This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
CommitLineData
d74e8afc 2X<operator>
a0d0e21e
LW
3
4perlop - Perl operators and precedence
5
d042e63d
MS
7
89d205f2 8=head2 Operator Precedence and Associativity
d74e8afc 9X<operator, precedence> X<precedence> X<associativity>
d042e63d
MS
10
11Operator precedence and associativity work in Perl more or less like
12they do in mathematics.
13
14I<Operator precedence> means some operators are evaluated before
15others. For example, in C<2 + 4 * 5>, the multiplication has higher
16precedence so C<4 * 5> is evaluated first yielding C<2 + 20 ==
1722> and not C<6 * 5 == 30>.
18
19I<Operator associativity> defines what happens if a sequence of the
20same operators is used one after another: whether the evaluator will
21evaluate the left operations first or the right. For example, in C<8
22- 4 - 2>, subtraction is left associative so Perl evaluates the
23expression left to right. C<8 - 4> is evaluated first making the
24expression C<4 - 2 == 2> and not C<8 - 2 == 6>.
a0d0e21e
LW
25
26Perl operators have the following associativity and precedence,
19799a22
GS
27listed from highest precedence to lowest. Operators borrowed from
28C keep the same precedence relationship with each other, even where
29C's precedence is slightly screwy. (This makes learning Perl easier
30for C folks.) With very few exceptions, these all operate on scalar
31values only, not array values.
a0d0e21e
LW
32
33 left terms and list operators (leftward)
34 left ->
35 nonassoc ++ --
36 right **
37 right ! ~ \ and unary + and -
54310121 38 left =~ !~
a0d0e21e
LW
39 left * / % x
40 left + - .
41 left << >>
42 nonassoc named unary operators
43 nonassoc < > <= >= lt gt le ge
0d863452 44 nonassoc == != <=> eq ne cmp ~~
a0d0e21e
LW
45 left &
46 left | ^
47 left &&
c963b151 48 left || //
137443ea 49 nonassoc .. ...
a0d0e21e
LW
50 right ?:
51 right = += -= *= etc.
52 left , =>
53 nonassoc list operators (rightward)
a5f75d66 54 right not
a0d0e21e 55 left and
f23102e2 56 left or xor
a0d0e21e
LW
57
58In the following sections, these operators are covered in precedence order.
59
5a964f20
TC
60Many operators can be overloaded for objects. See L<overload>.
61
a0d0e21e 62=head2 Terms and List Operators (Leftward)
d74e8afc 63X<list operator> X<operator, list> X<term>
a0d0e21e 64
62c18ce2 65A TERM has the highest precedence in Perl. They include variables,
5f05dabc 66quote and quote-like operators, any expression in parentheses,
a0d0e21e
LW
67and any function whose arguments are parenthesized. Actually, there
68aren't really functions in this sense, just list operators and unary
69operators behaving as functions because you put parentheses around
70the arguments. These are all documented in L<perlfunc>.
71
72If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
73is followed by a left parenthesis as the next token, the operator and
74arguments within parentheses are taken to be of highest precedence,
75just like a normal function call.
76
77In the absence of parentheses, the precedence of list operators such as
78C<print>, C<sort>, or C<chmod> is either very high or very low depending on
54310121 79whether you are looking at the left side or the right side of the operator.
a0d0e21e
LW
80For example, in
81
82 @ary = (1, 3, sort 4, 2);
83 print @ary; # prints 1324
84
19799a22
GS
85the commas on the right of the sort are evaluated before the sort,
86but the commas on the left are evaluated after. In other words,
87list operators tend to gobble up all arguments that follow, and
a0d0e21e 88then act like a simple TERM with regard to the preceding expression.
19799a22 89Be careful with parentheses:
a0d0e21e
LW
90
91 # These evaluate exit before doing the print:
92 print(\$foo, exit); # Obviously not what you want.
93 print \$foo, exit; # Nor is this.
94
95 # These do the print before evaluating exit:
96 (print \$foo), exit; # This is what you want.
97 print(\$foo), exit; # Or this.
98 print (\$foo), exit; # Or even this.
99
100Also note that
101
102 print (\$foo & 255) + 1, "\n";
103
d042e63d
MS
104probably doesn't do what you expect at first glance. The parentheses
105enclose the argument list for C<print> which is evaluated (printing
106the result of C<\$foo & 255>). Then one is added to the return value
107of C<print> (usually 1). The result is something like this:
108
109 1 + 1, "\n"; # Obviously not what you meant.
110
111To do what you meant properly, you must write:
112
113 print((\$foo & 255) + 1, "\n");
114
115See L<Named Unary Operators> for more discussion of this.
a0d0e21e
LW
116
117Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
54310121 118well as subroutine and method calls, and the anonymous
a0d0e21e
LW
119constructors C<[]> and C<{}>.
120
2ae324a7 121See also L<Quote and Quote-like Operators> toward the end of this section,
da87341d 122as well as L</"I/O Operators">.
a0d0e21e
LW
123
124=head2 The Arrow Operator
d74e8afc 125X<arrow> X<dereference> X<< -> >>
a0d0e21e 126
35f2feb0 127"C<< -> >>" is an infix dereference operator, just as it is in C
19799a22
GS
128and C++. If the right side is either a C<[...]>, C<{...}>, or a
129C<(...)> subscript, then the left side must be either a hard or
130symbolic reference to an array, a hash, or a subroutine respectively.
131(Or technically speaking, a location capable of holding a hard
132reference, if it's an array or hash reference being used for
133assignment.) See L<perlreftut> and L<perlref>.
a0d0e21e 134
19799a22
GS
135Otherwise, the right side is a method name or a simple scalar
136variable containing either the method name or a subroutine reference,
137and the left side must be either an object (a blessed reference)
138or a class name (that is, a package name). See L<perlobj>.
a0d0e21e 139
5f05dabc 140=head2 Auto-increment and Auto-decrement
d74e8afc 141X<increment> X<auto-increment> X<++> X<decrement> X<auto-decrement> X<-->
a0d0e21e 142
d042e63d
MS
143"++" and "--" work as in C. That is, if placed before a variable,
144they increment or decrement the variable by one before returning the
145value, and if placed after, increment or decrement after returning the
146value.
147
148 \$i = 0; \$j = 0;
149 print \$i++; # prints 0
150 print ++\$j; # prints 1
a0d0e21e 151
b033823e 152Note that just as in C, Perl doesn't define B<when> the variable is
89d205f2 153incremented or decremented. You just know it will be done sometime
b033823e
A
154before or after the value is returned. This also means that modifying
155a variable twice in the same statement will lead to undefined behaviour.
156Avoid statements like:
157
158 \$i = \$i ++;
159 print ++ \$i + \$i ++;
160
161Perl will not guarantee what the result of the above statements is.
162
54310121 163The auto-increment operator has a little extra builtin magic to it. If
a0d0e21e
LW
164you increment a variable that is numeric, or that has ever been used in
165a numeric context, you get a normal increment. If, however, the
5f05dabc 166variable has been used in only string contexts since it was set, and
5a964f20 167has a value that is not the empty string and matches the pattern
9c0670e1 168C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
a0d0e21e
LW
169character within its range, with carry:
170
171 print ++(\$foo = '99'); # prints '100'
172 print ++(\$foo = 'a0'); # prints 'a1'
173 print ++(\$foo = 'Az'); # prints 'Ba'
174 print ++(\$foo = 'zz'); # prints 'aaa'
175
6a61d433
HS
176C<undef> is always treated as numeric, and in particular is changed
177to C<0> before incrementing (so that a post-increment of an undef value
178will return C<0> rather than C<undef>).
179
5f05dabc 180The auto-decrement operator is not magical.
a0d0e21e
LW
181
d74e8afc 183X<**> X<exponentiation> X<power>
a0d0e21e 184
19799a22 185Binary "**" is the exponentiation operator. It binds even more
cb1a09d0
186tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
187implemented using C's pow(3) function, which actually works on doubles
188internally.)
a0d0e21e
LW
189
190=head2 Symbolic Unary Operators
d74e8afc 191X<unary operator> X<operator, unary>
a0d0e21e 192
5f05dabc 193Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
a0d0e21e 194precedence version of this.
d74e8afc 195X<!>
a0d0e21e 196
da2f94c5
FC
197Unary "-" performs arithmetic negation if the operand is numeric,
198including any string that looks like a number. If the operand is
199an identifier, a string consisting of a minus sign concatenated
200with the identifier is returned. Otherwise, if the string starts
201with a plus or minus, a string starting with the opposite sign is
202returned. One effect of these rules is that -bareword is equivalent
8705167b 203to the string "-bareword". If, however, the string begins with a
353c6505 204non-alphabetic character (excluding "+" or "-"), Perl will attempt to convert
06705523
SP
205the string to a numeric and the arithmetic negation is performed. If the
206string cannot be cleanly converted to a numeric, Perl will give the warning
207B<Argument "the string" isn't numeric in negation (-) at ...>.
d74e8afc 208X<-> X<negation, arithmetic>
a0d0e21e 209
972b05a9
JH
210Unary "~" performs bitwise negation, i.e., 1's complement. For
211example, C<0666 & ~027> is 0640. (See also L<Integer Arithmetic> and
212L<Bitwise String Operators>.) Note that the width of the result is
213platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
214bits wide on a 64-bit platform, so if you are expecting a certain bit
d042e63d 215width, remember to use the & operator to mask off the excess bits.
d74e8afc 216X<~> X<negation, binary>
a0d0e21e
LW
217
218Unary "+" has no effect whatsoever, even on strings. It is useful
219syntactically for separating a function name from a parenthesized expression
220that would otherwise be interpreted as the complete list of function
5ba421f6 221arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
d74e8afc 222X<+>
a0d0e21e 223
19799a22
GS
224Unary "\" creates a reference to whatever follows it. See L<perlreftut>
225and L<perlref>. Do not confuse this behavior with the behavior of
226backslash within a string, although both forms do convey the notion
227of protecting the next thing from interpolation.
d74e8afc 228X<\> X<reference> X<backslash>
a0d0e21e
LW
229
d74e8afc 231X<binding> X<operator, binding> X<=~> X<!~>
a0d0e21e 232
c07a80fd 233Binary "=~" binds a scalar expression to a pattern match. Certain operations
cb1a09d0
234search or modify the string \$_ by default. This operator makes that kind
235of operation work on some other string. The right argument is a search
TP
236pattern, substitution, or transliteration. The left argument is what is
237supposed to be searched, substituted, or transliterated instead of the default
f8bab1e9 238\$_. When used in scalar context, the return value generally indicates the
8ff32507
FC
239success of the operation. The exceptions are substitution (s///)
240and transliteration (y///) with the C</r> (non-destructive) option,
241which cause the B<r>eturn value to be the result of the substitution.
242Behavior in list context depends on the particular operator.
000c65fc
DG
243See L</"Regexp Quote-Like Operators"> for details and L<perlretut> for
244examples using these operators.
f8bab1e9
GS
245
246If the right argument is an expression rather than a search pattern,
2c268ad5 247substitution, or transliteration, it is interpreted as a search pattern at run
89d205f2
YO
248time. Note that this means that its contents will be interpolated twice, so
249
250 '\\' =~ q'\\';
251
252is not ok, as the regex engine will end up trying to compile the
253pattern C<\>, which it will consider a syntax error.
a0d0e21e
LW
254
255Binary "!~" is just like "=~" except the return value is negated in
256the logical sense.
257
8ff32507
FC
258Binary "!~" with a non-destructive substitution (s///r) or transliteration
259(y///r) is a syntax error.
4f4d7508 260
a0d0e21e 261=head2 Multiplicative Operators
d74e8afc 262X<operator, multiplicative>
a0d0e21e
LW
263
264Binary "*" multiplies two numbers.
d74e8afc 265X<*>
a0d0e21e
LW
266
267Binary "/" divides two numbers.
d74e8afc 268X</> X<slash>
a0d0e21e 269
f7918450
KW
270Binary "%" is the modulo operator, which computes the division
271remainder of its first argument with respect to its second argument.
272Given integer
54310121 273operands C<\$a> and C<\$b>: If C<\$b> is positive, then C<\$a % \$b> is
f7918450 274C<\$a> minus the largest multiple of C<\$b> less than or equal to
54310121 275C<\$a>. If C<\$b> is negative, then C<\$a % \$b> is C<\$a> minus the
276smallest multiple of C<\$b> that is not less than C<\$a> (i.e. the
89b4f0ad 277result will be less than or equal to zero). If the operands
4848a83b
TS
278C<\$a> and C<\$b> are floating point values and the absolute value of
279C<\$b> (that is C<abs(\$b)>) is less than C<(UV_MAX + 1)>, only
280the integer portion of C<\$a> and C<\$b> will be used in the operation
281(Note: here C<UV_MAX> means the maximum of the unsigned integer type).
