o{23072} [6,8] octal char (example: SMILEY)
1373 \033 [7,8] restricted range octal char (example: ESC)
5691ca5f
KW
1374
1375=over 4
1376
1377=item [1]
1378
2c4c1ff2
KW
1379The result is the character specified by the hexadecimal number between
1380the braces. See L</[8]> below for details on which character.
96448467
DG
1381
1382Only hexadecimal digits are valid between the braces. If an invalid
1383character is encountered, a warning will be issued and the invalid
1384character and all subsequent characters (valid or invalid) within the
1385braces will be discarded.
1386
1387If there are no valid digits between the braces, the generated character is
1388the NULL character (C<\x{00}>). However, an explicit empty brace (C<\x{}>)
c543c01b 1389will not cause a warning (currently).
40687185
KW
1390
1391=item [2]
1392
2c4c1ff2
KW
1393The result is the character specified by the hexadecimal number in the range
13940x00 to 0xFF. See L</[8]> below for details on which character.
96448467
DG
1395
1396Only hexadecimal digits are valid following C<\x>. When C<\x> is followed
2c4c1ff2 1397by fewer than two valid digits, any valid digits will be zero-padded. This
c543c01b 1398means that C<\x7> will be interpreted as C<\x07>, and a lone <\x> will be
2c4c1ff2 1399interpreted as C<\x00>. Except at the end of a string, having fewer than
c543c01b 1400two valid digits will result in a warning. Note that although the warning
96448467
DG
1401says the illegal character is ignored, it is only ignored as part of the
1402escape and will still be used as the subsequent character in the string.
1403For example:
1404
1405 Original Result Warns?
1406 "\x7" "\x07" no
1407 "\x" "\x00" no
1408 "\x7q" "\x07q" yes
1409 "\xq" "\x00q" yes
1410
40687185
KW
1411=item [3]
1412
fb121860 1413The result is the Unicode character or character sequence given by I<name>.
2c4c1ff2 1414See L<charnames>.
40687185
KW
1415
1416=item [4]
1417
2c4c1ff2
KW
1418C<\N{U+I<hexadecimal number>}> means the Unicode character whose Unicode code
1419point is I<hexadecimal number>.
40687185
KW
1420
1421=item [5]
1422
5691ca5f
KW
1423The character following C<\c> is mapped to some other character as shown in the
1424table:
1425
1426 Sequence Value
1427 \c@ chr(0)
1428 \cA chr(1)
1429 \ca chr(1)
1430 \cB chr(2)
1431 \cb chr(2)
1432 ...
1433 \cZ chr(26)
1434 \cz chr(26)
1435 \c[ chr(27)
1436 \c] chr(29)
1437 \c^ chr(30)
c3e9d7a9
KW
1438 \c_ chr(31)
1439 \c? chr(127) # (on ASCII platforms)
5691ca5f 1440
d813941f 1441In other words, it's the character whose code point has had 64 xor'd with
c3e9d7a9
KW
1442its uppercase. C<\c?> is DELETE on ASCII platforms because
1443S<C<ord("?") ^ 64>> is 127, and
d813941f
KW
1444C<\c@> is NULL because the ord of "@" is 64, so xor'ing 64 itself produces 0.
1445
5691ca5f
KW
1446Also, C<\c\I<X>> yields C< chr(28) . "I<X>"> for any I<X>, but cannot come at the
1447end of a string, because the backslash would be parsed as escaping the end
1448quote.
1449
1450On ASCII platforms, the resulting characters from the list above are the
1451complete set of ASCII controls. This isn't the case on EBCDIC platforms; see
c3e9d7a9
KW
1452L<perlebcdic/OPERATOR DIFFERENCES> for a full discussion of the
1453differences between these for ASCII versus EBCDIC platforms.
5691ca5f 1454
c3e9d7a9 1455Use of any other character following the C<"c"> besides those listed above is
63a63d81
KW
1456discouraged, and as of Perl v5.20, the only characters actually allowed
1457are the printable ASCII ones, minus the left brace C<"{">. What happens
1458for any of the allowed other characters is that the value is derived by
1459xor'ing with the seventh bit, which is 64, and a warning raised if
1460enabled. Using the non-allowed characters generates a fatal error.
5691ca5f
KW
1461
1462To get platform independent controls, you can use C<\N{...}>.
1463
40687185
KW
1464=item [6]
1465
2c4c1ff2
KW
1466The result is the character specified by the octal number between the braces.
1467See L</[8]> below for details on which character.
04341565
DG
1468
1469If a character that isn't an octal digit is encountered, a warning is raised,
1470and the value is based on the octal digits before it, discarding it and all
1471following characters up to the closing brace. It is a fatal error if there are
1472no octal digits at all.
1473
1474=item [7]
1475
c543c01b 1476The result is the character specified by the three-digit octal number in the
2c4c1ff2
KW
1477range 000 to 777 (but best to not use above 077, see next paragraph). See
1478L</[8]> below for details on which character.
1479
1480Some contexts allow 2 or even 1 digit, but any usage without exactly
40687185 1481three digits, the first being a zero, may give unintended results. (For
5db3e519
FC
1482example, in a regular expression it may be confused with a backreference;
1483see L<perlrebackslash/Octal escapes>.) Starting in Perl 5.14, you may
c543c01b 1484use C<\o{}> instead, which avoids all these problems. Otherwise, it is best to
04341565
DG
1485use this construct only for ordinals C<\077> and below, remembering to pad to
1486the left with zeros to make three digits. For larger ordinals, either use
9fef6a0d 1487C<\o{}>, or convert to something else, such as to hex and use C<\x{}>
04341565 1488instead.
40687185 1489
2c4c1ff2
KW
1490=item [8]
1491
c543c01b 1492Several constructs above specify a character by a number. That number
2c4c1ff2 1493gives the character's position in the character set encoding (indexed from 0).
c543c01b 1494This is called synonymously its ordinal, code position, or code point. Perl
2c4c1ff2
KW
1495works on platforms that have a native encoding currently of either ASCII/Latin1
1496or EBCDIC, each of which allow specification of 256 characters. In general, if
1497the number is 255 (0xFF, 0377) or below, Perl interprets this in the platform's
1498native encoding. If the number is 256 (0x100, 0400) or above, Perl interprets
c543c01b 1499it as a Unicode code point and the result is the corresponding Unicode
2c4c1ff2
KW
1500character. For example C<\x{50}> and C<\o{120}> both are the number 80 in
1501decimal, which is less than 256, so the number is interpreted in the native
1502character set encoding. In ASCII the character in the 80th position (indexed
1503from 0) is the letter "P", and in EBCDIC it is the ampersand symbol "&".
1504C<\x{100}> and C<\o{400}> are both 256 in decimal, so the number is interpreted
1505as a Unicode code point no matter what the native encoding is. The name of the
9fef6a0d 1506character in the 256th position (indexed by 0) in Unicode is
2c4c1ff2
KW
1507C<LATIN CAPITAL LETTER A WITH MACRON>.
1508
9fef6a0d 1509There are a couple of exceptions to the above rule. S<C<\N{U+I<hex number>}>> is
2c4c1ff2
KW
1510always interpreted as a Unicode code point, so that C<\N{U+0050}> is "P" even
1511on EBCDIC platforms. And if L<C<S<use encoding>>|encoding> is in effect, the
1512number is considered to be in that encoding, and is translated from that into
1513the platform's native encoding if there is a corresponding native character;
1514otherwise to Unicode.
1515
5691ca5f 1516=back
4c77eaa2 1517
e526e8bb 1518B<NOTE>: Unlike C and other languages, Perl has no C<\v> escape sequence for
8b312c40
KW
1519the vertical tab (VT, which is 11 in both ASCII and EBCDIC), but you may
1520use C<\ck> or
1521C<\x0b>. (C<\v>
e526e8bb
KW
1522does have meaning in regular expression patterns in Perl, see L<perlre>.)
1523
1524The following escape sequences are available in constructs that interpolate,
904501ec 1525but not in transliterations.
628253b8 1526X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q> X<\F>
904501ec 1527
c543c01b
TC
1528 \l lowercase next character only
1529 \u titlecase (not uppercase!) next character only
e4d34742
EB
1530 \L lowercase all characters till \E or end of string
1531 \U uppercase all characters till \E or end of string
628253b8 1532 \F foldcase all characters till \E or end of string
736fe711
KW
1533 \Q quote (disable) pattern metacharacters till \E or
1534 end of string
7e31b643 1535 \E end either case modification or quoted section
c543c01b
TC
1536 (whichever was last seen)
1537
736fe711
KW
1538See L<perlfunc/quotemeta> for the exact definition of characters that
1539are quoted by C<\Q>.
1540
628253b8 1541C<\L>, C<\U>, C<\F>, and C<\Q> can stack, in which case you need one
c543c01b
TC
1542C<\E> for each. For example:
1543
9fef6a0d
KW
1544 say"This \Qquoting \ubusiness \Uhere isn't quite\E done yet,\E is it?";
1545 This quoting\ Business\ HERE\ ISN\'T\ QUITE\ done\ yet\, is it?
a0d0e21e 1546
66cbab2c
KW
1547If C<use locale> is in effect (but not C<use locale ':not_characters'>),
1548the case map used by C<\l>, C<\L>,
c543c01b 1549C<\u>, and C<\U> is taken from the current locale. See L<perllocale>.
b6538e4f 1550If Unicode (for example, C<\N{}> or code points of 0x100 or
c543c01b
TC
1551beyond) is being used, the case map used by C<\l>, C<\L>, C<\u>, and
1552C<\U> is as defined by Unicode. That means that case-mapping
1553a single character can sometimes produce several characters.
31f05a37
KW
1554Under C<use locale>, C<\F> produces the same results as C<\L>
1555for all locales but a UTF-8 one, where it instead uses the Unicode
1556definition.
a034a98d 1557
5a964f20
TC
1558All systems use the virtual C<"\n"> to represent a line terminator,
1559called a "newline". There is no such thing as an unvarying, physical
19799a22 1560newline character. It is only an illusion that the operating system,
5a964f20
TC
1561device drivers, C libraries, and Perl all conspire to preserve. Not all
1562systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
c543c01b
TC
1563on the ancient Macs (pre-MacOS X) of yesteryear, these used to be reversed,
1564and on systems without line terminator,
1565printing C<"\n"> might emit no actual data. In general, use C<"\n"> when
5a964f20
TC
1566you mean a "newline" for your system, but use the literal ASCII when you
1567need an exact character. For example, most networking protocols expect
2a380090 1568and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
1569and although they often accept just C<"\012">, they seldom tolerate just
1570C<"\015">. If you get in the habit of using C<"\n"> for networking,
1571you may be burned some day.
d74e8afc
ITB
1572X<newline> X<line terminator> X<eol> X<end of line>
1573X<\n> X<\r> X<\r\n>
5a964f20 1574
904501ec
MG
1575For constructs that do interpolate, variables beginning with "C<$>"
1576or "C<@>" are interpolated. Subscripted variables such as C<$a[3]> or
ad0f383a
A
1577C<< $href->{key}[0] >> are also interpolated, as are array and hash slices.