282If the absolute value of the right operand (C<abs(\$b)>) is greater than
283or equal to C<(UV_MAX + 1)>, "%" computes the floating-point remainder
284C<\$r> in the equation C<(\$r = \$a - \$i*\$b)> where C<\$i> is a certain
f7918450 285integer that makes C<\$r> have the same sign as the right operand
4848a83b
TS
286C<\$b> (B<not> as the left operand C<\$a> like C function C<fmod()>)
287and the absolute value less than that of C<\$b>.
0412d526 288Note that when C<use integer> is in scope, "%" gives you direct access
f7918450 289to the modulo operator as implemented by your C compiler. This
55d729e4
GS
290operator is not as well defined for negative operands, but it will
291execute faster.
f7918450 292X<%> X<remainder> X<modulo> X<mod>
55d729e4 293
62d10b70
GS
294Binary "x" is the repetition operator. In scalar context or if the left
295operand is not enclosed in parentheses, it returns a string consisting
296of the left operand repeated the number of times specified by the right
297operand. In list context, if the left operand is enclosed in
3585017f
YST
298parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
299If the right operand is zero or negative, it returns an empty string
300or an empty list, depending on the context.
d74e8afc 301X<x>
a0d0e21e
LW
302
303 print '-' x 80; # print row of dashes
304
305 print "\t" x (\$tab/8), ' ' x (\$tab%8); # tab over
306
307 @ones = (1) x 80; # a list of 80 1's
308 @ones = (5) x @ones; # set all elements to 5
309
310
a0d0e21e
LW
313
314Binary "+" returns the sum of two numbers.
d74e8afc 315X<+>
a0d0e21e
LW
316
317Binary "-" returns the difference of two numbers.
d74e8afc 318X<->
a0d0e21e
LW
319
320Binary "." concatenates two strings.
d74e8afc
ITB
321X<string, concatenation> X<concatenation>
322X<cat> X<concat> X<concatenate> X<.>
a0d0e21e
LW
323
d74e8afc
ITB
325X<shift operator> X<operator, shift> X<<< << >>>
326X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
327X<shl> X<shr> X<shift, right> X<shift, left>
a0d0e21e 328
55497cff 329Binary "<<" returns the value of its left argument shifted left by the
330number of bits specified by the right argument. Arguments should be
a0d0e21e 332
55497cff 333Binary ">>" returns the value of its left argument shifted right by
334the number of bits specified by the right argument. Arguments should
982ce180 335be integers. (See also L<Integer Arithmetic>.)
a0d0e21e 336
b16cf6df
JH
337Note that both "<<" and ">>" in Perl are implemented directly using
338"<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
339in force then signed C integers are used, else unsigned C integers are
340used. Either way, the implementation isn't going to generate results
341larger than the size of the integer type Perl was built with (32 bits
342or 64 bits).
343
344The result of overflowing the range of the integers is undefined
345because it is undefined also in C. In other words, using 32-bit
346integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
347of bits is also undefined.
348
a0d0e21e 349=head2 Named Unary Operators
d74e8afc 350X<operator, named unary>
a0d0e21e
LW
351
352The various named unary operators are treated as functions with one
568e6d8b 353argument, with optional parentheses.
a0d0e21e
LW
354
355If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
356is followed by a left parenthesis as the next token, the operator and
357arguments within parentheses are taken to be of highest precedence,
3981b0eb
JA
358just like a normal function call. For example,
359because named unary operators are higher precedence than ||:
a0d0e21e
LW
360
361 chdir \$foo || die; # (chdir \$foo) || die
362 chdir(\$foo) || die; # (chdir \$foo) || die
363 chdir (\$foo) || die; # (chdir \$foo) || die
364 chdir +(\$foo) || die; # (chdir \$foo) || die
365
3981b0eb 366but, because * is higher precedence than named operators:
a0d0e21e
LW
367
368 chdir \$foo * 20; # chdir (\$foo * 20)
369 chdir(\$foo) * 20; # (chdir \$foo) * 20
370 chdir (\$foo) * 20; # (chdir \$foo) * 20
371 chdir +(\$foo) * 20; # chdir (\$foo * 20)
372
373 rand 10 * 20; # rand (10 * 20)
374 rand(10) * 20; # (rand 10) * 20
375 rand (10) * 20; # (rand 10) * 20
376 rand +(10) * 20; # rand (10 * 20)
377
568e6d8b
RGS
378Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
379treated like named unary operators, but they don't follow this functional
380parenthesis rule. That means, for example, that C<-f(\$file).".bak"> is
381equivalent to C<-f "\$file.bak">.
d74e8afc 382X<-X> X<filetest> X<operator, filetest>
568e6d8b 383
5ba421f6 384See also L<"Terms and List Operators (Leftward)">.
a0d0e21e
LW
385
d74e8afc 387X<relational operator> X<operator, relational>
a0d0e21e 388
35f2feb0 389Binary "<" returns true if the left argument is numerically less than
a0d0e21e 390the right argument.
d74e8afc 391X<< < >>
a0d0e21e 392
35f2feb0 393Binary ">" returns true if the left argument is numerically greater
a0d0e21e 394than the right argument.
d74e8afc 395X<< > >>
a0d0e21e 396
35f2feb0 397Binary "<=" returns true if the left argument is numerically less than
a0d0e21e 398or equal to the right argument.
d74e8afc 399X<< <= >>
a0d0e21e 400
35f2feb0 401Binary ">=" returns true if the left argument is numerically greater
a0d0e21e 402than or equal to the right argument.
d74e8afc 403X<< >= >>
a0d0e21e
LW
404
405Binary "lt" returns true if the left argument is stringwise less than
406the right argument.
d74e8afc 407X<< lt >>
a0d0e21e
LW
408
409Binary "gt" returns true if the left argument is stringwise greater
410than the right argument.
d74e8afc 411X<< gt >>
a0d0e21e
LW
412
413Binary "le" returns true if the left argument is stringwise less than
414or equal to the right argument.
d74e8afc 415X<< le >>
a0d0e21e
LW
416
417Binary "ge" returns true if the left argument is stringwise greater
418than or equal to the right argument.
d74e8afc 419X<< ge >>
a0d0e21e
LW
420
d74e8afc 422X<equality> X<equal> X<equals> X<operator, equality>
a0d0e21e
LW
423
424Binary "==" returns true if the left argument is numerically equal to
425the right argument.
d74e8afc 426X<==>
a0d0e21e
LW
427
428Binary "!=" returns true if the left argument is numerically not equal
429to the right argument.
d74e8afc 430X<!=>
a0d0e21e 431
35f2feb0 432Binary "<=>" returns -1, 0, or 1 depending on whether the left
6ee5d4e7 433argument is numerically less than, equal to, or greater than the right
d4ad863d 434argument. If your platform supports NaNs (not-a-numbers) as numeric
7d3a9d88
NC
435values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
436"<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
437returns true, as does NaN != anything else. If your platform doesn't
438support NaNs then NaN is just a string with numeric value 0.
d74e8afc 439X<< <=> >> X<spaceship>
7d3a9d88 440
2b54f59f
YST
441 perl -le '\$a = "NaN"; print "No NaN support here" if \$a == \$a'
442 perl -le '\$a = "NaN"; print "NaN support here" if \$a != \$a'
a0d0e21e
LW
443
444Binary "eq" returns true if the left argument is stringwise equal to
445the right argument.
d74e8afc 446X<eq>
a0d0e21e
LW
447
448Binary "ne" returns true if the left argument is stringwise not equal
449to the right argument.
d74e8afc 450X<ne>
a0d0e21e 451
JH
452Binary "cmp" returns -1, 0, or 1 depending on whether the left
453argument is stringwise less than, equal to, or greater than the right
454argument.
d74e8afc 455X<cmp>
a0d0e21e 456
0d863452 457Binary "~~" does a smart match between its arguments. Smart matching
0f7107a0 458is described in L<perlsyn/"Smart matching in detail">.
0d863452
RH
459X<~~>
460
a034a98d
DD
461"lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
462by the current locale if C<use locale> is in effect. See L<perllocale>.
463
a0d0e21e 464=head2 Bitwise And
d74e8afc 465X<operator, bitwise, and> X<bitwise and> X<&>
a0d0e21e 466
2cdc098b 467Binary "&" returns its operands ANDed together bit by bit.
a0d0e21e 469
2cdc098b
MG
470Note that "&" has lower priority than relational operators, so for example
471the brackets are essential in a test like
472
473 print "Even\n" if (\$x & 1) == 0;
474
a0d0e21e 475=head2 Bitwise Or and Exclusive Or
d74e8afc
ITB
476X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
477X<bitwise xor> X<^>
a0d0e21e 478
2cdc098b 479Binary "|" returns its operands ORed together bit by bit.
a0d0e21e 481
2cdc098b 482Binary "^" returns its operands XORed together bit by bit.
a0d0e21e 484
2cdc098b
MG
485Note that "|" and "^" have lower priority than relational operators, so
486for example the brackets are essential in a test like
487
488 print "false\n" if (8 | 2) != 10;
489
a0d0e21e 490=head2 C-style Logical And
d74e8afc 491X<&&> X<logical and> X<operator, logical, and>
a0d0e21e
LW
492
493Binary "&&" performs a short-circuit logical AND operation. That is,
494if the left operand is false, the right operand is not even evaluated.
495Scalar or list context propagates down to the right operand if it
496is evaluated.
497
498=head2 C-style Logical Or
d74e8afc 499X<||> X<operator, logical, or>
a0d0e21e
LW
500
501Binary "||" performs a short-circuit logical OR operation. That is,
502if the left operand is true, the right operand is not even evaluated.
503Scalar or list context propagates down to the right operand if it
504is evaluated.
505
c963b151 506=head2 C-style Logical Defined-Or
d74e8afc 507X<//> X<operator, logical, defined-or>
c963b151
BD
508
509Although it has no direct equivalent in C, Perl's C<//> operator is related
89d205f2 510to its C-style or. In fact, it's exactly the same as C<||>, except that it
c963b151 511tests the left hand side's definedness instead of its truth. Thus, C<\$a // \$b>
89d205f2 512is similar to C<defined(\$a) || \$b> (except that it returns the value of C<\$a>
d90d5a38 513rather than the value of C<defined(\$a)>) and yields the same result as
bdc7923b
RGS
514C<defined(\$a) ? \$a : \$b> (except that the ternary-operator form can be
515used as a lvalue, while C<\$a // \$b> cannot). This is very useful for
516providing default values for variables. If you actually want to test if
517at least one of C<\$a> and C<\$b> is defined, use C<defined(\$a // \$b)>.
c963b151 518
d042e63d
MS
519The C<||>, C<//> and C<&&> operators return the last value evaluated
520(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
521portable way to find out the home directory might be:
a0d0e21e 522
c963b151
BD
523 \$home = \$ENV{'HOME'} // \$ENV{'LOGDIR'} //
524 (getpwuid(\$<))[7] // die "You're homeless!\n";
a0d0e21e 525
5a964f20
TC
526In particular, this means that you shouldn't use this
527for selecting between two aggregates for assignment:
528
529 @a = @b || @c; # this is wrong
530 @a = scalar(@b) || @c; # really meant this
531 @a = @b ? @b : @c; # this works fine, though
532
f23102e2
RGS
533As more readable alternatives to C<&&> and C<||> when used for
534control flow, Perl provides the C<and> and C<or> operators (see below).
535The short-circuit behavior is identical. The precedence of "and"
c963b151 536and "or" is much lower, however, so that you can safely use them after a
5a964f20 537list operator without the need for parentheses:
a0d0e21e
LW
538
539 unlink "alpha", "beta", "gamma"
540 or gripe(), next LINE;
541
542With the C-style operators that would have been written like this:
543
544 unlink("alpha", "beta", "gamma")
545 || (gripe(), next LINE);
546
eeb6a2c9 547Using "or" for assignment is unlikely to do what you want; see below.
5a964f20
TC
548
d74e8afc 550X<operator, range> X<range> X<..> X<...>
a0d0e21e
LW
551
552Binary ".." is the range operator, which is really two different
fb53bbb2 553operators depending on the context. In list context, it returns a
54ae734e 554list of values counting (up by ones) from the left value to the right
2cdbc966 555value. If the left value is greater than the right value then it
fb53bbb2 556returns the empty list. The range operator is useful for writing
54ae734e 557C<foreach (1..10)> loops and for doing slice operations on arrays. In
2cdbc966
JD
558the current implementation, no temporary array is created when the
559range operator is used as the expression in C<foreach> loops, but older
560versions of Perl might burn a lot of memory when you write something
561like this:
a0d0e21e
LW
562
563 for (1 .. 1_000_000) {
564 # code
54310121 565 }
a0d0e21e 566
8f0f46f8 567The range operator also works on strings, using the magical
568auto-increment, see below.
54ae734e 569
5a964f20 570In scalar context, ".." returns a boolean value. The operator is
8f0f46f8 571bistable, like a flip-flop, and emulates the line-range (comma)
572operator of B<sed>, B<awk>, and various editors. Each ".." operator
573maintains its own boolean state, even across calls to a subroutine
574that contains it. It is false as long as its left operand is false.
a0d0e21e
LW
575Once the left operand is true, the range operator stays true until the
576right operand is true, I<AFTER> which the range operator becomes false
8f0f46f8 577again. It doesn't become false till the next time the range operator
578is evaluated. It can test the right operand and become false on the
579same evaluation it became true (as in B<awk>), but it still returns
580true once. If you don't want it to test the right operand until the
581next evaluation, as in B<sed>, just use three dots ("...") instead of
19799a22
GS
582two. In all other regards, "..." behaves just like ".." does.
583
584The right operand is not evaluated while the operator is in the
585"false" state, and the left operand is not evaluated while the
586operator is in the "true" state. The precedence is a little lower
587than || and &&. The value returned is either the empty string for
8f0f46f8 588false, or a sequence number (beginning with 1) for true. The sequence
589number is reset for each range encountered. The final sequence number
590in a range has the string "E0" appended to it, which doesn't affect
591its numeric value, but gives you something to search for if you want
592to exclude the endpoint. You can exclude the beginning point by
593waiting for the sequence number to be greater than 1.
df5f8116
CW
594
595If either operand of scalar ".." is a constant expression,
596that operand is considered true if it is equal (C<==>) to the current
597input line number (the C<\$.> variable).
598
599To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
600but that is only an issue if you use a floating point expression; when
601implicitly using C<\$.> as described in the previous paragraph, the
602comparison is C<int(EXPR) == int(\$.)> which is only an issue when C<\$.>
603is set to a floating point value and you are not reading from a file.
604Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
605you want in scalar context because each of the operands are evaluated
606using their integer representation.
607
608Examples:
a0d0e21e
LW
609
610As a scalar operator:
611
df5f8116 612 if (101 .. 200) { print; } # print 2nd hundred lines, short for
950b09ed 613 # if (\$. == 101 .. \$. == 200) { print; }
9f10b797
RGS
614
615 next LINE if (1 .. /^\$/); # skip header lines, short for
f343f960 616 # next LINE if (\$. == 1 .. /^\$/);
9f10b797
RGS
617 # (typically in a loop labeled LINE)
618
619 s/^/> / if (/^\$/ .. eof()); # quote body
a0d0e21e 620
5a964f20
TC
621 # parse mail messages
622 while (<>) {
623 \$in_header = 1 .. /^\$/;
df5f8116
CW
624 \$in_body = /^\$/ .. eof;
625 if (\$in_header) {
f343f960 626 # do something
df5f8116 627 } else { # in body
f343f960 628 # do something else
df5f8116 629 }
5a964f20 630 } continue {
df5f8116 631 close ARGV if eof; # reset \$. each file
5a964f20
TC
632 }
633
acf31ca5
SF
634Here's a simple example to illustrate the difference between
635the two range operators:
636
637 @lines = (" - Foo",
638 "01 - Bar",
639 "1 - Baz",
640 " - Quux");
641
9f10b797
RGS
642 foreach (@lines) {
643 if (/0/ .. /1/) {
acf31ca5
SF
644 print "\$_\n";
645 }
646 }
647
9f10b797
RGS
648This program will print only the line containing "Bar". If
649the range operator is changed to C<...>, it will also print the
acf31ca5
SF
650"Baz" line.
651
652And now some examples as a list operator:
a0d0e21e
LW
653
654 for (101 .. 200) { print; } # print \$_ 100 times
3e3baf6d 655 @foo = @foo[0 .. \$#foo]; # an expensive no-op
a0d0e21e
LW
656 @foo = @foo[\$#foo-4 .. \$#foo]; # slice last 5 items
657
5a964f20 658The range operator (in list context) makes use of the magical
5f05dabc 659auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
660can say
661
662 @alphabet = ('A' .. 'Z');
663
54ae734e 664to get all normal letters of the English alphabet, or
a0d0e21e
LW
665
666 \$hexdigit = (0 .. 9, 'a' .. 'f')[\$num & 15];
667
668to get a hexadecimal digit, or
669
670 @z2 = ('01' .. '31'); print \$z2[\$mday];
671
ea4f5703
YST
672to get dates with leading zeros.
673
674If the final value specified is not in the sequence that the magical
675increment would produce, the sequence goes until the next value would
676be longer than the final value specified.
677
678If the initial value specified isn't part of a magical increment
679sequence (that is, a non-empty string matching "/^[a-zA-Z]*[0-9]*\z/"),
680only the initial value will be returned. So the following will only
681return an alpha:
682
683 use charnames 'greek';
684 my @greek_small = ("\N{alpha}" .. "\N{omega}");
685
686To get lower-case greek letters, use this instead:
687
950b09ed
KW
688 my @greek_small = map { chr } ( ord("\N{alpha}") ..
689 ord("\N{omega}") );
a0d0e21e 690
df5f8116
CW
691Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
692return two elements in list context.
693
694 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
695
a0d0e21e 696=head2 Conditional Operator
d74e8afc 697X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
a0d0e21e
LW
698
699Ternary "?:" is the conditional operator, just as in C. It works much
700like an if-then-else. If the argument before the ? is true, the
701argument before the : is returned, otherwise the argument after the :
cb1a09d0
702is returned. For example:
703
54310121 704 printf "I have %d dog%s.\n", \$n,
cb1a09d0
705 (\$n == 1) ? '' : "s";
706
707Scalar or list context propagates downward into the 2nd
54310121 708or 3rd argument, whichever is selected.
cb1a09d0
709
710 \$a = \$ok ? \$b : \$c; # get a scalar
711 @a = \$ok ? @b : @c; # get an array
712 \$a = \$ok ? @b : @c; # oops, that's just a count!
713
714The operator may be assigned to if both the 2nd and 3rd arguments are
715legal lvalues (meaning that you can assign to them):
a0d0e21e
LW
716
717 (\$a_or_b ? \$a : \$b) = \$c;
718
5a964f20
TC
719Because this operator produces an assignable result, using assignments
720without parentheses will get you in trouble. For example, this:
721
722 \$a % 2 ? \$a += 10 : \$a += 2
723
724Really means this:
725
726 ((\$a % 2) ? (\$a += 10) : \$a) += 2
727
728Rather than this:
729
730 (\$a % 2) ? (\$a += 10) : (\$a += 2)
731
19799a22
GS
732That should probably be written more simply as:
733
734 \$a += (\$a % 2) ? 10 : 2;
735
4633a7c4 736=head2 Assignment Operators
d74e8afc 737X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
5ac3b81c 738X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
d74e8afc 739X<%=> X<^=> X<x=>
a0d0e21e
LW
740
741"=" is the ordinary assignment operator.
742
743Assignment operators work as in C. That is,
744
745 \$a += 2;
746
747is equivalent to
748
749 \$a = \$a + 2;
750
751although without duplicating any side effects that dereferencing the lvalue
54310121 752might trigger, such as from tie(). Other assignment operators work similarly.
753The following are recognized:
a0d0e21e
LW
754
755 **= += *= &= <<= &&=
9f10b797
RGS
756 -= /= |= >>= ||=
757 .= %= ^= //=
758 x=
a0d0e21e 759
19799a22 760Although these are grouped by family, they all have the precedence
a0d0e21e
LW
761of assignment.
762
b350dd2f
GS
763Unlike in C, the scalar assignment operator produces a valid lvalue.
764Modifying an assignment is equivalent to doing the assignment and
765then modifying the variable that was assigned to. This is useful
766for modifying a copy of something, like this:
a0d0e21e
LW
767
768 (\$tmp = \$global) =~ tr [A-Z] [a-z];
769
770Likewise,
771
772 (\$a += 2) *= 3;
773
774is equivalent to
775
776 \$a += 2;
777 \$a *= 3;
778
b350dd2f
GS
779Similarly, a list assignment in list context produces the list of
780lvalues assigned to, and a list assignment in scalar context returns
781the number of elements produced by the expression on the right hand
782side of the assignment.
783
748a9306 784=head2 Comma Operator
d74e8afc 785X<comma> X<operator, comma> X<,>
a0d0e21e 786
5a964f20 787Binary "," is the comma operator. In scalar context it evaluates
a0d0e21e
LW
788its left argument, throws that value away, then evaluates its right
789argument and returns that value. This is just like C's comma operator.
790
5a964f20 791In list context, it's just the list argument separator, and inserts
ed5c6d31
PJ
792both its arguments into the list. These arguments are also evaluated
793from left to right.
a0d0e21e 794
344f2c40
IG
795The C<< => >> operator is a synonym for the comma except that it causes
796its left operand to be interpreted as a string if it begins with a letter
797or underscore and is composed only of letters, digits and underscores.
798This includes operands that might otherwise be interpreted as operators,
799constants, single number v-strings or function calls. If in doubt about
800this behaviour, the left operand can be quoted explicitly.
801
802Otherwise, the C<< => >> operator behaves exactly as the comma operator
803or list argument separator, according to context.
804
805For example:
a44e5664
MS
806
807 use constant FOO => "something";
808
809 my %h = ( FOO => 23 );
810
811is equivalent to:
812
813 my %h = ("FOO", 23);
814
815It is I<NOT>:
816
817 my %h = ("something", 23);
818
719b43e8
RGS
819The C<< => >> operator is helpful in documenting the correspondence
820between keys and values in hashes, and other paired elements in lists.
748a9306 821
a44e5664
MS
822 %hash = ( \$key => \$value );
824
678ae90b
RGS
be25f609 827
e8163f9b 828The yada yada operator (noted C<...>) is a placeholder for code. Perl
829parses it without error, but when you try to execute a yada yada, it
830throws an exception with the text C<Unimplemented>:
831
832 sub unimplemented { ... }
833
834 eval { unimplemented() };
835 if( \$@ eq 'Unimplemented' ) {
836 print "I found the yada yada!\n";
837 }
838
839You can only use the yada yada to stand in for a complete statement.
840These examples of the yada yada work:
841
842 { ... }
843
844 sub foo { ... }
845
846 ...;
847
848 eval { ... };
849
850 sub foo {
851 my( \$self ) = shift;
852
853 ...;
854 }
855
856 do { my \$n; ...; print 'Hurrah!' };
857
858The yada yada cannot stand in for an expression that is part of a
859larger statement since the C<...> is also the three-dot version of the
860range operator (see L<Range Operators>). These examples of the yada
861yada are still syntax errors:
862
863 print ...;
864
865 open my(\$fh), '>', '/dev/passwd' or ...;
866
867 if( \$condition && ... ) { print "Hello\n" };
868
869There are some cases where Perl can't immediately tell the difference
870between an expression and a statement. For instance, the syntax for a
871block and an anonymous hash reference constructor look the same unless
872there's something in the braces that give Perl a hint. The yada yada
873is a syntax error if Perl doesn't guess that the C<{ ... }> is a
874block. In that case, it doesn't think the C<...> is the yada yada
875because it's expecting an expression instead of a statement:
876
877 my @transformed = map { ... } @input; # syntax error
878
879You can use a C<;> inside your block to denote that the C<{ ... }> is
880a block and not a hash reference constructor. Now the yada yada works:
881
882 my @transformed = map {; ... } @input; # ; disambiguates
883
884 my @transformed = map { ...; } @input; # ; disambiguates
be25f609 885
a0d0e21e 886=head2 List Operators (Rightward)
d74e8afc 887X<operator, list, rightward> X<list operator>
a0d0e21e
LW
888
889On the right side of a list operator, it has very low precedence,
890such that it controls all comma-separated expressions found there.
891The only operators with lower precedence are the logical operators
892"and", "or", and "not", which may be used to evaluate calls to list
893operators without the need for extra parentheses:
894
895 open HANDLE, "filename"
896 or die "Can't open: \$!\n";
897
5ba421f6 898See also discussion of list operators in L<Terms and List Operators (Leftward)>.
a0d0e21e
LW
899
d74e8afc 901X<operator, logical, not> X<not>
a0d0e21e
LW
902
903Unary "not" returns the logical negation of the expression to its right.
904It's the equivalent of "!" except for the very low precedence.
905
d74e8afc 907X<operator, logical, and> X<and>
a0d0e21e
LW
908
909Binary "and" returns the logical conjunction of the two surrounding
910expressions. It's equivalent to && except for the very low
5f05dabc 911precedence. This means that it short-circuits: i.e., the right
a0d0e21e
LW
912expression is evaluated only if the left expression is true.
913
c963b151 914=head2 Logical or, Defined or, and Exclusive Or
f23102e2 915X<operator, logical, or> X<operator, logical, xor>
d74e8afc 916X<operator, logical, defined or> X<operator, logical, exclusive or>
f23102e2 917X<or> X<xor>
a0d0e21e
LW
918
919Binary "or" returns the logical disjunction of the two surrounding
5a964f20
TC
920expressions. It's equivalent to || except for the very low precedence.