1578But method calls such as C<< $obj->meth >> are not.
af9219ee
MG
1579
1580Interpolating an array or slice interpolates the elements in order,
1581separated by the value of C<$">, so is equivalent to interpolating
c543c01b
TC
1582C<join $", @array>. "Punctuation" arrays such as C<@*> are usually
1583interpolated only if the name is enclosed in braces C<@{*}>, but the
1584arrays C<@_>, C<@+>, and C<@-> are interpolated even without braces.
af9219ee 1585
bc7b91c6
EB
1586For double-quoted strings, the quoting from C<\Q> is applied after
1587interpolation and escapes are processed.
1588
1589 "abc\Qfoo\tbar$s\Exyz"
1590
1591is equivalent to
1592
1593 "abc" . quotemeta("foo\tbar$s") . "xyz"
1594
1595For the pattern of regex operators (C<qr//>, C<m//> and C<s///>),
1596the quoting from C<\Q> is applied after interpolation is processed,
1597but before escapes are processed. This allows the pattern to match
1598literally (except for C<$> and C<@>). For example, the following matches:
1599
1600 '\s\t' =~ /\Q\s\t/
1601
1602Because C<$> or C<@> trigger interpolation, you'll need to use something
1603like C</\Quser\E\@\Qhost/> to match them literally.
1d2dff63 1604
a0d0e21e
LW
1605Patterns are subject to an additional level of interpretation as a
1606regular expression. This is done as a second pass, after variables are
1607interpolated, so that regular expressions may be incorporated into the
1608pattern from the variables. If this is not what you want, use C<\Q> to
1609interpolate a variable literally.
1610
19799a22
GS
1611Apart from the behavior described above, Perl does not expand
1612multiple levels of interpolation. In particular, contrary to the
1613expectations of shell programmers, back-quotes do I<NOT> interpolate
1614within double quotes, nor do single quotes impede evaluation of
1615variables when used within double quotes.
a0d0e21e 1616
5f05dabc 1617=head2 Regexp Quote-Like Operators
d74e8afc 1618X<operator, regexp>
cb1a09d0 1619
5f05dabc 1620Here are the quote-like operators that apply to pattern
cb1a09d0
AD
1621matching and related activities.
1622
a0d0e21e
LW
1623=over 8
1624
b6fa137b 1625=item qr/STRING/msixpodual
01c6f5f4 1626X<qr> X</i> X</m> X</o> X</s> X</x> X</p>
a0d0e21e 1627
87e95b7f
YO
1628This operator quotes (and possibly compiles) its I<STRING> as a regular
1629expression. I<STRING> is interpolated the same way as I<PATTERN>
1630in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1631is done. Returns a Perl value which may be used instead of the
f6050459 1632corresponding C</STRING/msixpodual> expression. The returned value is a
85dd5c8b 1633normalized version of the original pattern. It magically differs from
1c8ee595
CO
1634a string containing the same characters: C<ref(qr/x/)> returns "Regexp";
1635however, dereferencing it is not well defined (you currently get the
1636normalized version of the original pattern, but this may change).
1637
a0d0e21e 1638
87e95b7f
YO
1639For example,
1640
1641 $rex = qr/my.STRING/is;
85dd5c8b 1642 print $rex; # prints (?si-xm:my.STRING)
87e95b7f
YO
1643 s/$rex/foo/;
1644
1645is equivalent to
1646
1647 s/my.STRING/foo/is;
1648
1649The result may be used as a subpattern in a match:
1650
1651 $re = qr/$pattern/;
7188ca43
KW
1652 $string =~ /foo${re}bar/; # can be interpolated in other
1653 # patterns
87e95b7f
YO
1654 $string =~ $re; # or used standalone
1655 $string =~ /$re/; # or this way
1656
f6050459 1657Since Perl may compile the pattern at the moment of execution of the qr()
87e95b7f
YO
1658operator, using qr() may have speed advantages in some situations,
1659notably if the result of qr() is used standalone:
1660
1661 sub match {
1662 my $patterns = shift;
1663 my @compiled = map qr/$_/i, @$patterns;
1664 grep {
1665 my $success = 0;
1666 foreach my $pat (@compiled) {
1667 $success = 1, last if /$pat/;
1668 }
1669 $success;
1670 } @_;
5a964f20
TC
1671 }
1672
87e95b7f
YO
1673Precompilation of the pattern into an internal representation at
1674the moment of qr() avoids a need to recompile the pattern every
1675time a match C</$pat/> is attempted. (Perl has many other internal
1676optimizations, but none would be triggered in the above example if
1677we did not use qr() operator.)
1678
765fa144 1679Options (specified by the following modifiers) are:
87e95b7f
YO
1680
1681 m Treat string as multiple lines.
1682 s Treat string as single line. (Make . match a newline)
1683 i Do case-insensitive pattern matching.
1684 x Use extended regular expressions.
1685 p When matching preserve a copy of the matched string so
7188ca43
KW
1686 that ${^PREMATCH}, ${^MATCH}, ${^POSTMATCH} will be
1687 defined.
87e95b7f 1688 o Compile pattern only once.
7188ca43
KW
1689 a ASCII-restrict: Use ASCII for \d, \s, \w; specifying two
1690 a's further restricts /i matching so that no ASCII
48cbae4f
SK
1691 character will match a non-ASCII one.
1692 l Use the locale.
1693 u Use Unicode rules.
1694 d Use Unicode or native charset, as in 5.12 and earlier.
87e95b7f
YO
1695
1696If a precompiled pattern is embedded in a larger pattern then the effect
c543c01b 1697of "msixpluad" will be propagated appropriately. The effect the "o"
87e95b7f
YO
1698modifier has is not propagated, being restricted to those patterns
1699explicitly using it.
1700
b6fa137b 1701The last four modifiers listed above, added in Perl 5.14,
850b7ec9 1702control the character set rules, but C</a> is the only one you are likely
18509dec
KW
1703to want to specify explicitly; the other three are selected
1704automatically by various pragmas.
da392a17 1705
87e95b7f 1706See L<perlre> for additional information on valid syntax for STRING, and
5e2aa8f5 1707for a detailed look at the semantics of regular expressions. In
1ca345ed
TC
1708particular, all modifiers except the largely obsolete C</o> are further
1709explained in L<perlre/Modifiers>. C</o> is described in the next section.
a0d0e21e 1710
b6fa137b 1711=item m/PATTERN/msixpodualgc
89d205f2
YO
1712X<m> X<operator, match>
1713X<regexp, options> X<regexp> X<regex, options> X<regex>
01c6f5f4 1714X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c>
a0d0e21e 1715
b6fa137b 1716=item /PATTERN/msixpodualgc
a0d0e21e 1717
5a964f20 1718Searches a string for a pattern match, and in scalar context returns
19799a22
GS
1719true if it succeeds, false if it fails. If no string is specified
1720via the C<=~> or C<!~> operator, the $_ string is searched. (The
1721string specified with C<=~> need not be an lvalue--it may be the
1722result of an expression evaluation, but remember the C<=~> binds
006671a6 1723rather tightly.) See also L<perlre>.
a0d0e21e 1724
f6050459 1725Options are as described in C<qr//> above; in addition, the following match
01c6f5f4 1726process modifiers are available:
a0d0e21e 1727
950b09ed 1728 g Match globally, i.e., find all occurrences.
7188ca43
KW
1729 c Do not reset search position on a failed match when /g is
1730 in effect.
a0d0e21e 1731
725a61d7 1732If "/" is the delimiter then the initial C<m> is optional. With the C<m>
c543c01b 1733you can use any pair of non-whitespace (ASCII) characters
725a61d7
Z
1734as delimiters. This is particularly useful for matching path names
1735that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
1736the delimiter, then a match-only-once rule applies,
6ca3c6c6
JK
1737described in C<m?PATTERN?> below. If "'" (single quote) is the delimiter,
1738no interpolation is performed on the PATTERN.
ed02a3bf
DN
1739When using a character valid in an identifier, whitespace is required
1740after the C<m>.
a0d0e21e 1741
532c9e80
KW
1742PATTERN may contain variables, which will be interpolated
1743every time the pattern search is evaluated, except
1f247705
GS
1744for when the delimiter is a single quote. (Note that C<$(>, C<$)>, and
1745C<$|> are not interpolated because they look like end-of-string tests.)
532c9e80
KW
1746Perl will not recompile the pattern unless an interpolated
1747variable that it contains changes. You can force Perl to skip the
1748test and never recompile by adding a C</o> (which stands for "once")
1749after the trailing delimiter.
1750Once upon a time, Perl would recompile regular expressions
1751unnecessarily, and this modifier was useful to tell it not to do so, in the
5cc41653 1752interests of speed. But now, the only reasons to use C</o> are one of:
532c9e80
KW
1753
1754=over
1755
1756=item 1
1757
1758The variables are thousands of characters long and you know that they
1759don't change, and you need to wring out the last little bit of speed by
1760having Perl skip testing for that. (There is a maintenance penalty for
1761doing this, as mentioning C</o> constitutes a promise that you won't
18509dec 1762change the variables in the pattern. If you do change them, Perl won't
532c9e80
KW
1763even notice.)
1764
1765=item 2
1766
1767you want the pattern to use the initial values of the variables
1768regardless of whether they change or not. (But there are saner ways
1769of accomplishing this than using C</o>.)
1770
fa9b8686
DM
1771=item 3
1772
1773If the pattern contains embedded code, such as
1774
1775 use re 'eval';
1776 $code = 'foo(?{ $x })';
1777 /$code/
1778
1779then perl will recompile each time, even though the pattern string hasn't
1780changed, to ensure that the current value of C<$x> is seen each time.
1781Use C</o> if you want to avoid this.
1782
532c9e80 1783=back
a0d0e21e 1784
18509dec
KW
1785The bottom line is that using C</o> is almost never a good idea.
1786
e9d89077
DN
1787=item The empty pattern //
1788
5a964f20 1789If the PATTERN evaluates to the empty string, the last
d65afb4b 1790I<successfully> matched regular expression is used instead. In this
c543c01b 1791case, only the C<g> and C<c> flags on the empty pattern are honored;
d65afb4b
HS
1792the other flags are taken from the original pattern. If no match has
1793previously succeeded, this will (silently) act instead as a genuine
1794empty pattern (which will always match).
a0d0e21e 1795
89d205f2
YO
1796Note that it's possible to confuse Perl into thinking C<//> (the empty
1797regex) is really C<//> (the defined-or operator). Perl is usually pretty
1798good about this, but some pathological cases might trigger this, such as
1799C<$a///> (is that C<($a) / (//)> or C<$a // />?) and C<print $fh //>
1800(C<print $fh(//> or C<print($fh //>?). In all of these examples, Perl
1801will assume you meant defined-or. If you meant the empty regex, just
1802use parentheses or spaces to disambiguate, or even prefix the empty
c963b151
BD
1803regex with an C<m> (so C<//> becomes C<m//>).