921This makes it useful for control flow
922
923 print FH \$data or die "Can't write to FH: \$!";
924
925This means that it short-circuits: i.e., the right expression is evaluated
926only if the left expression is false. Due to its precedence, you should
927probably avoid using this for assignment, only for control flow.
928
929 \$a = \$b or \$c; # bug: this is wrong
930 (\$a = \$b) or \$c; # really means this
931 \$a = \$b || \$c; # better written this way
932
19799a22 933However, when it's a list-context assignment and you're trying to use
5a964f20
TC
934"||" for control flow, you probably need "or" so that the assignment
935takes higher precedence.
936
937 @info = stat(\$file) || die; # oops, scalar sense of stat!
938 @info = stat(\$file) or die; # better, now @info gets its due
939
c963b151
BD
940Then again, you could always use parentheses.
941
a0d0e21e
LW
942Binary "xor" returns the exclusive-OR of the two surrounding expressions.
943It cannot short circuit, of course.
944
945=head2 C Operators Missing From Perl
d74e8afc
ITB
946X<operator, missing from perl> X<&> X<*>
947X<typecasting> X<(TYPE)>
a0d0e21e
LW
948
949Here is what C has that Perl doesn't:
950
951=over 8
952
953=item unary &
954
955Address-of operator. (But see the "\" operator for taking a reference.)
956
957=item unary *
958
54310121 959Dereference-address operator. (Perl's prefix dereferencing
a0d0e21e
LW
960operators are typed: \$, @, %, and &.)
961
962=item (TYPE)
963
19799a22 964Type-casting operator.
a0d0e21e
LW
965
966=back
967
5f05dabc 968=head2 Quote and Quote-like Operators
89d205f2 969X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
d74e8afc
ITB
970X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
971X<escape sequence> X<escape>
972
a0d0e21e
LW
973
974While we usually think of quotes as literal values, in Perl they
975function as operators, providing various kinds of interpolating and
976pattern matching capabilities. Perl provides customary quote characters
977for these behaviors, but also provides a way for you to choose your
978quote character for any of them. In the following table, a C<{}> represents
9f10b797 979any pair of delimiters you choose.
a0d0e21e 980
TP
981 Customary Generic Meaning Interpolates
982 '' q{} Literal no
983 "" qq{} Literal yes
af9219ee 984 `` qx{} Command yes*
2c268ad5 985 qw{} Word list no
af9219ee
MG
986 // m{} Pattern match yes*
987 qr{} Pattern yes*
988 s{}{} Substitution yes*
2c268ad5 989 tr{}{} Transliteration no (but see below)
7e3b091d 990 <<EOF here-doc yes*
a0d0e21e 991
af9219ee
MG
992 * unless the delimiter is ''.
993
87275199
GS
994Non-bracketing delimiters use the same character fore and aft, but the four
995sorts of brackets (round, angle, square, curly) will all nest, which means
9f10b797 996that
87275199 997
9f10b797 998 q{foo{bar}baz}
35f2feb0 999
9f10b797 1000is the same as
87275199
GS
1001
1002 'foo{bar}baz'
1003
1004Note, however, that this does not always work for quoting Perl code:
1005
1006 \$s = q{ if(\$a eq "}") ... }; # WRONG
1007
83df6a1d
JH
1008is a syntax error. The C<Text::Balanced> module (from CPAN, and
1009starting from Perl 5.8 part of the standard distribution) is able
1010to do this properly.
87275199 1011
19799a22 1012There can be whitespace between the operator and the quoting
fb73857a 1013characters, except when C<#> is being used as the quoting character.
19799a22
GS
1014C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
1015operator C<q> followed by a comment. Its argument will be taken
1016from the next line. This allows you to write:
fb73857a 1017
1018 s {foo} # Replace foo
1019 {bar} # with bar.
1020
904501ec
MG
1021The following escape sequences are available in constructs that interpolate
1022and in transliterations.
5691ca5f 1023X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N> X<\N{}>
04341565 1024X<\o{}>
5691ca5f 1025
2c4c1ff2
KW
1026 Sequence Note Description
1027 \t tab (HT, TAB)
1028 \n newline (NL)
1029 \r return (CR)
1030 \f form feed (FF)
1031 \b backspace (BS)
1032 \a alarm (bell) (BEL)
1033 \e escape (ESC)
1034 \x{263a} [1,8] hex char (example: SMILEY)
1035 \x1b [2,8] restricted range hex char (example: ESC)
fb121860 1036 \N{name} [3] named Unicode character or character sequence
2c4c1ff2
KW
1037 \N{U+263D} [4,8] Unicode character (example: FIRST QUARTER MOON)
1038 \c[ [5] control char (example: chr(27))
1039 \o{23072} [6,8] octal char (example: SMILEY)
1040 \033 [7,8] restricted range octal char (example: ESC)
5691ca5f
KW
1041
1042=over 4
1043
1044=item [1]
1045
2c4c1ff2
KW
1046The result is the character specified by the hexadecimal number between
1047the braces. See L</[8]> below for details on which character.
96448467
DG
1048
1049Only hexadecimal digits are valid between the braces. If an invalid
1050character is encountered, a warning will be issued and the invalid
1051character and all subsequent characters (valid or invalid) within the
1052braces will be discarded.
1053
1054If there are no valid digits between the braces, the generated character is
1055the NULL character (C<\x{00}>). However, an explicit empty brace (C<\x{}>)
1056will not cause a warning.
40687185
KW
1057
1058=item [2]
1059
2c4c1ff2
KW
1060The result is the character specified by the hexadecimal number in the range
10610x00 to 0xFF. See L</[8]> below for details on which character.
96448467
DG
1062
1063Only hexadecimal digits are valid following C<\x>. When C<\x> is followed
2c4c1ff2 1064by fewer than two valid digits, any valid digits will be zero-padded. This
96448467 1065means that C<\x7> will be interpreted as C<\x07> and C<\x> alone will be
2c4c1ff2 1066interpreted as C<\x00>. Except at the end of a string, having fewer than
96448467
DG
1067two valid digits will result in a warning. Note that while the warning
1068says the illegal character is ignored, it is only ignored as part of the
1069escape and will still be used as the subsequent character in the string.
1070For example:
1071
1072 Original Result Warns?
1073 "\x7" "\x07" no
1074 "\x" "\x00" no
1075 "\x7q" "\x07q" yes
1076 "\xq" "\x00q" yes
1077
40687185
KW
1078=item [3]
1079
fb121860 1080The result is the Unicode character or character sequence given by I<name>.
2c4c1ff2 1081See L<charnames>.
40687185
KW
1082
1083=item [4]
1084
2c4c1ff2
KW
1085C<\N{U+I<hexadecimal number>}> means the Unicode character whose Unicode code
1086point is I<hexadecimal number>.
40687185
KW
1087
1088=item [5]
1089
5691ca5f
KW
1090The character following C<\c> is mapped to some other character as shown in the
1091table:
1092
1093 Sequence Value
1094 \c@ chr(0)
1095 \cA chr(1)
1096 \ca chr(1)
1097 \cB chr(2)
1098 \cb chr(2)
1099 ...
1100 \cZ chr(26)
1101 \cz chr(26)
1102 \c[ chr(27)
1103 \c] chr(29)
1104 \c^ chr(30)
1105 \c? chr(127)
1106
1107Also, C<\c\I<X>> yields C< chr(28) . "I<X>"> for any I<X>, but cannot come at the
1108end of a string, because the backslash would be parsed as escaping the end
1109quote.
1110
1111On ASCII platforms, the resulting characters from the list above are the
1112complete set of ASCII controls. This isn't the case on EBCDIC platforms; see
1113L<perlebcdic/OPERATOR DIFFERENCES> for the complete list of what these
1114sequences mean on both ASCII and EBCDIC platforms.
1115
1116Use of any other character following the "c" besides those listed above is
17a3df4c
KW
1117discouraged, and some are deprecated with the intention of removing
1118those in Perl 5.16. What happens for any of these
6ae9f32a
KW
1119other characters currently though, is that the value is derived by inverting
1120the 7th bit (0x40).
5691ca5f
KW
1121
1122To get platform independent controls, you can use C<\N{...}>.
1123
40687185
KW
1124=item [6]
1125
2c4c1ff2
KW
1126The result is the character specified by the octal number between the braces.
1127See L</[8]> below for details on which character.
04341565
DG
1128
1129If a character that isn't an octal digit is encountered, a warning is raised,
1130and the value is based on the octal digits before it, discarding it and all
1131following characters up to the closing brace. It is a fatal error if there are
1132no octal digits at all.
1133
1134=item [7]
1135
2c4c1ff2
KW
1136The result is the character specified by the three digit octal number in the
1137range 000 to 777 (but best to not use above 077, see next paragraph). See
1138L</[8]> below for details on which character.
1139
1140Some contexts allow 2 or even 1 digit, but any usage without exactly
40687185 1141three digits, the first being a zero, may give unintended results. (For
04341565
DG
1142example, see L<perlrebackslash/Octal escapes>.) Starting in Perl 5.14, you may
1143use C<\o{}> instead which avoids all these problems. Otherwise, it is best to
1144use this construct only for ordinals C<\077> and below, remembering to pad to
1145the left with zeros to make three digits. For larger ordinals, either use
d90d5a38 1146C<\o{}> , or convert to something else, such as to hex and use C<\x{}>
40687185 1148
40687185
KW
1149Having fewer than 3 digits may lead to a misleading warning message that says
1150that what follows is ignored. For example, C<"\128"> in the ASCII character set
1151is equivalent to the two characters C<"\n8">, but the warning C<Illegal octal
1152digit '8' ignored> will be thrown. To avoid this warning, make sure to pad
f6050459 1153your octal number with C<0>'s: C<"\0128">.
5691ca5f 1154
2c4c1ff2
KW
1155=item [8]
1156
1157Several of the constructs above specify a character by a number. That number
1158gives the character's position in the character set encoding (indexed from 0).
1159This is called synonymously its ordinal, code position, or code point). Perl
1160works on platforms that have a native encoding currently of either ASCII/Latin1
1161or EBCDIC, each of which allow specification of 256 characters. In general, if
1162the number is 255 (0xFF, 0377) or below, Perl interprets this in the platform's
1163native encoding. If the number is 256 (0x100, 0400) or above, Perl interprets
1164it as as a Unicode code point and the result is the corresponding Unicode
1165character. For example C<\x{50}> and C<\o{120}> both are the number 80 in
1166decimal, which is less than 256, so the number is interpreted in the native
1167character set encoding. In ASCII the character in the 80th position (indexed
1168from 0) is the letter "P", and in EBCDIC it is the ampersand symbol "&".
1169C<\x{100}> and C<\o{400}> are both 256 in decimal, so the number is interpreted
1170as a Unicode code point no matter what the native encoding is. The name of the
1171character in the 100th position (indexed by 0) in Unicode is
1172C<LATIN CAPITAL LETTER A WITH MACRON>.
1173
1174There are a couple of exceptions to the above rule. C<\N{U+I<hex number>}> is
1175always interpreted as a Unicode code point, so that C<\N{U+0050}> is "P" even
1176on EBCDIC platforms. And if L<C<S<use encoding>>|encoding> is in effect, the
1177number is considered to be in that encoding, and is translated from that into
1178the platform's native encoding if there is a corresponding native character;
1179otherwise to Unicode.
1180
5691ca5f 1181=back
4c77eaa2 1182
e526e8bb
KW
1183B<NOTE>: Unlike C and other languages, Perl has no C<\v> escape sequence for
1184the vertical tab (VT - ASCII 11), but you may use C<\ck> or C<\x0b>. (C<\v>
1185does have meaning in regular expression patterns in Perl, see L<perlre>.)
1186
1187The following escape sequences are available in constructs that interpolate,
904501ec 1188but not in transliterations.
d74e8afc 1189X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q>
904501ec 1190
a0d0e21e
LW
1191 \l lowercase next char
1192 \u uppercase next char
1193 \L lowercase till \E
1194 \U uppercase till \E
1d2dff63 1195 \Q quote non-word characters till \E
7e31b643 1196 \E end either case modification or quoted section
a0d0e21e 1197
95cc3e0c
JH
1198If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
1199C<\u> and C<\U> is taken from the current locale. See L<perllocale>.
b6538e4f 1200If Unicode (for example, C<\N{}> or code points of 0x100 or
95cc3e0c 1201beyond) is being used, the case map used by C<\l>, C<\L>, C<\u> and
e526e8bb 1202C<\U> is as defined by Unicode.
a034a98d 1203
5a964f20
TC
1204All systems use the virtual C<"\n"> to represent a line terminator,
1205called a "newline". There is no such thing as an unvarying, physical
19799a22 1206newline character. It is only an illusion that the operating system,
5a964f20
TC
1207device drivers, C libraries, and Perl all conspire to preserve. Not all
1208systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
1209on a Mac, these are reversed, and on systems without line terminator,
1210printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
1211you mean a "newline" for your system, but use the literal ASCII when you
1212need an exact character. For example, most networking protocols expect
2a380090 1213and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
1214and although they often accept just C<"\012">, they seldom tolerate just
1215C<"\015">. If you get in the habit of using C<"\n"> for networking,
1216you may be burned some day.
d74e8afc
ITB
1217X<newline> X<line terminator> X<eol> X<end of line>
1218X<\n> X<\r> X<\r\n>
5a964f20 1219
904501ec
MG
1220For constructs that do interpolate, variables beginning with "C<\$>"
1221or "C<@>" are interpolated. Subscripted variables such as C<\$a[3]> or
A
1222C<< \$href->{key}[0] >> are also interpolated, as are array and hash slices.
1223But method calls such as C<< \$obj->meth >> are not.
af9219ee
MG
1224
1225Interpolating an array or slice interpolates the elements in order,
1226separated by the value of C<\$">, so is equivalent to interpolating
6deea57f
TS
1227C<join \$", @array>. "Punctuation" arrays such as C<@*> are only
1228interpolated if the name is enclosed in braces C<@{*}>, but special
1229arrays C<@_>, C<@+>, and C<@-> are interpolated, even without braces.
af9219ee 1230
bc7b91c6
EB
1231For double-quoted strings, the quoting from C<\Q> is applied after
1232interpolation and escapes are processed.
1233
1234 "abc\Qfoo\tbar\$s\Exyz"
1235
1236is equivalent to
1237
1238 "abc" . quotemeta("foo\tbar\$s") . "xyz"
1239
1240For the pattern of regex operators (C<qr//>, C<m//> and C<s///>),
1241the quoting from C<\Q> is applied after interpolation is processed,
1242but before escapes are processed. This allows the pattern to match
1243literally (except for C<\$> and C<@>). For example, the following matches:
1244
1245 '\s\t' =~ /\Q\s\t/
1246
1247Because C<\$> or C<@> trigger interpolation, you'll need to use something
1248like C</\Quser\E\@\Qhost/> to match them literally.
1d2dff63 1249
a0d0e21e
LW
1250Patterns are subject to an additional level of interpretation as a
1251regular expression. This is done as a second pass, after variables are
1252interpolated, so that regular expressions may be incorporated into the
1253pattern from the variables. If this is not what you want, use C<\Q> to
1254interpolate a variable literally.
1255
19799a22
GS
1256Apart from the behavior described above, Perl does not expand
1257multiple levels of interpolation. In particular, contrary to the
1258expectations of shell programmers, back-quotes do I<NOT> interpolate
1259within double quotes, nor do single quotes impede evaluation of
1260variables when used within double quotes.
a0d0e21e 1261
5f05dabc 1262=head2 Regexp Quote-Like Operators
d74e8afc 1263X<operator, regexp>
cb1a09d0 1264
5f05dabc 1265Here are the quote-like operators that apply to pattern
cb1a09d0
1266matching and related activities.
1267
a0d0e21e
LW
1268=over 8
1269
b6fa137b 1270=item qr/STRING/msixpodual
01c6f5f4 1271X<qr> X</i> X</m> X</o> X</s> X</x> X</p>
a0d0e21e 1272
87e95b7f
YO
1273This operator quotes (and possibly compiles) its I<STRING> as a regular
1274expression. I<STRING> is interpolated the same way as I<PATTERN>
1275in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1276is done. Returns a Perl value which may be used instead of the
f6050459 1277corresponding C</STRING/msixpodual> expression. The returned value is a
85dd5c8b 1278normalized version of the original pattern. It magically differs from
64c5a566 1279a string containing the same characters: C<ref(qr/x/)> returns "Regexp",
85dd5c8b 1280even though dereferencing the result returns undef.
a0d0e21e 1281
87e95b7f
YO
1282For example,
1283
1284 \$rex = qr/my.STRING/is;
85dd5c8b 1285 print \$rex; # prints (?si-xm:my.STRING)
87e95b7f
YO
1286 s/\$rex/foo/;
1287
1288is equivalent to
1289
1290 s/my.STRING/foo/is;
1291
1292The result may be used as a subpattern in a match:
1293
1294 \$re = qr/\$pattern/;
1295 \$string =~ /foo\${re}bar/; # can be interpolated in other patterns
1296 \$string =~ \$re; # or used standalone
1297 \$string =~ /\$re/; # or this way
1298
f6050459 1299Since Perl may compile the pattern at the moment of execution of the qr()
87e95b7f
YO
1300operator, using qr() may have speed advantages in some situations,
1301notably if the result of qr() is used standalone:
1302
1303 sub match {
1304 my \$patterns = shift;
1305 my @compiled = map qr/\$_/i, @\$patterns;
1306 grep {
1307 my \$success = 0;
1308 foreach my \$pat (@compiled) {
1309 \$success = 1, last if /\$pat/;
1310 }
1311 \$success;
1312 } @_;
5a964f20
TC
1313 }
1314
87e95b7f
YO
1315Precompilation of the pattern into an internal representation at
1316the moment of qr() avoids a need to recompile the pattern every
1317time a match C</\$pat/> is attempted. (Perl has many other internal
1318optimizations, but none would be triggered in the above example if
1319we did not use qr() operator.)
1320
765fa144 1321Options (specified by the following modifiers) are:
87e95b7f
YO
1322
1323 m Treat string as multiple lines.
1324 s Treat string as single line. (Make . match a newline)
1325 i Do case-insensitive pattern matching.
1326 x Use extended regular expressions.
1327 p When matching preserve a copy of the matched string so
1328 that \${^PREMATCH}, \${^MATCH}, \${^POSTMATCH} will be defined.
1329 o Compile pattern only once.
b6fa137b 1330 l Use the locale
5e2aa8f5
KW
1331 u Use Unicode rules
1332 a Use ASCII for \d, \s, \w; specifying two a's further restricts
1333 /i matching so that no ASCII character will match a non-ASCII
1334 one
b6fa137b 1335 d Use Unicode or native charset, as in 5.12 and earlier
87e95b7f
YO
1336
1337If a precompiled pattern is embedded in a larger pattern then the effect
f6050459 1338of 'msixpluad' will be propagated appropriately. The effect the 'o'
87e95b7f
YO
1339modifier has is not propagated, being restricted to those patterns
1340explicitly using it.
1341
b6fa137b 1342The last four modifiers listed above, added in Perl 5.14,
5e2aa8f5 1343control the character set semantics.
da392a17 1344
87e95b7f 1345See L<perlre> for additional information on valid syntax for STRING, and
5e2aa8f5 1346for a detailed look at the semantics of regular expressions. In
f6050459
KW
1347particular, all the modifiers execpt C</o> are further explained in
1348L<perlre/Modifiers>. C</o> is described in the next section.
a0d0e21e 1349
b6fa137b 1350=item m/PATTERN/msixpodualgc
89d205f2
YO
1351X<m> X<operator, match>
1352X<regexp, options> X<regexp> X<regex, options> X<regex>
01c6f5f4 1353X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c>
a0d0e21e 1354
b6fa137b 1355=item /PATTERN/msixpodualgc
a0d0e21e 1356
5a964f20 1357Searches a string for a pattern match, and in scalar context returns
19799a22
GS
1358true if it succeeds, false if it fails. If no string is specified
1359via the C<=~> or C<!~> operator, the \$_ string is searched. (The
1360string specified with C<=~> need not be an lvalue--it may be the
1361result of an expression evaluation, but remember the C<=~> binds
a0d0e21e 1363
f6050459 1364Options are as described in C<qr//> above; in addition, the following match
01c6f5f4 1365process modifiers are available:
a0d0e21e 1366
950b09ed
KW
1367 g Match globally, i.e., find all occurrences.
1368 c Do not reset search position on a failed match when /g is in effect.
a0d0e21e 1369
725a61d7
Z
1370If "/" is the delimiter then the initial C<m> is optional. With the C<m>
1371you can use any pair of non-whitespace characters
1372as delimiters. This is particularly useful for matching path names
1373that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
1374the delimiter, then a match-only-once rule applies,
1375described in C<m?PATTERN?> below.
19799a22 1376If "'" is the delimiter, no interpolation is performed on the PATTERN.
ed02a3bf
DN
1377When using a character valid in an identifier, whitespace is required
1378after the C<m>.
a0d0e21e 1379
532c9e80
KW
1380PATTERN may contain variables, which will be interpolated
1381every time the pattern search is evaluated, except
1f247705
GS
1382for when the delimiter is a single quote. (Note that C<\$(>, C<\$)>, and
1383C<\$|> are not interpolated because they look like end-of-string tests.)
532c9e80
KW
1384Perl will not recompile the pattern unless an interpolated
1385variable that it contains changes. You can force Perl to skip the
1386test and never recompile by adding a C</o> (which stands for "once")
1387after the trailing delimiter.
1388Once upon a time, Perl would recompile regular expressions
1389unnecessarily, and this modifier was useful to tell it not to do so, in the
1390interests of speed. But now, the only reasons to use C</o> are either:
1391
1392=over
1393
1394=item 1
1395
1396The variables are thousands of characters long and you know that they
1397don't change, and you need to wring out the last little bit of speed by
1398having Perl skip testing for that. (There is a maintenance penalty for
1399doing this, as mentioning C</o> constitutes a promise that you won't
1400change the variables in the pattern. If you change them, Perl won't
1401even notice.)
1402
1403=item 2
1404
1405you want the pattern to use the initial values of the variables
1406regardless of whether they change or not. (But there are saner ways
1407of accomplishing this than using C</o>.)
1408
1409=back
a0d0e21e 1410
e9d89077
DN
1411=item The empty pattern //
1412
5a964f20 1413If the PATTERN evaluates to the empty string, the last
d65afb4b
HS
1414I<successfully> matched regular expression is used instead. In this
1415case, only the C<g> and C<c> flags on the empty pattern is honoured -
1416the other flags are taken from the original pattern. If no match has
1417previously succeeded, this will (silently) act instead as a genuine
1418empty pattern (which will always match).
a0d0e21e 1419
89d205f2
YO
1420Note that it's possible to confuse Perl into thinking C<//> (the empty
1421regex) is really C<//> (the defined-or operator). Perl is usually pretty
1422good about this, but some pathological cases might trigger this, such as
1423C<\$a///> (is that C<(\$a) / (//)> or C<\$a // />?) and C<print \$fh //>
1424(C<print \$fh(//> or C<print(\$fh //>?). In all of these examples, Perl
1425will assume you meant defined-or. If you meant the empty regex, just
1426use parentheses or spaces to disambiguate, or even prefix the empty
c963b151
BD
1427regex with an C<m> (so C<//> becomes C<m//>).
1428
e9d89077
DN
1429=item Matching in list context
1430
19799a22 1431If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1432list consisting of the subexpressions matched by the parentheses in the
f7e33566
GS
1433pattern, i.e., (C<\$1>, C<\$2>, C<\$3>...). (Note that here C<\$1> etc. are
1434also set, and that this differs from Perl 4's behavior.) When there are
1435no parentheses in the pattern, the return value is the list C<(1)> for
1436success. With or without parentheses, an empty list is returned upon
1437failure.
a0d0e21e
LW
1438
1439Examples:
1440
1441 open(TTY, '/dev/tty');
1442 <TTY> =~ /^y/i && foo(); # do foo if desired
1443
1444 if (/Version: *([0-9.]*)/) { \$version = \$1; }
1445
1446 next if m#^/usr/spool/uucp#;
1447
1448 # poor man's grep
1449 \$arg = shift;
1450 while (<>) {
1451 print if /\$arg/o; # compile only once
1452 }
1453
1454 if ((\$F1, \$F2, \$Etc) = (\$foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
1455
1456This last example splits \$foo into the first two words and the
5f05dabc 1457remainder of the line, and assigns those three fields to \$F1, \$F2, and
1458\$Etc. The conditional is true if any variables were assigned, i.e., if
a0d0e21e
LW
1459the pattern matched.
1460
19799a22 1461The C</g> modifier specifies global pattern matching--that is,
3dd93342 1462matching as many times as possible within the string. How it behaves
1463depends on the context. In list context, it returns a list of the
19799a22 1464substrings matched by any capturing parentheses in the regular
3dd93342 1465expression. If there are no parentheses, it returns a list of all
19799a22
GS
1466the matched strings, as if there were parentheses around the whole
1467pattern.
a0d0e21e 1468
7e86de3e 1469In scalar context, each execution of C<m//g> finds the next match,
19799a22 1470returning true if it matches, and false if there is no further match.