1804
e9d89077
DN
1805=item Matching in list context
1806
19799a22 1807If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1808list consisting of the subexpressions matched by the parentheses in the
3ff8ecf9
BF
1809pattern, that is, (C<$1>, C<$2>, C<$3>...) (Note that here C<$1> etc. are
1810also set). When there are no parentheses in the pattern, the return
1811value is the list C<(1)> for success.
1812With or without parentheses, an empty list is returned upon failure.
a0d0e21e
LW
1813
1814Examples:
1815
7188ca43
KW
1816 open(TTY, "+</dev/tty")
1817 || die "can't access /dev/tty: $!";
c543c01b 1818
7188ca43 1819 <TTY> =~ /^y/i && foo(); # do foo if desired
a0d0e21e 1820
7188ca43 1821 if (/Version: *([0-9.]*)/) { $version = $1; }
a0d0e21e 1822
7188ca43 1823 next if m#^/usr/spool/uucp#;
a0d0e21e 1824
7188ca43
KW
1825 # poor man's grep
1826 $arg = shift;
1827 while (<>) {
1828 print if /$arg/o; # compile only once (no longer needed!)
1829 }
a0d0e21e 1830
7188ca43 1831 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
a0d0e21e
LW
1832
1833This last example splits $foo into the first two words and the
5f05dabc 1834remainder of the line, and assigns those three fields to $F1, $F2, and
c543c01b
TC
1835$Etc. The conditional is true if any variables were assigned; that is,
1836if the pattern matched.
a0d0e21e 1837
19799a22 1838The C</g> modifier specifies global pattern matching--that is,
3dd93342 1839matching as many times as possible within the string. How it behaves
1840depends on the context. In list context, it returns a list of the
19799a22 1841substrings matched by any capturing parentheses in the regular
3dd93342 1842expression. If there are no parentheses, it returns a list of all
19799a22
GS
1843the matched strings, as if there were parentheses around the whole
1844pattern.
a0d0e21e 1845
7e86de3e 1846In scalar context, each execution of C<m//g> finds the next match,
19799a22 1847returning true if it matches, and false if there is no further match.
3dd93342 1848The position after the last match can be read or set using the C<pos()>
1849function; see L<perlfunc/pos>. A failed match normally resets the
7e86de3e 1850search position to the beginning of the string, but you can avoid that
1ca345ed 1851by adding the C</c> modifier (for example, C<m//gc>). Modifying the target
7e86de3e 1852string also resets the search position.
c90c0ff4 1853
e9d89077
DN
1854=item \G assertion
1855
c90c0ff4 1856You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
3dd93342 1857zero-width assertion that matches the exact position where the
1858previous C<m//g>, if any, left off. Without the C</g> modifier, the
1859C<\G> assertion still anchors at C<pos()> as it was at the start of
1860the operation (see L<perlfunc/pos>), but the match is of course only
1861attempted once. Using C<\G> without C</g> on a target string that has
1862not previously had a C</g> match applied to it is the same as using
1863the C<\A> assertion to match the beginning of the string. Note also
1864that, currently, C<\G> is only properly supported when anchored at the
1865very beginning of the pattern.
c90c0ff4
PP
1866
1867Examples:
a0d0e21e
LW
1868
1869 # list context
1870 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1871
1872 # scalar context
c543c01b
TC
1873 local $/ = "";
1874 while ($paragraph = <>) {
1875 while ($paragraph =~ /\p{Ll}['")]*[.!?]+['")]*\s/g) {
19799a22 1876 $sentences++;
a0d0e21e
LW
1877 }
1878 }
c543c01b
TC
1879 say $sentences;
1880
1881Here's another way to check for sentences in a paragraph:
1882
7188ca43
KW
1883 my $sentence_rx = qr{
1884 (?: (?<= ^ ) | (?<= \s ) ) # after start-of-string or
1885 # whitespace
1886 \p{Lu} # capital letter
1887 .*? # a bunch of anything
1888 (?<= \S ) # that ends in non-
1889 # whitespace
1890 (?<! \b [DMS]r ) # but isn't a common abbr.
1891 (?<! \b Mrs )
1892 (?<! \b Sra )
1893 (?<! \b St )
1894 [.?!] # followed by a sentence
1895 # ender
1896 (?= $ | \s ) # in front of end-of-string
1897 # or whitespace
1898 }sx;
1899 local $/ = "";
1900 while (my $paragraph = <>) {
1901 say "NEW PARAGRAPH";
1902 my $count = 0;
1903 while ($paragraph =~ /($sentence_rx)/g) {
1904 printf "\tgot sentence %d: <%s>\n", ++$count, $1;
c543c01b 1905 }
7188ca43 1906 }
c543c01b
TC
1907
1908Here's how to use C<m//gc> with C<\G>:
a0d0e21e 1909
137443ea 1910 $_ = "ppooqppqq";
44a8e56a
PP
1911 while ($i++ < 2) {
1912 print "1: '";
c90c0ff4 1913 print $1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1914 print "2: '";
c90c0ff4 1915 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1916 print "3: '";
c90c0ff4 1917 print $1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1918 }
5d43e42d 1919 print "Final: '$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a
PP
1920
1921The last example should print:
1922
1923 1: 'oo', pos=4
137443ea 1924 2: 'q', pos=5
44a8e56a
PP
1925 3: 'pp', pos=7
1926 1: '', pos=7
137443ea
PP
1927 2: 'q', pos=8
1928 3: '', pos=8
5d43e42d
DC
1929 Final: 'q', pos=8
1930
1931Notice that the final match matched C<q> instead of C<p>, which a match
1932without the C<\G> anchor would have done. Also note that the final match
ac036724 1933did not update C<pos>. C<pos> is only updated on a C</g> match. If the
c543c01b
TC
1934final match did indeed match C<p>, it's a good bet that you're running a
1935very old (pre-5.6.0) version of Perl.
44a8e56a 1936
c90c0ff4 1937A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1938combine several regexps like this to process a string part-by-part,
c90c0ff4
PP
1939doing different actions depending on which regexp matched. Each
1940regexp tries to match where the previous one leaves off.
e7ea3e70 1941
3fe9a6f1 1942 $_ = <<'EOL';
7188ca43
KW
1943 $url = URI::URL->new( "http://example.com/" );
1944 die if $url eq "xXx";
3fe9a6f1 1945 EOL
c543c01b
TC
1946
1947 LOOP: {
950b09ed 1948 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
7188ca43
KW
1949 print(" lowercase"), redo LOOP
1950 if /\G\p{Ll}+\b[,.;]?\s*/gc;
1951 print(" UPPERCASE"), redo LOOP
1952 if /\G\p{Lu}+\b[,.;]?\s*/gc;
1953 print(" Capitalized"), redo LOOP
1954 if /\G\p{Lu}\p{Ll}+\b[,.;]?\s*/gc;
c543c01b 1955 print(" MiXeD"), redo LOOP if /\G\pL+\b[,.;]?\s*/gc;
7188ca43
KW
1956 print(" alphanumeric"), redo LOOP
1957 if /\G[\p{Alpha}\pN]+\b[,.;]?\s*/gc;
c543c01b 1958 print(" line-noise"), redo LOOP if /\G\W+/gc;
950b09ed 1959 print ". That's all!\n";
c543c01b 1960 }
e7ea3e70
IZ
1961
1962Here is the output (split into several lines):
1963
7188ca43
KW
1964 line-noise lowercase line-noise UPPERCASE line-noise UPPERCASE
1965 line-noise lowercase line-noise lowercase line-noise lowercase
1966 lowercase line-noise lowercase lowercase line-noise lowercase
1967 lowercase line-noise MiXeD line-noise. That's all!
44a8e56a 1968
c543c01b 1969=item m?PATTERN?msixpodualgc
725a61d7 1970X<?> X<operator, match-once>
87e95b7f 1971
c543c01b 1972=item ?PATTERN?msixpodualgc
55d389e7 1973
725a61d7
Z
1974This is just like the C<m/PATTERN/> search, except that it matches
1975only once between calls to the reset() operator. This is a useful
87e95b7f 1976optimization when you want to see only the first occurrence of
ceb131e8 1977something in each file of a set of files, for instance. Only C<m??>
87e95b7f
YO
1978patterns local to the current package are reset.
1979
1980 while (<>) {
ceb131e8 1981 if (m?^$?) {
87e95b7f
YO
1982 # blank line between header and body
1983 }
1984 } continue {
725a61d7 1985 reset if eof; # clear m?? status for next file
87e95b7f
YO
1986 }
1987
c543c01b
TC
1988Another example switched the first "latin1" encoding it finds
1989to "utf8" in a pod file:
1990
1991 s//utf8/ if m? ^ =encoding \h+ \K latin1 ?x;
1992
1993The match-once behavior is controlled by the match delimiter being
725a61d7
Z
1994C<?>; with any other delimiter this is the normal C<m//> operator.
1995
1996For historical reasons, the leading C<m> in C<m?PATTERN?> is optional,
1997but the resulting C<?PATTERN?> syntax is deprecated, will warn on
c543c01b
TC
1998usage and might be removed from a future stable release of Perl (without
1999further notice!).
87e95b7f 2000
b6fa137b 2001=item s/PATTERN/REPLACEMENT/msixpodualgcer
87e95b7f 2002X<substitute> X<substitution> X<replace> X<regexp, replace>
4f4d7508 2003X<regexp, substitute> X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c> X</e> X</r>
87e95b7f
YO
2004
2005Searches a string for a pattern, and if found, replaces that pattern
2006with the replacement text and returns the number of substitutions
2007made. Otherwise it returns false (specifically, the empty string).
2008
c543c01b 2009If the C</r> (non-destructive) option is used then it runs the
679563bb
KW
2010substitution on a copy of the string and instead of returning the
2011number of substitutions, it returns the copy whether or not a
c543c01b
TC
2012substitution occurred. The original string is never changed when
2013C</r> is used. The copy will always be a plain string, even if the
2014input is an object or a tied variable.
4f4d7508 2015
87e95b7f 2016If no string is specified via the C<=~> or C<!~> operator, the C<$_>
c543c01b
TC
2017variable is searched and modified. Unless the C</r> option is used,
2018the string specified must be a scalar variable, an array element, a
2019hash element, or an assignment to one of those; that is, some sort of
2020scalar lvalue.
87e95b7f
YO
2021
2022If the delimiter chosen is a single quote, no interpolation is
2023done on either the PATTERN or the REPLACEMENT. Otherwise, if the
2024PATTERN contains a $ that looks like a variable rather than an
2025end-of-string test, the variable will be interpolated into the pattern
2026at run-time. If you want the pattern compiled only once the first time
2027the variable is interpolated, use the C</o> option. If the pattern
2028evaluates to the empty string, the last successfully executed regular
2029expression is used instead. See L<perlre> for further explanation on these.