3dd93342 1471The position after the last match can be read or set using the C<pos()>
1472function; see L<perlfunc/pos>. A failed match normally resets the
7e86de3e 1473search position to the beginning of the string, but you can avoid that
3dd93342 1474by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
7e86de3e 1475string also resets the search position.
c90c0ff4 1476
e9d89077
DN
1477=item \G assertion
1478
c90c0ff4 1479You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
3dd93342 1480zero-width assertion that matches the exact position where the
1481previous C<m//g>, if any, left off. Without the C</g> modifier, the
1482C<\G> assertion still anchors at C<pos()> as it was at the start of
1483the operation (see L<perlfunc/pos>), but the match is of course only
1484attempted once. Using C<\G> without C</g> on a target string that has
1485not previously had a C</g> match applied to it is the same as using
1486the C<\A> assertion to match the beginning of the string. Note also
1487that, currently, C<\G> is only properly supported when anchored at the
1488very beginning of the pattern.
c90c0ff4 1489
1490Examples:
a0d0e21e
LW
1491
1492 # list context
1493 (\$one,\$five,\$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1494
1495 # scalar context
5d43e42d 1496 \$/ = "";
19799a22
GS
1497 while (defined(\$paragraph = <>)) {
1498 while (\$paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
1499 \$sentences++;
a0d0e21e
LW
1500 }
1501 }
1502 print "\$sentences\n";
1503
c90c0ff4 1504 # using m//gc with \G
137443ea 1505 \$_ = "ppooqppqq";
44a8e56a 1506 while (\$i++ < 2) {
1507 print "1: '";
c90c0ff4 1508 print \$1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1509 print "2: '";
c90c0ff4 1510 print \$1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1511 print "3: '";
c90c0ff4 1512 print \$1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1513 }
5d43e42d 1514 print "Final: '\$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a 1515
1516The last example should print:
1517
1518 1: 'oo', pos=4
137443ea 1519 2: 'q', pos=5
44a8e56a 1520 3: 'pp', pos=7
1521 1: '', pos=7
137443ea 1522 2: 'q', pos=8
1523 3: '', pos=8
5d43e42d
DC
1524 Final: 'q', pos=8
1525
1526Notice that the final match matched C<q> instead of C<p>, which a match
1527without the C<\G> anchor would have done. Also note that the final match
ac036724 1528did not update C<pos>. C<pos> is only updated on a C</g> match. If the
5d43e42d
DC
1529final match did indeed match C<p>, it's a good bet that you're running an
1530older (pre-5.6.0) Perl.
44a8e56a 1531
c90c0ff4 1532A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1533combine several regexps like this to process a string part-by-part,
c90c0ff4 1534doing different actions depending on which regexp matched. Each
1535regexp tries to match where the previous one leaves off.
e7ea3e70 1536
3fe9a6f1 1537 \$_ = <<'EOL';
950b09ed 1538 \$url = URI::URL->new( "http://example.com/" ); die if \$url eq "xXx";
3fe9a6f1 1539 EOL
1540 LOOP:
e7ea3e70 1541 {
950b09ed
KW
1542 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
1543 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
1544 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
1545 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
1546 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
1547 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
1548 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
1549 print ". That's all!\n";
e7ea3e70
IZ
1550 }
1551
1552Here is the output (split into several lines):
1553
51622cce 1554 line-noise lowercase line-noise UPPERCASE line-noise UPPERCASE
1555 line-noise lowercase line-noise lowercase line-noise lowercase
1556 lowercase line-noise lowercase lowercase line-noise lowercase
1557 lowercase line-noise MiXeD line-noise. That's all!
44a8e56a 1558
ceb131e8 1559=item m?PATTERN?
725a61d7 1560X<?> X<operator, match-once>
87e95b7f 1561
55d389e7
DG
1562=item ?PATTERN?
1563
725a61d7
Z
1564This is just like the C<m/PATTERN/> search, except that it matches
1565only once between calls to the reset() operator. This is a useful
87e95b7f 1566optimization when you want to see only the first occurrence of
ceb131e8 1567something in each file of a set of files, for instance. Only C<m??>
87e95b7f
YO
1568patterns local to the current package are reset.
1569
1570 while (<>) {
ceb131e8 1571 if (m?^\$?) {
87e95b7f
YO
1572 # blank line between header and body
1573 }
1574 } continue {
725a61d7 1575 reset if eof; # clear m?? status for next file
87e95b7f
YO
1576 }
1577
725a61d7
Z
1578The match-once behaviour is controlled by the match delimiter being
1579C<?>; with any other delimiter this is the normal C<m//> operator.
1580
1581For historical reasons, the leading C<m> in C<m?PATTERN?> is optional,
1582but the resulting C<?PATTERN?> syntax is deprecated, will warn on
1583usage and may be removed from a future stable release of Perl without
55d389e7 1584further notice.
87e95b7f 1585
b6fa137b 1586=item s/PATTERN/REPLACEMENT/msixpodualgcer
87e95b7f 1587X<substitute> X<substitution> X<replace> X<regexp, replace>
4f4d7508 1588X<regexp, substitute> X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c> X</e> X</r>
87e95b7f
YO
1589
1590Searches a string for a pattern, and if found, replaces that pattern
1591with the replacement text and returns the number of substitutions
1592made. Otherwise it returns false (specifically, the empty string).
1593
4f4d7508 1594If the C</r> (non-destructive) option is used then it will perform the
679563bb
KW
1595substitution on a copy of the string and instead of returning the
1596number of substitutions, it returns the copy whether or not a
4f4d7508 1597substitution occurred. The original string will always remain unchanged in
c8dbf8cd 1598this case. The copy will always be a plain string, even if the input is an
4f4d7508
DC
1599object or a tied variable.
1600
87e95b7f
YO
1601If no string is specified via the C<=~> or C<!~> operator, the C<\$_>
1602variable is searched and modified. (The string specified with C<=~> must
1603be scalar variable, an array element, a hash element, or an assignment
1604to one of those, i.e., an lvalue.)
1605
1606If the delimiter chosen is a single quote, no interpolation is
1607done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1608PATTERN contains a \$ that looks like a variable rather than an
1609end-of-string test, the variable will be interpolated into the pattern
1610at run-time. If you want the pattern compiled only once the first time
1611the variable is interpolated, use the C</o> option. If the pattern
1612evaluates to the empty string, the last successfully executed regular
1613expression is used instead. See L<perlre> for further explanation on these.
87e95b7f
YO
1614
1615Options are as with m// with the addition of the following replacement
1616specific options:
1617
1618 e Evaluate the right side as an expression.
4f4d7508
DC
1619 ee Evaluate the right side as a string then eval the result.
1620 r Return substitution and leave the original string untouched.
87e95b7f 1621
ed02a3bf
DN
1622Any non-whitespace delimiter may replace the slashes. Add space after
1623the C<s> when using a character allowed in identifiers. If single quotes
1624are used, no interpretation is done on the replacement string (the C</e>
1625modifier overrides this, however). Unlike Perl 4, Perl 5 treats backticks
1626as normal delimiters; the replacement text is not evaluated as a command.
1627If the PATTERN is delimited by bracketing quotes, the REPLACEMENT has
1628its own pair of quotes, which may or may not be bracketing quotes, e.g.,
87e95b7f
YO
1629C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
1630replacement portion to be treated as a full-fledged Perl expression
1631and evaluated right then and there. It is, however, syntax checked at
1632compile-time. A second C<e> modifier will cause the replacement portion
1633to be C<eval>ed before being run as a Perl expression.
1634
1635Examples:
1636
1637 s/\bgreen\b/mauve/g; # don't change wintergreen
1638
1639 \$path =~ s|/usr/bin|/usr/local/bin|;
1640
1641 s/Login: \$foo/Login: \$bar/; # run-time pattern
1642
1643 (\$foo = \$bar) =~ s/this/that/; # copy first, then change
4f4d7508
DC
1644 (\$foo = "\$bar") =~ s/this/that/; # convert to string, copy, then change
1645 \$foo = \$bar =~ s/this/that/r; # Same as above using /r
1646 \$foo = \$bar =~ s/this/that/r
1647 =~ s/that/the other/r; # Chained substitutes using /r
1648 @foo = map { s/this/that/r } @bar # /r is very useful in maps
87e95b7f
YO
1649
1650 \$count = (\$paragraph =~ s/Mister\b/Mr./g); # get change-count
1651
1652 \$_ = 'abc123xyz';
1653 s/\d+/\$&*2/e; # yields 'abc246xyz'
1654 s/\d+/sprintf("%5d",\$&)/e; # yields 'abc 246xyz'
1655 s/\w/\$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1656
1657 s/%(.)/\$percent{\$1}/g; # change percent escapes; no /e
1658 s/%(.)/\$percent{\$1} || \$&/ge; # expr now, so /e
1659 s/^=(\w+)/pod(\$1)/ge; # use function call
1660
4f4d7508
DC
1661 \$_ = 'abc123xyz';
1662 \$a = s/abc/def/r; # \$a is 'def123xyz' and
1663 # \$_ remains 'abc123xyz'.
1664
87e95b7f
YO
1665 # expand variables in \$_, but dynamics only, using
1666 # symbolic dereferencing
1667 s/\\$(\w+)/\${\$1}/g;
1668
1669 # Add one to the value of any numbers in the string
1670 s/(\d+)/1 + \$1/eg;
1671
1672 # This will expand any embedded scalar variable
1673 # (including lexicals) in \$_ : First \$1 is interpolated
1674 # to the variable name, and then evaluated
1675 s/(\\$\w+)/\$1/eeg;
1676
1677 # Delete (most) C comments.
1678 \$program =~ s {
1679 /\* # Match the opening delimiter.
1680 .*? # Match a minimal number of characters.
1681 \*/ # Match the closing delimiter.
1682 } []gsx;
1683
1684 s/^\s*(.*?)\s*\$/\$1/; # trim whitespace in \$_, expensively
1685
1686 for (\$variable) { # trim whitespace in \$variable, cheap
1687 s/^\s+//;
1688 s/\s+\$//;
1689 }
1690
1691 s/([^ ]*) *([^ ]*)/\$2 \$1/; # reverse 1st two fields
1692
1693Note the use of \$ instead of \ in the last example. Unlike
1694B<sed>, we use the \<I<digit>> form in only the left hand side.
1695Anywhere else it's \$<I<digit>>.
1696
1697Occasionally, you can't use just a C</g> to get all the changes
1698to occur that you might want. Here are two common cases:
1699
1700 # put commas in the right places in an integer
1701 1 while s/(\d)(\d\d\d)(?!\d)/\$1,\$2/g;
1702
1703 # expand tabs to 8-column spacing
1704 1 while s/\t+/' ' x (length(\$&)*8 - length(\$`)%8)/e;
1705
1c424184
FC
1706C<s///le> is treated as a substitution followed by the C<le> operator, not
1707the C</le> flags. This may change in a future version of Perl. It
1708produces a warning if warnings are enabled. To disambiguate, use a space
1709or change the order of the flags:
1710
1711 s/foo/bar/ le 5; # "le" infix operator
1712 s/foo/bar/el; # "e" and "l" flags
1713
87e95b7f
YO
1714=back
1715
1717X<operator, quote-like>
1718
01c6f5f4
RGS
1719=over 4
1720
a0d0e21e 1721=item q/STRING/
5d44bfff 1722X<q> X<quote, single> X<'> X<''>
a0d0e21e 1723
5d44bfff 1724=item 'STRING'
a0d0e21e 1725
19799a22 1726A single-quoted, literal string. A backslash represents a backslash
68dc0745 1727unless followed by the delimiter or another backslash, in which case
1728the delimiter or backslash is interpolated.
a0d0e21e
LW
1729
1730 \$foo = q!I said, "You said, 'She said it.'"!;
1731 \$bar = q('This is it.');
68dc0745 1732 \$baz = '\n'; # a two-character string
a0d0e21e
LW
1733
1734=item qq/STRING/
d74e8afc 1735X<qq> X<quote, double> X<"> X<"">
a0d0e21e
LW
1736
1737=item "STRING"
1738
1739A double-quoted, interpolated string.
1740
1741 \$_ .= qq
1742 (*** The previous line contains the naughty word "\$1".\n)
19799a22 1743 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 1744 \$baz = "\n"; # a one-character string
a0d0e21e
LW
1745
1746=item qx/STRING/
d74e8afc 1747X<qx> X<`> X<``> X<backtick>
a0d0e21e
LW
1748
1749=item `STRING`
1750
43dd4d21
JH
1751A string which is (possibly) interpolated and then executed as a
1752system command with C</bin/sh> or its equivalent. Shell wildcards,
1753pipes, and redirections will be honored. The collected standard
1754output of the command is returned; standard error is unaffected. In
1755scalar context, it comes back as a single (potentially multi-line)
1756string, or undef if the command failed. In list context, returns a
1757list of lines (however you've defined lines with \$/ or
1758\$INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
5a964f20
TC
1759
1760Because backticks do not affect standard error, use shell file descriptor
1761syntax (assuming the shell supports this) if you care to address this.
1762To capture a command's STDERR and STDOUT together:
a0d0e21e 1763
5a964f20
TC
1764 \$output = `cmd 2>&1`;
1765
1766To capture a command's STDOUT but discard its STDERR:
1767
1768 \$output = `cmd 2>/dev/null`;
1769
1770To capture a command's STDERR but discard its STDOUT (ordering is
1771important here):
1772
1773 \$output = `cmd 2>&1 1>/dev/null`;
1774
1775To exchange a command's STDOUT and STDERR in order to capture the STDERR
1776but leave its STDOUT to come out the old STDERR:
1777
1778 \$output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1779
1780To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
1781to redirect them separately to files, and then read from those files
1782when the program is done:
5a964f20 1783
2359510d 1784 system("program args 1>program.stdout 2>program.stderr");
5a964f20 1785
30398227
SP
1786The STDIN filehandle used by the command is inherited from Perl's STDIN.
1787For example:
1788
1789 open BLAM, "blam" || die "Can't open: \$!";
1790 open STDIN, "<&BLAM";
1791 print `sort`;
1792
1793will print the sorted contents of the file "blam".
1794
5a964f20
TC
1795Using single-quote as a delimiter protects the command from Perl's
1796double-quote interpolation, passing it on to the shell instead:
1797
1798 \$perl_info = qx(ps \$\$); # that's Perl's \$\$
1799 \$shell_info = qx'ps \$\$'; # that's the new shell's \$\$
1800
19799a22 1801How that string gets evaluated is entirely subject to the command
5a964f20
TC
1802interpreter on your system. On most platforms, you will have to protect
1803shell metacharacters if you want them treated literally. This is in
1804practice difficult to do, as it's unclear how to escape which characters.
1805See L<perlsec> for a clean and safe example of a manual fork() and exec()
1806to emulate backticks safely.
a0d0e21e 1807
bb32b41a
GS
1808On some platforms (notably DOS-like ones), the shell may not be
1809capable of dealing with multiline commands, so putting newlines in
1810the string may not get you what you want. You may be able to evaluate
1811multiple commands in a single line by separating them with the command
1812separator character, if your shell supports that (e.g. C<;> on many Unix
1813shells; C<&> on the Windows NT C<cmd> shell).
1814
0f897271
GS
1815Beginning with v5.6.0, Perl will attempt to flush all files opened for
1816output before starting the child process, but this may not be supported
1817on some platforms (see L<perlport>). To be safe, you may need to set
1818C<\$|> (\$AUTOFLUSH in English) or call the C<autoflush()> method of
1819C<IO::Handle> on any open handles.
1820
bb32b41a
GS
1821Beware that some command shells may place restrictions on the length
1822of the command line. You must ensure your strings don't exceed this
1823limit after any necessary interpolations. See the platform-specific
1824release notes for more details about your particular environment.
1825
5a964f20
TC
1826Using this operator can lead to programs that are difficult to port,
1827because the shell commands called vary between systems, and may in
1828fact not be present at all. As one example, the C<type> command under
1829the POSIX shell is very different from the C<type> command under DOS.
1830That doesn't mean you should go out of your way to avoid backticks
1831when they're the right way to get something done. Perl was made to be
1832a glue language, and one of the things it glues together is commands.
1833Just understand what you're getting yourself into.
bb32b41a 1834
da87341d 1835See L</"I/O Operators"> for more discussion.
a0d0e21e 1836
945c54fd 1837=item qw/STRING/
d74e8afc 1838X<qw> X<quote, list> X<quote, words>
945c54fd
JH
1839
1840Evaluates to a list of the words extracted out of STRING, using embedded
1841whitespace as the word delimiters. It can be understood as being roughly
1842equivalent to:
1843
1844 split(' ', q/STRING/);
1845
efb1e162
CW
1846the differences being that it generates a real list at compile time, and
1847in scalar context it returns the last element in the list. So
945c54fd
JH
1848this expression:
1849
1850 qw(foo bar baz)
1851
1852is semantically equivalent to the list:
1853
1854 'foo', 'bar', 'baz'
1855
1856Some frequently seen examples:
1857
1858 use POSIX qw( setlocale localeconv )
1859 @EXPORT = qw( foo bar baz );
1860
1861A common mistake is to try to separate the words with comma or to
1862put comments into a multi-line C<qw>-string. For this reason, the
89d205f2 1863C<use warnings> pragma and the B<-w> switch (that is, the C<\$^W> variable)
945c54fd
JH
1864produces warnings if the STRING contains the "," or the "#" character.
1865
a0d0e21e 1866
8ff32507 1867=item tr/SEARCHLIST/REPLACEMENTLIST/cdsr
d74e8afc 1868X<tr> X<y> X<transliterate> X</c> X</d> X</s>
a0d0e21e 1869
8ff32507 1870=item y/SEARCHLIST/REPLACEMENTLIST/cdsr
a0d0e21e 1871
2c268ad5 1872Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
1873with the corresponding character in the replacement list. It returns
1874the number of characters replaced or deleted. If no string is
2c268ad5 1875specified via the =~ or !~ operator, the \$_ string is transliterated. (The
54310121 1876string specified with =~ must be a scalar variable, an array element, a
1877hash element, or an assignment to one of those, i.e., an lvalue.)
8ff32507
FC
1879If the C</r> (non-destructive) option is used then it will perform the
1880replacement on a copy of the string and return the copy whether or not it
1881was modified. The original string will always remain unchanged in
1882this case. The copy will always be a plain string, even if the input is an
1883object or a tied variable.
1884
89d205f2 1885A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
2c268ad5 1886does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121 1887For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1888SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1889its own pair of quotes, which may or may not be bracketing quotes,
2c268ad5 1890e.g., C<tr[A-Z][a-z]> or C<tr(+\-*/)/ABCD/>.
a0d0e21e 1891
cc255d5f 1892Note that C<tr> does B<not> do regular expression character classes
e0c83546 1893such as C<\d> or C<[:lower:]>. The C<tr> operator is not equivalent to
cc255d5f
JH
1894the tr(1) utility. If you want to map strings between lower/upper
1895cases, see L<perlfunc/lc> and L<perlfunc/uc>, and in general consider
1896using the C<s> operator if you need regular expressions.
1897
JH
1898Note also that the whole range idea is rather unportable between
1899character sets--and even within character sets they may cause results
1900you probably didn't expect. A sound principle is to use only ranges
1901that begin from and end at either alphabets of equal case (a-e, A-E),
1902or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1903character sets in full.
1904
a0d0e21e
LW
1905Options:
1906
1907 c Complement the SEARCHLIST.
1908 d Delete found but unreplaced characters.
1909 s Squash duplicate replaced characters.
8ff32507
FC
1910 r Return the modified string and leave the original string
1911 untouched.
a0d0e21e 1912
19799a22
GS
1913If the C</c> modifier is specified, the SEARCHLIST character set
1914is complemented. If the C</d> modifier is specified, any characters
1915specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
1916(Note that this is slightly more flexible than the behavior of some
1917B<tr> programs, which delete anything they find in the SEARCHLIST,
1918period.) If the C</s> modifier is specified, sequences of characters
1919that were transliterated to the same character are squashed down
1920to a single instance of the character.
a0d0e21e
LW
1921
1922If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1923exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1924than the SEARCHLIST, the final character is replicated till it is long
5a964f20 1925enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
a0d0e21e
LW
1926This latter is useful for counting characters in a class or for
1927squashing character sequences in a class.
1928
1929Examples:
1930
1931 \$ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1932
1933 \$cnt = tr/*/*/; # count the stars in \$_
1934
1935 \$cnt = \$sky =~ tr/*/*/; # count the stars in \$sky
1936
1937 \$cnt = tr/0-9//; # count the digits in \$_
1938
1939 tr/a-zA-Z//s; # bookkeeper -> bokeper
1940
1941 (\$HOST = \$host) =~ tr/a-z/A-Z/;
8ff32507
FC
1942 \$HOST = \$host =~ tr/a-z/A-Z/r; # same thing
1943
1944 \$HOST = \$host =~ tr/a-z/A-Z/r # chained with s///
1945 =~ s/:/ -p/r;
a0d0e21e
LW
1946
1947 tr/a-zA-Z/ /cs; # change non-alphas to single space
1948
8ff32507
FC
1949 @stripped = map tr/a-zA-Z/ /csr, @original;
1950 # /r with map
1951
a0d0e21e
LW
1952 tr [\200-\377]
1953 [\000-\177]; # delete 8th bit
1954
19799a22
GS
1955If multiple transliterations are given for a character, only the
1956first one is used:
748a9306
LW
1957
1958 tr/AAA/XYZ/
1959
2c268ad5 1960will transliterate any A to X.
748a9306 1961
19799a22 1962Because the transliteration table is built at compile time, neither
a0d0e21e 1963the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
19799a22
GS
1964interpolation. That means that if you want to use variables, you
1965must use an eval():
a0d0e21e
LW
1966
1967 eval "tr/\$oldlist/\$newlist/";
1968 die \$@ if \$@;
1969
1970 eval "tr/\$oldlist/\$newlist/, 1" or die \$@;
1971
7e3b091d 1972=item <<EOF
d74e8afc 1973X<here-doc> X<heredoc> X<here-document> X<<< << >>>
7e3b091d
DA
1974
1975A line-oriented form of quoting is based on the shell "here-document"
1976syntax. Following a C<< << >> you specify a string to terminate
1977the quoted material, and all lines following the current line down to
89d205f2
YO
1978the terminating string are the value of the item.
1979
1980The terminating string may be either an identifier (a word), or some
1981quoted text. An unquoted identifier works like double quotes.
1982There may not be a space between the C<< << >> and the identifier,
1983unless the identifier is explicitly quoted. (If you put a space it
1984will be treated as a null identifier, which is valid, and matches the
1985first empty line.) The terminating string must appear by itself
1986(unquoted and with no surrounding whitespace) on the terminating line.
1987
1988If the terminating string is quoted, the type of quotes used determine
1989the treatment of the text.
1990
1991=over 4
1992
1993=item Double Quotes
1994
1995Double quotes indicate that the text will be interpolated using exactly
1996the same rules as normal double quoted strings.
7e3b091d
DA
1997
1998 print <<EOF;
1999 The price is \$Price.
2000 EOF
2001
2002 print << "EOF"; # same as above
2003 The price is \$Price.
2004 EOF
2005
89d205f2
YO
2006
2007=item Single Quotes
2008
2009Single quotes indicate the text is to be treated literally with no
2010interpolation of its content. This is similar to single quoted
2011strings except that backslashes have no special meaning, with C<\\>
2012being treated as two backslashes and not one as they would in every
2013other quoting construct.
2014
2015This is the only form of quoting in perl where there is no need
2016to worry about escaping content, something that code generators
2017can and do make good use of.
2018
2019=item Backticks
2020
2021The content of the here doc is treated just as it would be if the
2022string were embedded in backticks. Thus the content is interpolated
2023as though it were double quoted and then executed via the shell, with
2024the results of the execution returned.
2025
2026 print << `EOC`; # execute command and get results
7e3b091d 2027 echo hi there
7e3b091d
DA
2028 EOC
2029
89d205f2
YO
2030=back
2031
2032It is possible to stack multiple here-docs in a row:
2033
7e3b091d
DA
2034 print <<"foo", <<"bar"; # you can stack them
2035 I said foo.
2036 foo
2037 I said bar.
2038 bar
2039
2040 myfunc(<< "THIS", 23, <<'THAT');
2041 Here's a line
2042 or two.
2043 THIS
2044 and here's another.
2045 THAT
2046
2047Just don't forget that you have to put a semicolon on the end
2048to finish the statement, as Perl doesn't know you're not going to
2049try to do this:
2050
2051 print <<ABC
2052 179231
2053 ABC
2054 + 20;
2055
872d7e53
TS
2056If you want to remove the line terminator from your here-docs,
2057use C<chomp()>.
2058
2059 chomp(\$string = <<'END');
2060 This is a string.
2061 END
2062
2063If you want your here-docs to be indented with the rest of the code,
2064you'll need to remove leading whitespace from each line manually:
7e3b091d
DA
2065
2066 (\$quote = <<'FINIS') =~ s/^\s+//gm;
89d205f2 2067 The Road goes ever on and on,
7e3b091d
DA
2068 down from the door where it began.
2069 FINIS
2070
2071If you use a here-doc within a delimited construct, such as in C<s///eg>,
2072the quoted material must come on the lines following the final delimiter.
2074
2075 s/this/<<E . 'that'
2076 the other
2077 E
2078 . 'more '/eg;
2079
2080you have to write
2081
89d205f2
YO
2082 s/this/<<E . 'that'
2083 . 'more '/eg;
2084 the other
2085 E
7e3b091d
DA
2086
2087If the terminating identifier is on the last line of the program, you
2088must be sure there is a newline after it; otherwise, Perl will give the
2089warning B<Can't find string terminator "END" anywhere before EOF...>.