87e95b7f
YO
2030
2031Options are as with m// with the addition of the following replacement
2032specific options:
2033
2034 e Evaluate the right side as an expression.
7188ca43
KW
2035 ee Evaluate the right side as a string then eval the
2036 result.
2037 r Return substitution and leave the original string
2038 untouched.
87e95b7f 2039
ed02a3bf
DN
2040Any non-whitespace delimiter may replace the slashes. Add space after
2041the C<s> when using a character allowed in identifiers. If single quotes
2042are used, no interpretation is done on the replacement string (the C</e>
3ff8ecf9 2043modifier overrides this, however). Note that Perl treats backticks
ed02a3bf
DN
2044as normal delimiters; the replacement text is not evaluated as a command.
2045If the PATTERN is delimited by bracketing quotes, the REPLACEMENT has
1ca345ed 2046its own pair of quotes, which may or may not be bracketing quotes, for example,
87e95b7f
YO
2047C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
2048replacement portion to be treated as a full-fledged Perl expression
2049and evaluated right then and there. It is, however, syntax checked at
2050compile-time. A second C<e> modifier will cause the replacement portion
2051to be C<eval>ed before being run as a Perl expression.
2052
2053Examples:
2054
7188ca43 2055 s/\bgreen\b/mauve/g; # don't change wintergreen
87e95b7f
YO
2056
2057 $path =~ s|/usr/bin|/usr/local/bin|;
2058
2059 s/Login: $foo/Login: $bar/; # run-time pattern
2060
7188ca43
KW
2061 ($foo = $bar) =~ s/this/that/; # copy first, then
2062 # change
2063 ($foo = "$bar") =~ s/this/that/; # convert to string,
2064 # copy, then change
4f4d7508
DC
2065 $foo = $bar =~ s/this/that/r; # Same as above using /r
2066 $foo = $bar =~ s/this/that/r
7188ca43
KW
2067 =~ s/that/the other/r; # Chained substitutes
2068 # using /r
2069 @foo = map { s/this/that/r } @bar # /r is very useful in
2070 # maps
87e95b7f 2071
7188ca43 2072 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-cnt
87e95b7f
YO
2073
2074 $_ = 'abc123xyz';
2075 s/\d+/$&*2/e; # yields 'abc246xyz'
2076 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
2077 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
2078
2079 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
2080 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
2081 s/^=(\w+)/pod($1)/ge; # use function call
2082
4f4d7508
DC
2083 $_ = 'abc123xyz';
2084 $a = s/abc/def/r; # $a is 'def123xyz' and
2085 # $_ remains 'abc123xyz'.
2086
87e95b7f
YO
2087 # expand variables in $_, but dynamics only, using
2088 # symbolic dereferencing
2089 s/\$(\w+)/${$1}/g;
2090
2091 # Add one to the value of any numbers in the string
2092 s/(\d+)/1 + $1/eg;
2093
c543c01b
TC
2094 # Titlecase words in the last 30 characters only
2095 substr($str, -30) =~ s/\b(\p{Alpha}+)\b/\u\L$1/g;
2096
87e95b7f
YO
2097 # This will expand any embedded scalar variable
2098 # (including lexicals) in $_ : First $1 is interpolated
2099 # to the variable name, and then evaluated
2100 s/(\$\w+)/$1/eeg;
2101
2102 # Delete (most) C comments.
2103 $program =~ s {
2104 /\* # Match the opening delimiter.
2105 .*? # Match a minimal number of characters.
2106 \*/ # Match the closing delimiter.
2107 } []gsx;
2108
7188ca43
KW
2109 s/^\s*(.*?)\s*$/$1/; # trim whitespace in $_,
2110 # expensively
87e95b7f 2111
7188ca43
KW
2112 for ($variable) { # trim whitespace in $variable,
2113 # cheap
87e95b7f
YO
2114 s/^\s+//;
2115 s/\s+$//;
2116 }
2117
2118 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
2119
2120Note the use of $ instead of \ in the last example. Unlike
2121B<sed>, we use the \<I<digit>> form in only the left hand side.
2122Anywhere else it's $<I<digit>>.
2123
2124Occasionally, you can't use just a C</g> to get all the changes
2125to occur that you might want. Here are two common cases:
2126
2127 # put commas in the right places in an integer
2128 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
2129
2130 # expand tabs to 8-column spacing
2131 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
2132
2133=back
2134
2135=head2 Quote-Like Operators
2136X<operator, quote-like>
2137
01c6f5f4
RGS
2138=over 4
2139
a0d0e21e 2140=item q/STRING/
5d44bfff 2141X<q> X<quote, single> X<'> X<''>
a0d0e21e 2142
5d44bfff 2143=item 'STRING'
a0d0e21e 2144
19799a22 2145A single-quoted, literal string. A backslash represents a backslash
68dc0745
PP
2146unless followed by the delimiter or another backslash, in which case
2147the delimiter or backslash is interpolated.
a0d0e21e
LW
2148
2149 $foo = q!I said, "You said, 'She said it.'"!;
2150 $bar = q('This is it.');
68dc0745 2151 $baz = '\n'; # a two-character string
a0d0e21e
LW
2152
2153=item qq/STRING/
d74e8afc 2154X<qq> X<quote, double> X<"> X<"">
a0d0e21e
LW
2155
2156=item "STRING"
2157
2158A double-quoted, interpolated string.
2159
2160 $_ .= qq
2161 (*** The previous line contains the naughty word "$1".\n)
19799a22 2162 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 2163 $baz = "\n"; # a one-character string
a0d0e21e
LW
2164
2165=item qx/STRING/
d74e8afc 2166X<qx> X<`> X<``> X<backtick>
a0d0e21e
LW
2167
2168=item `STRING`
2169
43dd4d21 2170A string which is (possibly) interpolated and then executed as a
f703fc96 2171system command with F</bin/sh> or its equivalent. Shell wildcards,
43dd4d21
JH
2172pipes, and redirections will be honored. The collected standard
2173output of the command is returned; standard error is unaffected. In
2174scalar context, it comes back as a single (potentially multi-line)
2175string, or undef if the command failed. In list context, returns a
2176list of lines (however you've defined lines with $/ or
2177$INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
5a964f20
TC
2178
2179Because backticks do not affect standard error, use shell file descriptor
2180syntax (assuming the shell supports this) if you care to address this.
2181To capture a command's STDERR and STDOUT together:
a0d0e21e 2182
5a964f20
TC
2183 $output = `cmd 2>&1`;
2184
2185To capture a command's STDOUT but discard its STDERR:
2186
2187 $output = `cmd 2>/dev/null`;
2188
2189To capture a command's STDERR but discard its STDOUT (ordering is
2190important here):
2191
2192 $output = `cmd 2>&1 1>/dev/null`;
2193
2194To exchange a command's STDOUT and STDERR in order to capture the STDERR
2195but leave its STDOUT to come out the old STDERR:
2196
2197 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
2198
2199To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
2200to redirect them separately to files, and then read from those files
2201when the program is done:
5a964f20 2202
2359510d 2203 system("program args 1>program.stdout 2>program.stderr");
5a964f20 2204
30398227
SP
2205The STDIN filehandle used by the command is inherited from Perl's STDIN.
2206For example:
2207
c543c01b
TC
2208 open(SPLAT, "stuff") || die "can't open stuff: $!";
2209 open(STDIN, "<&SPLAT") || die "can't dupe SPLAT: $!";
40bbb707 2210 print STDOUT `sort`;
30398227 2211
40bbb707 2212will print the sorted contents of the file named F<"stuff">.
30398227 2213
5a964f20
TC
2214Using single-quote as a delimiter protects the command from Perl's
2215double-quote interpolation, passing it on to the shell instead:
2216
2217 $perl_info = qx(ps $$); # that's Perl's $$
2218 $shell_info = qx'ps $$'; # that's the new shell's $$
2219
19799a22 2220How that string gets evaluated is entirely subject to the command
5a964f20
TC
2221interpreter on your system. On most platforms, you will have to protect
2222shell metacharacters if you want them treated literally. This is in
2223practice difficult to do, as it's unclear how to escape which characters.
2224See L<perlsec> for a clean and safe example of a manual fork() and exec()
2225to emulate backticks safely.
a0d0e21e 2226
bb32b41a
GS
2227On some platforms (notably DOS-like ones), the shell may not be
2228capable of dealing with multiline commands, so putting newlines in
2229the string may not get you what you want. You may be able to evaluate
2230multiple commands in a single line by separating them with the command
1ca345ed
TC
2231separator character, if your shell supports that (for example, C<;> on
2232many Unix shells and C<&> on the Windows NT C<cmd> shell).
bb32b41a 2233
3ff8ecf9 2234Perl will attempt to flush all files opened for
0f897271
GS
2235output before starting the child process, but this may not be supported
2236on some platforms (see L<perlport>). To be safe, you may need to set
2237C<$|> ($AUTOFLUSH in English) or call the C<autoflush()> method of
2238C<IO::Handle> on any open handles.
2239
bb32b41a
GS
2240Beware that some command shells may place restrictions on the length
2241of the command line. You must ensure your strings don't exceed this
2242limit after any necessary interpolations. See the platform-specific
2243release notes for more details about your particular environment.
2244
5a964f20
TC
2245Using this operator can lead to programs that are difficult to port,
2246because the shell commands called vary between systems, and may in
2247fact not be present at all. As one example, the C<type> command under
2248the POSIX shell is very different from the C<type> command under DOS.
2249That doesn't mean you should go out of your way to avoid backticks
2250when they're the right way to get something done. Perl was made to be
2251a glue language, and one of the things it glues together is commands.
2252Just understand what you're getting yourself into.
bb32b41a 2253
da87341d 2254See L</"I/O Operators"> for more discussion.
a0d0e21e 2255
945c54fd 2256=item qw/STRING/
d74e8afc 2257X<qw> X<quote, list> X<quote, words>
945c54fd
JH
2258
2259Evaluates to a list of the words extracted out of STRING, using embedded
2260whitespace as the word delimiters. It can be understood as being roughly
2261equivalent to:
2262
c543c01b 2263 split(" ", q/STRING/);
945c54fd 2264
efb1e162
CW
2265the differences being that it generates a real list at compile time, and
2266in scalar context it returns the last element in the list. So
945c54fd
JH
2267this expression:
2268
2269 qw(foo bar baz)
2270
2271is semantically equivalent to the list:
2272
c543c01b 2273 "foo", "bar", "baz"
945c54fd
JH
2274
2275Some frequently seen examples:
2276
2277 use POSIX qw( setlocale localeconv )
2278 @EXPORT = qw( foo bar baz );
2279
2280A common mistake is to try to separate the words with comma or to
2281put comments into a multi-line C<qw>-string. For this reason, the
89d205f2 2282C<use warnings> pragma and the B<-w> switch (that is, the C<$^W> variable)
945c54fd
JH
2283produces warnings if the STRING contains the "," or the "#" character.
2284
8ff32507 2285=item tr/SEARCHLIST/REPLACEMENTLIST/cdsr
d74e8afc 2286X<tr> X<y> X<transliterate> X</c> X</d> X</s>
a0d0e21e 2287
8ff32507 2288=item y/SEARCHLIST/REPLACEMENTLIST/cdsr
a0d0e21e 2289
2c268ad5 2290Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
2291with the corresponding character in the replacement list. It returns
2292the number of characters replaced or deleted. If no string is
c543c01b
TC
2293specified via the C<=~> or C<!~> operator, the $_ string is transliterated.
2294
2295If the C</r> (non-destructive) option is present, a new copy of the string
2296is made and its characters transliterated, and this copy is returned no
2297matter whether it was modified or not: the original string is always
2298left unchanged. The new copy is always a plain string, even if the input
2299string is an object or a tied variable.
8ada0baa 2300
c543c01b
TC
2301Unless the C</r> option is used, the string specified with C<=~> must be a
2302scalar variable, an array element, a hash element, or an assignment to one
2303of those; in other words, an lvalue.
8ff32507 2304
89d205f2 2305A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
2c268ad5 2306does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121
PP
2307For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
2308SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
c543c01b
TC
2309its own pair of quotes, which may or may not be bracketing quotes;
2310for example, C<tr[aeiouy][yuoiea]> or C<tr(+\-*/)/ABCD/>.
2311
2312Note that C<tr> does B<not> do regular expression character classes such as
2313C<\d> or C<\pL>. The C<tr> operator is not equivalent to the tr(1)
2314utility. If you want to map strings between lower/upper cases, see
2315L<perlfunc/lc> and L<perlfunc/uc>, and in general consider using the C<s>
2316operator if you need regular expressions. The C<\U>, C<\u>, C<\L>, and
2317C<\l> string-interpolation escapes on the right side of a substitution
2318operator will perform correct case-mappings, but C<tr[a-z][A-Z]> will not
2319(except sometimes on legacy 7-bit data).
cc255d5f 2320
8ada0baa
JH
2321Note also that the whole range idea is rather unportable between
2322character sets--and even within character sets they may cause results
2323you probably didn't expect. A sound principle is to use only ranges
2324that begin from and end at either alphabets of equal case (a-e, A-E),
2325or digits (0-4). Anything else is unsafe. If in doubt, spell out the
2326character sets in full.
2327
a0d0e21e
LW
2328Options:
2329
2330 c Complement the SEARCHLIST.
2331 d Delete found but unreplaced characters.
2332 s Squash duplicate replaced characters.
8ff32507
FC
2333 r Return the modified string and leave the original string
2334 untouched.
a0d0e21e 2335
19799a22
GS
2336If the C</c> modifier is specified, the SEARCHLIST character set
2337is complemented. If the C</d> modifier is specified, any characters
2338specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
2339(Note that this is slightly more flexible than the behavior of some
2340B<tr> programs, which delete anything they find in the SEARCHLIST,
2341period.) If the C</s> modifier is specified, sequences of characters
2342that were transliterated to the same character are squashed down
2343to a single instance of the character.
a0d0e21e
LW
2344
2345If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
2346exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
2347than the SEARCHLIST, the final character is replicated till it is long
5a964f20 2348enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
a0d0e21e
LW
2349This latter is useful for counting characters in a class or for
2350squashing character sequences in a class.
2351
2352Examples:
2353
c543c01b 2354 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case ASCII
a0d0e21e
LW
2355
2356 $cnt = tr/*/*/; # count the stars in $_
2357
2358 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
2359
2360 $cnt = tr/0-9//; # count the digits in $_
2361
2362 tr/a-zA-Z//s; # bookkeeper -> bokeper
2363
2364 ($HOST = $host) =~ tr/a-z/A-Z/;
c543c01b 2365 $HOST = $host =~ tr/a-z/A-Z/r; # same thing
8ff32507 2366
c543c01b 2367 $HOST = $host =~ tr/a-z/A-Z/r # chained with s///r
8ff32507 2368 =~ s/:/ -p/r;
a0d0e21e
LW
2369
2370 tr/a-zA-Z/ /cs; # change non-alphas to single space
2371
8ff32507
FC
2372 @stripped = map tr/a-zA-Z/ /csr, @original;
2373 # /r with map
2374
a0d0e21e 2375 tr [\200-\377]
c543c01b 2376 [\000-\177]; # wickedly delete 8th bit
a0d0e21e 2377
19799a22
GS
2378If multiple transliterations are given for a character, only the
2379first one is used:
748a9306
LW
2380
2381 tr/AAA/XYZ/
2382
2c268ad5 2383will transliterate any A to X.
748a9306 2384
19799a22 2385Because the transliteration table is built at compile time, neither
a0d0e21e 2386the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
19799a22
GS
2387interpolation. That means that if you want to use variables, you
2388must use an eval():
a0d0e21e
LW
2389
2390 eval "tr/$oldlist/$newlist/";
2391 die $@ if $@;
2392
2393 eval "tr/$oldlist/$newlist/, 1" or die $@;
2394
7e3b091d 2395=item <<EOF
d74e8afc 2396X<here-doc> X<heredoc> X<here-document> X<<< << >>>
7e3b091d
DA
2397
2398A line-oriented form of quoting is based on the shell "here-document"
2399syntax. Following a C<< << >> you specify a string to terminate
2400the quoted material, and all lines following the current line down to
89d205f2
YO
2401the terminating string are the value of the item.
2402
2403The terminating string may be either an identifier (a word), or some
2404quoted text. An unquoted identifier works like double quotes.
2405There may not be a space between the C<< << >> and the identifier,
2406unless the identifier is explicitly quoted. (If you put a space it
2407will be treated as a null identifier, which is valid, and matches the
2408first empty line.) The terminating string must appear by itself
2409(unquoted and with no surrounding whitespace) on the terminating line.
2410
2411If the terminating string is quoted, the type of quotes used determine
2412the treatment of the text.
2413
2414=over 4
2415
2416=item Double Quotes
2417
2418Double quotes indicate that the text will be interpolated using exactly
2419the same rules as normal double quoted strings.
7e3b091d
DA
2420
2421 print <<EOF;
2422 The price is $Price.
2423 EOF
2424
2425 print << "EOF"; # same as above
2426 The price is $Price.
2427 EOF
2428
89d205f2
YO
2429
2430=item Single Quotes
2431
2432Single quotes indicate the text is to be treated literally with no
2433interpolation of its content. This is similar to single quoted
2434strings except that backslashes have no special meaning, with C<\\>
2435being treated as two backslashes and not one as they would in every
2436other quoting construct.
2437
c543c01b
TC
2438Just as in the shell, a backslashed bareword following the C<<< << >>>
2439means the same thing as a single-quoted string does:
2440
2441 $cost = <<'VISTA'; # hasta la ...
2442 That'll be $10 please, ma'am.
2443 VISTA
2444
2445 $cost = <<\VISTA; # Same thing!
2446 That'll be $10 please, ma'am.
2447 VISTA
2448
89d205f2
YO
2449This is the only form of quoting in perl where there is no need
2450to worry about escaping content, something that code generators
2451can and do make good use of.
2452
2453=item Backticks
2454
2455The content of the here doc is treated just as it would be if the
2456string were embedded in backticks. Thus the content is interpolated
2457as though it were double quoted and then executed via the shell, with
2458the results of the execution returned.
2459
2460 print << `EOC`; # execute command and get results
7e3b091d 2461 echo hi there
7e3b091d
DA
2462 EOC
2463
89d205f2
YO
2464=back
2465
2466It is possible to stack multiple here-docs in a row:
2467
7e3b091d
DA
2468 print <<"foo", <<"bar"; # you can stack them
2469 I said foo.
2470 foo
2471 I said bar.
2472 bar
2473
2474 myfunc(<< "THIS", 23, <<'THAT');
2475 Here's a line
2476 or two.
2477 THIS
2478 and here's another.
2479 THAT
2480
2481Just don't forget that you have to put a semicolon on the end
2482to finish the statement, as Perl doesn't know you're not going to
2483try to do this:
2484
2485 print <<ABC
2486 179231
2487 ABC
2488 + 20;
2489
872d7e53
ST
2490If you want to remove the line terminator from your here-docs,
2491use C<chomp()>.
2492
2493 chomp($string = <<'END');
2494 This is a string.
2495 END
2496
2497If you want your here-docs to be indented with the rest of the code,
2498you'll need to remove leading whitespace from each line manually:
7e3b091d
DA
2499
2500 ($quote = <<'FINIS') =~ s/^\s+//gm;
89d205f2 2501 The Road goes ever on and on,
7e3b091d
DA
2502 down from the door where it began.
2503 FINIS
2504
2505If you use a here-doc within a delimited construct, such as in C<s///eg>,
1bf48760
FC
2506the quoted material must still come on the line following the
2507C<<< <<FOO >>> marker, which means it may be inside the delimited
2508construct:
7e3b091d
DA
2509
2510 s/this/<<E . 'that'
2511 the other
2512 E
2513 . 'more '/eg;
2514
1bf48760
FC
2515It works this way as of Perl 5.18. Historically, it was inconsistent, and
2516you would have to write
7e3b091d 2517
89d205f2
YO
2518 s/this/<<E . 'that'
2519 . 'more '/eg;
2520 the other
2521 E
7e3b091d 2522
1bf48760
FC
2523outside of string evals.
2524
c543c01b
TC
2525Additionally, quoting rules for the end-of-string identifier are
2526unrelated to Perl's quoting rules. C<q()>, C<qq()>, and the like are not
89d205f2
YO
2527supported in place of C<''> and C<"">, and the only interpolation is for
2528backslashing the quoting character:
7e3b091d
DA
2529
2530 print << "abc\"def";
2531 testing...
2532 abc"def
2533
2534Finally, quoted strings cannot span multiple lines. The general rule is
2535that the identifier must be a string literal. Stick with that, and you
2536should be safe.
2537
a0d0e21e
LW
2538=back
2539
75e14d17 2540=head2 Gory details of parsing quoted constructs
d74e8afc 2541X<quote, gory details>
75e14d17 2542
19799a22
GS
2543When presented with something that might have several different
2544interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
2545principle to pick the most probable interpretation. This strategy
2546is so successful that Perl programmers often do not suspect the
2547ambivalence of what they write. But from time to time, Perl's
2548notions differ substantially from what the author honestly meant.
2549
2550This section hopes to clarify how Perl handles quoted constructs.
2551Although the most common reason to learn this is to unravel labyrinthine
2552regular expressions, because the initial steps of parsing are the
2553same for all quoting operators, they are all discussed together.
2554
2555The most important Perl parsing rule is the first one discussed
2556below: when processing a quoted construct, Perl first finds the end
2557of that construct, then interprets its contents. If you understand
2558this rule, you may skip the rest of this section on the first
2559reading. The other rules are likely to contradict the user's
2560expectations much less frequently than this first one.
2561
2562Some passes discussed below are performed concurrently, but because
2563their results are the same, we consider them individually. For different
2564quoting constructs, Perl performs different numbers of passes, from
6deea57f 2565one to four, but these passes are always performed in the same order.
75e14d17 2566
13a2d996 2567=over 4
75e14d17
IZ
2568
2569=item Finding the end
2570
6deea57f
ST
2571The first pass is finding the end of the quoted construct, where
2572the information about the delimiters is used in parsing.
2573During this search, text between the starting and ending delimiters
2574is copied to a safe location. The text copied gets delimiter-independent.
2575
2576If the construct is a here-doc, the ending delimiter is a line
2577that has a terminating string as the content. Therefore C<<<EOF> is
2578terminated by C<EOF> immediately followed by C<"\n"> and starting
2579from the first column of the terminating line.
2580When searching for the terminating line of a here-doc, nothing
2581is skipped. In other words, lines after the here-doc syntax
2582are compared with the terminating string line by line.
2583
2584For the constructs except here-docs, single characters are used as starting
2585and ending delimiters. If the starting delimiter is an opening punctuation
2586(that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
2587corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
2588If the starting delimiter is an unpaired character like C</> or a closing
2589punctuation, the ending delimiter is same as the starting delimiter.
2590Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
2591C<qq[]> and C<qq]]> constructs.
2592
2593When searching for single-character delimiters, escaped delimiters
1ca345ed 2594and C<\\> are skipped. For example, while searching for terminating C</>,
6deea57f
ST
2595combinations of C<\\> and C<\/> are skipped. If the delimiters are
2596bracketing, nested pairs are also skipped. For example, while searching
2597for closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
2598and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
2599However, when backslashes are used as the delimiters (like C<qq\\> and
2600C<tr\\\>), nothing is skipped.
32581033 2601During the search for the end, backslashes that escape delimiters or
7188ca43 2602other backslashes are removed (exactly speaking, they are not copied to the
32581033 2603safe location).
75e14d17 2604
19799a22
GS
2605For constructs with three-part delimiters (C<s///>, C<y///>, and
2606C<tr///>), the search is repeated once more.
6deea57f
ST
2607If the first delimiter is not an opening punctuation, three delimiters must
2608be same such as C<s!!!> and C<tr)))>, in which case the second delimiter
2609terminates the left part and starts the right part at once.
b6538e4f 2610If the left part is delimited by bracketing punctuation (that is C<()>,
6deea57f 2611C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
b6538e4f 2612delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespace
6deea57f 2613and comments are allowed between both parts, though the comment must follow
b6538e4f
TC
2614at least one whitespace character; otherwise a character expected as the
2615start of the comment may be regarded as the starting delimiter of the right part.
75e14d17 2616
19799a22
GS
2617During this search no attention is paid to the semantics of the construct.
2618Thus:
75e14d17
IZ
2619
2620 "$hash{"$foo/$bar"}"
2621
2a94b7ce 2622or:
75e14d17 2623
89d205f2 2624 m/
2a94b7ce 2625 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
2626 /x
2627
19799a22
GS
2628do not form legal quoted expressions. The quoted part ends on the
2629first C<"> and C</>, and the rest happens to be a syntax error.
2630Because the slash that terminated C<m//> was followed by a C<SPACE>,
2631the example above is not C<m//x>, but rather C<m//> with no C</x>
2632modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 2633
89d205f2
YO
2634Also no attention is paid to C<\c\> (multichar control char syntax) during
2635this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
2636of C<\/>, and the following C</> is not recognized as a delimiter.
0d594e51
ST
2637Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
2638
75e14d17 2639=item Interpolation
d74e8afc 2640X<interpolation>
75e14d17 2641
19799a22 2642The next step is interpolation in the text obtained, which is now
89d205f2 2643delimiter-independent. There are multiple cases.
75e14d17 2644
13a2d996 2645=over 4
75e14d17 2646
89d205f2 2647=item C<<<'EOF'>
75e14d17
IZ
2648
2649No interpolation is performed.
6deea57f
ST
2650Note that the combination C<\\> is left intact, since escaped delimiters
2651are not available for here-docs.
75e14d17 2652
6deea57f 2653=item C<m''>, the pattern of C<s'''>
89d205f2 2654
6deea57f
ST
2655No interpolation is performed at this stage.
2656Any backslashed sequences including C<\\> are treated at the stage
2657to L</"parsing regular expressions">.
89d205f2 2658
6deea57f 2659=item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
75e14d17 2660
89d205f2 2661The only interpolation is removal of C<\> from pairs of C<\\>.
6deea57f
ST
2662Therefore C<-> in C<tr'''> and C<y'''> is treated literally
2663as a hyphen and no character range is available.
2664C<\1> in the replacement of C<s'''> does not work as C<$1>.
89d205f2
YO
2665
2666=item C<tr///>, C<y///>
2667
6deea57f
ST
2668No variable interpolation occurs. String modifying combinations for
2669case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
2670The other escape sequences such as C<\200> and C<\t> and backslashed
2671characters such as C<\\> and C<\-> are converted to appropriate literals.
89d205f2
YO
2672The character C<-> is treated specially and therefore C<\-> is treated
2673as a literal C<->.
75e14d17 2674
89d205f2 2675=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
75e14d17 2676
628253b8 2677C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> (possibly paired with C<\E>) are
19799a22
GS
2678converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
2679is converted to C<$foo . (quotemeta("baz" . $bar))> internally.
6deea57f
ST
2680The other escape sequences such as C<\200> and C<\t> and backslashed
2681characters such as C<\\> and C<\-> are replaced with appropriate
2682expansions.
2a94b7ce 2683
19799a22
GS
2684Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
2685is interpolated in the usual way. Something like C<"\Q\\E"> has
48cbae4f 2686no C<\E> inside. Instead, it has C<\Q>, C<\\>, and C<E>, so the
19799a22
GS
2687result is the same as for C<"\\\\E">. As a general rule, backslashes
2688between C<\Q> and C<\E> may lead to counterintuitive results. So,
2689C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
2690as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
2691
2692 $str = '\t';
2693 return "\Q$str";
2694
2695may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
2696
19799a22 2697Interpolated scalars and arrays are converted internally to the C<join> and
92d29cee 2698C<.> catenation operations. Thus, C<"$foo XXX '@arr'"> becomes:
75e14d17 2699
19799a22 2700 $foo . " XXX '" . (join $", @arr) . "'";
75e14d17 2701
19799a22 2702All operations above are performed simultaneously, left to right.
75e14d17 2703
19799a22
GS
2704Because the result of C<"\Q STRING \E"> has all metacharacters
2705quoted, there is no way to insert a literal C<$> or C<@> inside a
2706C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to became
2707C<"\\\$">; if not, it is interpreted as the start of an interpolated
2708scalar.
75e14d17 2709
19799a22 2710Note also that the interpolation code needs to make a decision on
89d205f2 2711where the interpolated scalar ends. For instance, whether
35f2feb0 2712C<< "a $b -> {c}" >> really means:
75e14d17
IZ
2713
2714 "a " . $b . " -> {c}";
2715
2a94b7ce 2716or:
75e14d17
IZ
2717
2718 "a " . $b -> {c};
2719
19799a22
GS
2720Most of the time, the longest possible text that does not include
2721spaces between components and which contains matching braces or
2722brackets. because the outcome may be determined by voting based
2723on heuristic estimators, the result is not strictly predictable.
2724Fortunately, it's usually correct for ambiguous cases.
75e14d17 2725
6deea57f 2726=item the replacement of C<s///>
75e14d17 2727
628253b8 2728Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> and interpolation
6deea57f
ST
2729happens as with C<qq//> constructs.
2730
2731It is at this step that C<\1> is begrudgingly converted to C<$1> in
2732the replacement text of C<s///>, in order to correct the incorrigible
2733I<sed> hackers who haven't picked up the saner idiom yet. A warning
2734is emitted if the C<use warnings> pragma or the B<-w> command-line flag
2735(that is, the C<$^W> variable) was set.
2736
2737=item C<RE> in C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
2738
628253b8 2739Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F>, C<\E>,
cc74c5bd
ST
2740and interpolation happens (almost) as with C<qq//> constructs.
2741
5d03b57c
KW
2742Processing of C<\N{...}> is also done here, and compiled into an intermediate
2743form for the regex compiler. (This is because, as mentioned below, the regex
2744compilation may be done at execution time, and C<\N{...}> is a compile-time
2745construct.)
2746
cc74c5bd
ST
2747However any other combinations of C<\> followed by a character
2748are not substituted but only skipped, in order to parse them
2749as regular expressions at the following step.
6deea57f 2750As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
1749ea0d 2751treated as an array symbol (for example C<@foo>),
6deea57f 2752even though the same text in C<qq//> gives interpolation of C<\c@>.
6deea57f 2753
e128ab2c
DM
2754Code blocks such as C<(?{BLOCK})> are handled by temporarily passing control
2755back to the perl parser, in a similar way that an interpolated array
2756subscript expression such as C<"foo$array[1+f("[xyz")]bar"> would be.
2757
6deea57f 2758Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
19799a22
GS
2759a C<#>-comment in a C<//x>-regular expression, no processing is
2760performed whatsoever. This is the first step at which the presence
2761of the C<//x> modifier is relevant.
2762
1749ea0d
ST
2763Interpolation in patterns has several quirks: C<$|>, C<$(>, C<$)>, C<@+>
2764and C<@-> are not interpolated, and constructs C<$var[SOMETHING]> are
2765voted (by several different estimators) to be either an array element
2766or C<$var> followed by an RE alternative. This is where the notation
19799a22
GS
2767C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
2768array element C<-9>, not as a regular expression from the variable
2769C<$arr> followed by a digit, which would be the interpretation of
2770C</$arr[0-9]/>. Since voting among different estimators may occur,
2771the result is not predictable.
2772
19799a22
GS
2773The lack of processing of C<\\> creates specific restrictions on
2774the post-processed text. If the delimiter is C</>, one cannot get
2775the combination C<\/> into the result of this step. C</> will
2776finish the regular expression, C<\/> will be stripped to C</> on
2777the previous step, and C<\\/> will be left as is. Because C</> is
2778equivalent to C<\/> inside a regular expression, this does not
2779matter unless the delimiter happens to be character special to the
2780RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
2781alphanumeric char, as in:
2a94b7ce
IZ
2782
2783 m m ^ a \s* b mmx;
2784
19799a22 2785In the RE above, which is intentionally obfuscated for illustration, the
6deea57f 2786delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
89d205f2 2787RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
19799a22
GS
2788reason you're encouraged to restrict your delimiters to non-alphanumeric,
2789non-whitespace choices.
75e14d17
IZ
2790
2791=back
2792
19799a22 2793This step is the last one for all constructs except regular expressions,
75e14d17
IZ
2794which are processed further.
2795
6deea57f
ST
2796=item parsing regular expressions
2797X<regexp, parse>
75e14d17 2798
19799a22 2799Previous steps were performed during the compilation of Perl code,
ac036724 2800but this one happens at run time, although it may be optimized to
19799a22 2801be calculated at compile time if appropriate. After preprocessing
6deea57f 2802described above, and possibly after evaluation if concatenation,
19799a22
GS
2803joining, casing translation, or metaquoting are involved, the
2804resulting I<string> is passed to the RE engine for compilation.
2805
2806Whatever happens in the RE engine might be better discussed in L<perlre>,
2807but for the sake of continuity, we shall do so here.
2808
2809This is another step where the presence of the C<//x> modifier is
2810relevant. The RE engine scans the string from left to right and
2811converts it to a finite automaton.
2812
2813Backslashed characters are either replaced with corresponding
2814literal strings (as with C<\{>), or else they generate special nodes
2815in the finite automaton (as with C<\b>). Characters special to the
2816RE engine (such as C<|>) generate corresponding nodes or groups of
2817nodes. C<(?#...)> comments are ignored. All the rest is either
2818converted to literal strings to match, or else is ignored (as is
2819whitespace and C<#>-style comments if C<//x> is present).
2820
2821Parsing of the bracketed character class construct, C<[...]>, is
2822rather different than the rule used for the rest of the pattern.
2823The terminator of this construct is found using the same rules as
2824for finding the terminator of a C<{}>-delimited construct, the only
2825exception being that C<]> immediately following C<[> is treated as
e128ab2c
DM
2826though preceded by a backslash.
2827
2828The terminator of runtime C<(?{...})> is found by temporarily switching
2829control to the perl parser, which should stop at the point where the
2830logically balancing terminating C<}> is found.
19799a22
GS
2831
2832It is possible to inspect both the string given to RE engine and the
2833resulting finite automaton. See the arguments C<debug>/C<debugcolor>
2834in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 2835switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
2836
2837=item Optimization of regular expressions
d74e8afc 2838X<regexp, optimization>
75e14d17 2839
7522fed5 2840This step is listed for completeness only. Since it does not change
75e14d17 2841semantics, details of this step are not documented and are subject
19799a22
GS
2842to change without notice. This step is performed over the finite
2843automaton that was generated during the previous pass.
2a94b7ce 2844
19799a22
GS
2845It is at this stage that C<split()> silently optimizes C</^/> to
2846mean C</^/m>.
75e14d17
IZ
2847
2848=back
2849
a0d0e21e 2850=head2 I/O Operators
d74e8afc
ITB
2851X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
2852X<< <> >> X<@ARGV>
a0d0e21e 2853
54310121 2854There are several I/O operators you should know about.
fbad3eb5 2855
7b8d334a 2856A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
2857double-quote interpolation. It is then interpreted as an external
2858command, and the output of that command is the value of the
e9c56f9b
JH
2859backtick string, like in a shell. In scalar context, a single string
2860consisting of all output is returned. In list context, a list of
2861values is returned, one per line of output. (You can set C<$/> to use
2862a different line terminator.) The command is executed each time the
2863pseudo-literal is evaluated. The status value of the command is
2864returned in C<$?> (see L<perlvar> for the interpretation of C<$?>).
2865Unlike in B<csh>, no translation is done on the return data--newlines
2866remain newlines. Unlike in any of the shells, single quotes do not
2867hide variable names in the command from interpretation. To pass a
2868literal dollar-sign through to the shell you need to hide it with a
2869backslash. The generalized form of backticks is C<qx//>. (Because
2870backticks always undergo shell expansion as well, see L<perlsec> for
2871security concerns.)
d74e8afc 2872X<qx> X<`> X<``> X<backtick> X<glob>
19799a22
GS
2873
2874In scalar context, evaluating a filehandle in angle brackets yields
2875the next line from that file (the newline, if any, included), or
2876C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
2877(sometimes known as file-slurp mode) and the file is empty, it
2878returns C<''> the first time, followed by C<undef> subsequently.
2879
2880Ordinarily you must assign the returned value to a variable, but
2881there is one situation where an automatic assignment happens. If
2882and only if the input symbol is the only thing inside the conditional
2883of a C<while> statement (even if disguised as a C<for(;;)> loop),
2884the value is automatically assigned to the global variable $_,
2885destroying whatever was there previously. (This may seem like an
2886odd thing to you, but you'll use the construct in almost every Perl
17b829fa 2887script you write.) The $_ variable is not implicitly localized.
19799a22
GS
2888You'll have to put a C<local $_;> before the loop if you want that
2889to happen.
2890
2891The following lines are equivalent:
a0d0e21e 2892
748a9306 2893 while (defined($_ = <STDIN>)) { print; }
7b8d334a 2894 while ($_ = <STDIN>) { print; }
a0d0e21e
LW
2895 while (<STDIN>) { print; }
2896 for (;<STDIN>;) { print; }
748a9306 2897 print while defined($_ = <STDIN>);
7b8d334a 2898 print while ($_ = <STDIN>);
a0d0e21e
LW
2899 print while <STDIN>;
2900
1ca345ed
TC
2901This also behaves similarly, but assigns to a lexical variable
2902instead of to C<$_>:
7b8d334a 2903
89d205f2 2904 while (my $line = <STDIN>) { print $line }
7b8d334a 2905
19799a22
GS
2906In these loop constructs, the assigned value (whether assignment
2907is automatic or explicit) is then tested to see whether it is
1ca345ed
TC
2908defined. The defined test avoids problems where the line has a string
2909value that would be treated as false by Perl; for example a "" or
19799a22
GS
2910a "0" with no trailing newline. If you really mean for such values
2911to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
2912
2913 while (($_ = <STDIN>) ne '0') { ... }
2914 while (<STDIN>) { last unless $_; ... }
2915
1ca345ed 2916In other boolean contexts, C<< <FILEHANDLE> >> without an
5ef4d93e 2917explicit C<defined> test or comparison elicits a warning if the
9f1b1f2d 2918C<use warnings> pragma or the B<-w>
19799a22 2919command-line switch (the C<$^W> variable) is in effect.
7b8d334a 2920
5f05dabc 2921The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
2922filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
2923in packages, where they would be interpreted as local identifiers
2924rather than global.) Additional filehandles may be created with
2925the open() function, amongst others. See L<perlopentut> and
2926L<perlfunc/open> for details on this.
d74e8afc 2927X<stdin> X<stdout> X<sterr>
a0d0e21e 2928
35f2feb0 2929If a <FILEHANDLE> is used in a context that is looking for
19799a22
GS
2930a list, a list comprising all input lines is returned, one line per
2931list element. It's easy to grow to a rather large data space this
2932way, so use with care.
a0d0e21e 2933
35f2feb0 2934<FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
19799a22 2935See L<perlfunc/readline>.
fbad3eb5 2936
35f2feb0 2937The null filehandle <> is special: it can be used to emulate the
1ca345ed
TC
2938behavior of B<sed> and B<awk>, and any other Unix filter program
2939that takes a list of filenames, doing the same to each line
2940of input from all of them. Input from <> comes either from
a0d0e21e 2941standard input, or from each file listed on the command line. Here's
35f2feb0 2942how it works: the first time <> is evaluated, the @ARGV array is
5a964f20 2943checked, and if it is empty, C<$ARGV[0]> is set to "-", which when opened
a0d0e21e
LW
2944gives you standard input. The @ARGV array is then processed as a list
2945of filenames. The loop
2946
2947 while (<>) {
2948 ... # code for each line
2949 }
2950
2951is equivalent to the following Perl-like pseudo code:
2952
3e3baf6d 2953 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
2954 while ($ARGV = shift) {
2955 open(ARGV, $ARGV);
2956 while (<ARGV>) {
2957 ... # code for each line
2958 }
2959 }
2960
19799a22
GS
2961except that it isn't so cumbersome to say, and will actually work.
2962It really does shift the @ARGV array and put the current filename
2963into the $ARGV variable. It also uses filehandle I<ARGV>
ac036724 2964internally. <> is just a synonym for <ARGV>, which
19799a22 2965is magical. (The pseudo code above doesn't work because it treats
35f2feb0 2966<ARGV> as non-magical.)
a0d0e21e 2967
48ab5743
ML
2968Since the null filehandle uses the two argument form of L<perlfunc/open>
2969it interprets special characters, so if you have a script like this:
2970
2971 while (<>) {
2972 print;
2973 }
2974
2975and call it with C<perl dangerous.pl 'rm -rfv *|'>, it actually opens a
2976pipe, executes the C<rm> command and reads C<rm>'s output from that pipe.
2977If you want all items in C<@ARGV> to be interpreted as file names, you
2978can use the module C<ARGV::readonly> from CPAN.
2979
35f2feb0 2980You can modify @ARGV before the first <> as long as the array ends up
a0d0e21e 2981containing the list of filenames you really want. Line numbers (C<$.>)
19799a22
GS
2982continue as though the input were one big happy file. See the example
2983in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20 2984
89d205f2 2985If you want to set @ARGV to your own list of files, go right ahead.
5a964f20
TC
2986This sets @ARGV to all plain text files if no @ARGV was given:
2987
2988 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 2989
5a964f20
TC
2990You can even set them to pipe commands. For example, this automatically
2991filters compressed arguments through B<gzip>:
2992
2993 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
2994
2995If you want to pass switches into your script, you can use one of the
a0d0e21e
LW
2996Getopts modules or put a loop on the front like this:
2997
2998 while ($_ = $ARGV[0], /^-/) {
2999 shift;
3000 last if /^--$/;
3001 if (/^-D(.*)/) { $debug = $1 }
3002 if (/^-v/) { $verbose++ }
5a964f20 3003 # ... # other switches
a0d0e21e 3004 }
5a964f20 3005
a0d0e21e 3006 while (<>) {
5a964f20 3007 # ... # code for each line
a0d0e21e
LW
3008 }
3009
89d205f2
YO
3010The <> symbol will return C<undef> for end-of-file only once.
3011If you call it again after this, it will assume you are processing another
19799a22 3012@ARGV list, and if you haven't set @ARGV, will read input from STDIN.
a0d0e21e 3013
1ca345ed 3014If what the angle brackets contain is a simple scalar variable (for example,
35f2feb0 3015<$foo>), then that variable contains the name of the
19799a22
GS
3016filehandle to input from, or its typeglob, or a reference to the
3017same. For example:
cb1a09d0
AD
3018
3019 $fh = \*STDIN;
3020 $line = <$fh>;
a0d0e21e 3021
5a964f20
TC
3022If what's within the angle brackets is neither a filehandle nor a simple
3023scalar variable containing a filehandle name, typeglob, or typeglob
3024reference, it is interpreted as a filename pattern to be globbed, and
3025either a list of filenames or the next filename in the list is returned,
19799a22 3026depending on context. This distinction is determined on syntactic
35f2feb0
GS
3027grounds alone. That means C<< <$x> >> is always a readline() from
3028an indirect handle, but C<< <$hash{key}> >> is always a glob().
5a964f20 3029That's because $x is a simple scalar variable, but C<$hash{key}> is
ef191992
YST
3030not--it's a hash element. Even C<< <$x > >> (note the extra space)
3031is treated as C<glob("$x ")>, not C<readline($x)>.
5a964f20
TC
3032
3033One level of double-quote interpretation is done first, but you can't
35f2feb0 3034say C<< <$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
3035in the previous paragraph. (In older versions of Perl, programmers
3036would insert curly brackets to force interpretation as a filename glob:
35f2feb0 3037C<< <${foo}> >>. These days, it's considered cleaner to call the
5a964f20 3038internal function directly as C<glob($foo)>, which is probably the right
19799a22 3039way to have done it in the first place.) For example:
a0d0e21e
LW
3040
3041 while (<*.c>) {
3042 chmod 0644, $_;
3043 }
3044
3a4b19e4 3045is roughly equivalent to:
a0d0e21e
LW
3046
3047 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
3048 while (<FOO>) {
5b3eff12 3049 chomp;
a0d0e21e
LW
3050 chmod 0644, $_;
3051 }
3052
3a4b19e4
GS
3053except that the globbing is actually done internally using the standard
3054C<File::Glob> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
3055
3056 chmod 0644, <*.c>;
3057
19799a22
GS
3058A (file)glob evaluates its (embedded) argument only when it is
3059starting a new list. All values must be read before it will start
3060over. In list context, this isn't important because you automatically
3061get them all anyway. However, in scalar context the operator returns
069e01df 3062the next value each time it's called, or C<undef> when the list has
19799a22
GS
3063run out. As with filehandle reads, an automatic C<defined> is
3064generated when the glob occurs in the test part of a C<while>,
1ca345ed
TC
3065because legal glob returns (for example,
3066a file called F<0>) would otherwise
19799a22
GS
3067terminate the loop. Again, C<undef> is returned only once. So if
3068you're expecting a single value from a glob, it is much better to
3069say
4633a7c4
LW
3070
3071 ($file) = <blurch*>;
3072
3073than
3074
3075 $file = <blurch*>;
3076
3077because the latter will alternate between returning a filename and
19799a22 3078returning false.
4633a7c4 3079
b159ebd3 3080If you're trying to do variable interpolation, it's definitely better
4633a7c4 3081to use the glob() function, because the older notation can cause people
e37d713d 3082to become confused with the indirect filehandle notation.
4633a7c4
LW
3083
3084 @files = glob("$dir/*.[ch]");
3085 @files = glob($files[$i]);
3086
a0d0e21e 3087=head2 Constant Folding
d74e8afc 3088X<constant folding> X<folding>
a0d0e21e
LW
3089
3090Like C, Perl does a certain amount of expression evaluation at
19799a22 3091compile time whenever it determines that all arguments to an
a0d0e21e
LW
3092operator are static and have no side effects. In particular, string
3093concatenation happens at compile time between literals that don't do
19799a22 3094variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
3095compile time. You can say
3096
1ca345ed
TC
3097 'Now is the time for all'
3098 . "\n"
3099 . 'good men to come to.'
a0d0e21e 3100
54310121 3101and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
3102you say
3103
3104 foreach $file (@filenames) {
5a964f20 3105 if (-s $file > 5 + 100 * 2**16) { }
54310121 3106 }
a0d0e21e 3107
1ca345ed 3108the compiler precomputes the number which that expression
19799a22 3109represents so that the interpreter won't have to.
a0d0e21e 3110
fd1abbef 3111=head2 No-ops
d74e8afc 3112X<no-op> X<nop>
fd1abbef
DN
3113
3114Perl doesn't officially have a no-op operator, but the bare constants
1ca345ed 3115C<0> and C<1> are special-cased not to produce a warning in void
fd1abbef
DN
3116context, so you can for example safely do
3117
3118 1 while foo();
3119
2c268ad5 3120=head2 Bitwise String Operators
d74e8afc 3121X<operator, bitwise, string>
2c268ad5
TP
3122
3123Bitstrings of any size may be manipulated by the bitwise operators
3124(C<~ | & ^>).
3125
19799a22
GS
3126If the operands to a binary bitwise op are strings of different
3127sizes, B<|> and B<^> ops act as though the shorter operand had
3128additional zero bits on the right, while the B<&> op acts as though
3129the longer operand were truncated to the length of the shorter.
3130The granularity for such extension or truncation is one or more
3131bytes.
2c268ad5 3132
89d205f2 3133 # ASCII-based examples
2c268ad5
TP
3134 print "j p \n" ^ " a h"; # prints "JAPH\n"
3135 print "JA" | " ph\n"; # prints "japh\n"
3136 print "japh\nJunk" & '_____'; # prints "JAPH\n";
3137 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
3138
19799a22 3139If you are intending to manipulate bitstrings, be certain that
2c268ad5 3140you're supplying bitstrings: If an operand is a number, that will imply
19799a22 3141a B<numeric> bitwise operation. You may explicitly show which type of
2c268ad5
TP
3142operation you intend by using C<""> or C<0+>, as in the examples below.
3143
4358a253
SS
3144 $foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
3145 $foo = '150' | 105; # yields 255
2c268ad5
TP
3146 $foo = 150 | '105'; # yields 255
3147 $foo = '150' | '105'; # yields string '155' (under ASCII)
3148
3149 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
3150 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
a0d0e21e 3151
1ae175c8
GS
3152See L<perlfunc/vec> for information on how to manipulate individual bits
3153in a bit vector.
3154
55497cff 3155=head2 Integer Arithmetic
d74e8afc 3156X<integer>
a0d0e21e 3157
19799a22 3158By default, Perl assumes that it must do most of its arithmetic in
a0d0e21e
LW
3159floating point. But by saying
3160
3161 use integer;
3162
3eab78e3
CW
3163you may tell the compiler to use integer operations
3164(see L<integer> for a detailed explanation) from here to the end of
3165the enclosing BLOCK. An inner BLOCK may countermand this by saying
a0d0e21e
LW
3166
3167 no integer;
3168
19799a22 3169which lasts until the end of that BLOCK. Note that this doesn't
3eab78e3
CW
3170mean everything is an integer, merely that Perl will use integer
3171operations for arithmetic, comparison, and bitwise operators. For
3172example, even under C<use integer>, if you take the C<sqrt(2)>, you'll
3173still get C<1.4142135623731> or so.
19799a22
GS
3174
3175Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
89d205f2 3176and ">>") always produce integral results. (But see also
13a2d996 3177L<Bitwise String Operators>.) However, C<use integer> still has meaning for
19799a22
GS
3178them. By default, their results are interpreted as unsigned integers, but
3179if C<use integer> is in effect, their results are interpreted
3180as signed integers. For example, C<~0> usually evaluates to a large
0be96356 3181integral value. However, C<use integer; ~0> is C<-1> on two's-complement
19799a22 3182machines.
68dc0745
PP
3183
3184=head2 Floating-point Arithmetic
06ce2fa3 3185
d74e8afc 3186X<floating-point> X<floating point> X<float> X<real>
68dc0745
PP
3187
3188While C<use integer> provides integer-only arithmetic, there is no
19799a22
GS
3189analogous mechanism to provide automatic rounding or truncation to a
3190certain number of decimal places. For rounding to a certain number
3191of digits, sprintf() or printf() is usually the easiest route.
3192See L<perlfaq4>.
68dc0745 3193
5a964f20
TC
3194Floating-point numbers are only approximations to what a mathematician
3195would call real numbers. There are infinitely more reals than floats,
3196so some corners must be cut. For example:
3197
3198 printf "%.20g\n", 123456789123456789;
3199 # produces 123456789123456784
3200
8548cb57
RGS
3201Testing for exact floating-point equality or inequality is not a
3202good idea. Here's a (relatively expensive) work-around to compare
5a964f20
TC
3203whether two floating-point numbers are equal to a particular number of
3204decimal places. See Knuth, volume II, for a more robust treatment of
3205this topic.
3206
3207 sub fp_equal {
3208 my ($X, $Y, $POINTS) = @_;
3209 my ($tX, $tY);
3210 $tX = sprintf("%.${POINTS}g", $X);
3211 $tY = sprintf("%.${POINTS}g", $Y);
3212 return $tX eq $tY;
3213 }
3214
68dc0745 3215The POSIX module (part of the standard perl distribution) implements
19799a22
GS
3216ceil(), floor(), and other mathematical and trigonometric functions.
3217The Math::Complex module (part of the standard perl distribution)
3218defines mathematical functions that work on both the reals and the
3219imaginary numbers. Math::Complex not as efficient as POSIX, but
68dc0745
PP
3220POSIX can't work with complex numbers.
3221
3222Rounding in financial applications can have serious implications, and
3223the rounding method used should be specified precisely. In these
3224cases, it probably pays not to trust whichever system rounding is
3225being used by Perl, but to instead implement the rounding function you
3226need yourself.
5a964f20
TC
3227
3228=head2 Bigger Numbers
d74e8afc 3229X<number, arbitrary precision>
5a964f20 3230
c543c01b 3231The standard C<Math::BigInt>, C<Math::BigRat>, and C<Math::BigFloat> modules,
fb1a95c6 3232along with the C<bignum>, C<bigint>, and C<bigrat> pragmas, provide
19799a22 3233variable-precision arithmetic and overloaded operators, although
cd5c4fce 3234they're currently pretty slow. At the cost of some space and
19799a22
GS
3235considerable speed, they avoid the normal pitfalls associated with
3236limited-precision representations.
5a964f20 3237
c543c01b
TC
3238 use 5.010;
3239 use bigint; # easy interface to Math::BigInt
3240 $x = 123456789123456789;
3241 say $x * $x;
3242 +15241578780673678515622620750190521
3243
3244Or with rationals:
3245
3246 use 5.010;
3247 use bigrat;
3248 $a = 3/22;
3249 $b = 4/6;
3250 say "a/b is ", $a/$b;
3251 say "a*b is ", $a*$b;
3252 a/b is 9/44
3253 a*b is 1/11
3254
3255Several modules let you calculate with (bound only by memory and CPU time)
3256unlimited or fixed precision. There are also some non-standard modules that
3257provide faster implementations via external C libraries.
cd5c4fce
T
3258
3259Here is a short, but incomplete summary:
3260
950b09ed
KW
3261 Math::String treat string sequences like numbers
3262 Math::FixedPrecision calculate with a fixed precision
3263 Math::Currency for currency calculations
3264 Bit::Vector manipulate bit vectors fast (uses C)
3265 Math::BigIntFast Bit::Vector wrapper for big numbers
3266 Math::Pari provides access to the Pari C library
70c45be3
FC
3267 Math::Cephes uses the external Cephes C library (no
3268 big numbers)
950b09ed
KW
3269 Math::Cephes::Fraction fractions via the Cephes library
3270 Math::GMP another one using an external C library
70c45be3
FC
3271 Math::GMPz an alternative interface to libgmp's big ints
3272 Math::GMPq an interface to libgmp's fraction numbers
3273 Math::GMPf an interface to libgmp's floating point numbers
cd5c4fce
T
3274
3275Choose wisely.
16070b82
GS
3276
3277=cut