2090
89d205f2 2091Additionally, the quoting rules for the end of string identifier are not
ac036724 2092related to Perl's quoting rules. C<q()>, C<qq()>, and the like are not
89d205f2
YO
2093supported in place of C<''> and C<"">, and the only interpolation is for
2094backslashing the quoting character:
7e3b091d
DA
2095
2096 print << "abc\"def";
2097 testing...
2098 abc"def
2099
2100Finally, quoted strings cannot span multiple lines. The general rule is
2101that the identifier must be a string literal. Stick with that, and you
2102should be safe.
2103
a0d0e21e
LW
2104=back
2105
75e14d17 2106=head2 Gory details of parsing quoted constructs
d74e8afc 2107X<quote, gory details>
75e14d17 2108
19799a22
GS
2109When presented with something that might have several different
2110interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
2111principle to pick the most probable interpretation. This strategy
2112is so successful that Perl programmers often do not suspect the
2113ambivalence of what they write. But from time to time, Perl's
2114notions differ substantially from what the author honestly meant.
2115
2116This section hopes to clarify how Perl handles quoted constructs.
2117Although the most common reason to learn this is to unravel labyrinthine
2118regular expressions, because the initial steps of parsing are the
2119same for all quoting operators, they are all discussed together.
2120
2121The most important Perl parsing rule is the first one discussed
2122below: when processing a quoted construct, Perl first finds the end
2123of that construct, then interprets its contents. If you understand
2124this rule, you may skip the rest of this section on the first
2125reading. The other rules are likely to contradict the user's
2126expectations much less frequently than this first one.
2127
2128Some passes discussed below are performed concurrently, but because
2129their results are the same, we consider them individually. For different
2130quoting constructs, Perl performs different numbers of passes, from
6deea57f 2131one to four, but these passes are always performed in the same order.
75e14d17 2132
13a2d996 2133=over 4
75e14d17
IZ
2134
2135=item Finding the end
2136
6deea57f
TS
2137The first pass is finding the end of the quoted construct, where
2138the information about the delimiters is used in parsing.
2139During this search, text between the starting and ending delimiters
2140is copied to a safe location. The text copied gets delimiter-independent.
2141
2142If the construct is a here-doc, the ending delimiter is a line
2143that has a terminating string as the content. Therefore C<<<EOF> is
2144terminated by C<EOF> immediately followed by C<"\n"> and starting
2145from the first column of the terminating line.
2146When searching for the terminating line of a here-doc, nothing
2147is skipped. In other words, lines after the here-doc syntax
2148are compared with the terminating string line by line.
2149
2150For the constructs except here-docs, single characters are used as starting
2151and ending delimiters. If the starting delimiter is an opening punctuation
2152(that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
2153corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
2154If the starting delimiter is an unpaired character like C</> or a closing
2155punctuation, the ending delimiter is same as the starting delimiter.
2156Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
2157C<qq[]> and C<qq]]> constructs.
2158
2159When searching for single-character delimiters, escaped delimiters
2160and C<\\> are skipped. For example, while searching for terminating C</>,
2161combinations of C<\\> and C<\/> are skipped. If the delimiters are
2162bracketing, nested pairs are also skipped. For example, while searching
2163for closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
2164and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
2165However, when backslashes are used as the delimiters (like C<qq\\> and
2166C<tr\\\>), nothing is skipped.
2167During the search for the end, backslashes that escape delimiters
2168are removed (exactly speaking, they are not copied to the safe location).
75e14d17 2169
19799a22
GS
2170For constructs with three-part delimiters (C<s///>, C<y///>, and
2171C<tr///>), the search is repeated once more.
6deea57f
TS
2172If the first delimiter is not an opening punctuation, three delimiters must
2173be same such as C<s!!!> and C<tr)))>, in which case the second delimiter
2174terminates the left part and starts the right part at once.
b6538e4f 2175If the left part is delimited by bracketing punctuation (that is C<()>,
6deea57f 2176C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
b6538e4f 2177delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespace
6deea57f 2178and comments are allowed between both parts, though the comment must follow
b6538e4f
TC
2179at least one whitespace character; otherwise a character expected as the
2180start of the comment may be regarded as the starting delimiter of the right part.
75e14d17 2181
19799a22
GS
2182During this search no attention is paid to the semantics of the construct.
2183Thus:
75e14d17
IZ
2184
2185 "\$hash{"\$foo/\$bar"}"
2186
2a94b7ce 2187or:
75e14d17 2188
89d205f2 2189 m/
2a94b7ce 2190 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
2191 /x
2192
19799a22
GS
2193do not form legal quoted expressions. The quoted part ends on the
2194first C<"> and C</>, and the rest happens to be a syntax error.
2195Because the slash that terminated C<m//> was followed by a C<SPACE>,
2196the example above is not C<m//x>, but rather C<m//> with no C</x>
2197modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 2198
89d205f2
YO
2199Also no attention is paid to C<\c\> (multichar control char syntax) during
2200this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
2201of C<\/>, and the following C</> is not recognized as a delimiter.
0d594e51
TS
2202Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
2203
75e14d17 2204=item Interpolation
d74e8afc 2205X<interpolation>
75e14d17 2206
19799a22 2207The next step is interpolation in the text obtained, which is now
89d205f2 2208delimiter-independent. There are multiple cases.
75e14d17 2209
13a2d996 2210=over 4
75e14d17 2211
89d205f2 2212=item C<<<'EOF'>
75e14d17
IZ
2213
2214No interpolation is performed.
6deea57f
TS
2215Note that the combination C<\\> is left intact, since escaped delimiters
2216are not available for here-docs.
75e14d17 2217
6deea57f 2218=item C<m''>, the pattern of C<s'''>
89d205f2 2219
6deea57f
TS
2220No interpolation is performed at this stage.
2221Any backslashed sequences including C<\\> are treated at the stage
2222to L</"parsing regular expressions">.
89d205f2 2223
6deea57f 2224=item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
75e14d17 2225
89d205f2 2226The only interpolation is removal of C<\> from pairs of C<\\>.
6deea57f
TS
2227Therefore C<-> in C<tr'''> and C<y'''> is treated literally
2228as a hyphen and no character range is available.
2229C<\1> in the replacement of C<s'''> does not work as C<\$1>.
89d205f2
YO
2230
2231=item C<tr///>, C<y///>
2232
6deea57f
TS
2233No variable interpolation occurs. String modifying combinations for
2234case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
2235The other escape sequences such as C<\200> and C<\t> and backslashed
2236characters such as C<\\> and C<\-> are converted to appropriate literals.
89d205f2
YO
2237The character C<-> is treated specially and therefore C<\-> is treated
2238as a literal C<->.
75e14d17 2239
89d205f2 2240=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
75e14d17 2241
19799a22
GS
2242C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
2243converted to corresponding Perl constructs. Thus, C<"\$foo\Qbaz\$bar">
2244is converted to C<\$foo . (quotemeta("baz" . \$bar))> internally.
6deea57f
TS
2245The other escape sequences such as C<\200> and C<\t> and backslashed
2246characters such as C<\\> and C<\-> are replaced with appropriate
2247expansions.
2a94b7ce 2248
19799a22
GS
2249Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
2250is interpolated in the usual way. Something like C<"\Q\\E"> has
2251no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
2252result is the same as for C<"\\\\E">. As a general rule, backslashes
2253between C<\Q> and C<\E> may lead to counterintuitive results. So,
2254C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
2255as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
2256
2257 \$str = '\t';
2258 return "\Q\$str";
2259
2260may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
2261
19799a22 2262Interpolated scalars and arrays are converted internally to the C<join> and
92d29cee 2263C<.> catenation operations. Thus, C<"\$foo XXX '@arr'"> becomes:
75e14d17 2264
19799a22 2265 \$foo . " XXX '" . (join \$", @arr) . "'";
75e14d17 2266
19799a22 2267All operations above are performed simultaneously, left to right.
75e14d17 2268
19799a22
GS
2269Because the result of C<"\Q STRING \E"> has all metacharacters
2270quoted, there is no way to insert a literal C<\$> or C<@> inside a
2271C<\Q\E> pair. If protected by C<\>, C<\$> will be quoted to became
2272C<"\\\\$">; if not, it is interpreted as the start of an interpolated
2273scalar.
75e14d17 2274
19799a22 2275Note also that the interpolation code needs to make a decision on
89d205f2 2276where the interpolated scalar ends. For instance, whether
35f2feb0 2277C<< "a \$b -> {c}" >> really means:
75e14d17
IZ
2278
2279 "a " . \$b . " -> {c}";
2280
2a94b7ce 2281or:
75e14d17
IZ
2282
2283 "a " . \$b -> {c};
2284
19799a22
GS
2285Most of the time, the longest possible text that does not include
2286spaces between components and which contains matching braces or
2287brackets. because the outcome may be determined by voting based
2288on heuristic estimators, the result is not strictly predictable.
2289Fortunately, it's usually correct for ambiguous cases.
75e14d17 2290
6deea57f 2291=item the replacement of C<s///>
75e14d17 2292
19799a22 2293Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
6deea57f
TS
2294happens as with C<qq//> constructs.
2295
2296It is at this step that C<\1> is begrudgingly converted to C<\$1> in
2297the replacement text of C<s///>, in order to correct the incorrigible
2298I<sed> hackers who haven't picked up the saner idiom yet. A warning
2299is emitted if the C<use warnings> pragma or the B<-w> command-line flag
2300(that is, the C<\$^W> variable) was set.
2301
2302=item C<RE> in C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
2303
cc74c5bd
TS
2304Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\E>,
2305and interpolation happens (almost) as with C<qq//> constructs.
2306
5d03b57c
KW
2307Processing of C<\N{...}> is also done here, and compiled into an intermediate
2308form for the regex compiler. (This is because, as mentioned below, the regex
2309compilation may be done at execution time, and C<\N{...}> is a compile-time
2310construct.)
2311
cc74c5bd
TS
2312However any other combinations of C<\> followed by a character
2313are not substituted but only skipped, in order to parse them
2314as regular expressions at the following step.
6deea57f 2315As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
1749ea0d 2316treated as an array symbol (for example C<@foo>),
6deea57f 2317even though the same text in C<qq//> gives interpolation of C<\c@>.
6deea57f
TS
2318
2319Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
19799a22
GS
2320a C<#>-comment in a C<//x>-regular expression, no processing is
2321performed whatsoever. This is the first step at which the presence
2322of the C<//x> modifier is relevant.
2323
1749ea0d
TS
2324Interpolation in patterns has several quirks: C<\$|>, C<\$(>, C<\$)>, C<@+>
2325and C<@-> are not interpolated, and constructs C<\$var[SOMETHING]> are
2326voted (by several different estimators) to be either an array element
2327or C<\$var> followed by an RE alternative. This is where the notation
19799a22
GS
2328C<\${arr[\$bar]}> comes handy: C</\${arr[0-9]}/> is interpreted as
2329array element C<-9>, not as a regular expression from the variable
2330C<\$arr> followed by a digit, which would be the interpretation of
2331C</\$arr[0-9]/>. Since voting among different estimators may occur,
2332the result is not predictable.
2333
19799a22
GS
2334The lack of processing of C<\\> creates specific restrictions on
2335the post-processed text. If the delimiter is C</>, one cannot get
2336the combination C<\/> into the result of this step. C</> will
2337finish the regular expression, C<\/> will be stripped to C</> on
2338the previous step, and C<\\/> will be left as is. Because C</> is
2339equivalent to C<\/> inside a regular expression, this does not
2340matter unless the delimiter happens to be character special to the
2341RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
2342alphanumeric char, as in:
2a94b7ce
IZ
2343
2344 m m ^ a \s* b mmx;
2345
19799a22 2346In the RE above, which is intentionally obfuscated for illustration, the
6deea57f 2347delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
89d205f2 2348RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
19799a22
GS
2349reason you're encouraged to restrict your delimiters to non-alphanumeric,
2350non-whitespace choices.
75e14d17
IZ
2351
2352=back
2353
19799a22 2354This step is the last one for all constructs except regular expressions,
75e14d17
IZ
2355which are processed further.
2356
6deea57f
TS
2357=item parsing regular expressions
2358X<regexp, parse>
75e14d17 2359
19799a22 2360Previous steps were performed during the compilation of Perl code,
ac036724 2361but this one happens at run time, although it may be optimized to
19799a22 2362be calculated at compile time if appropriate. After preprocessing
6deea57f 2363described above, and possibly after evaluation if concatenation,
19799a22
GS
2364joining, casing translation, or metaquoting are involved, the
2365resulting I<string> is passed to the RE engine for compilation.
2366
2367Whatever happens in the RE engine might be better discussed in L<perlre>,
2368but for the sake of continuity, we shall do so here.
2369
2370This is another step where the presence of the C<//x> modifier is
2371relevant. The RE engine scans the string from left to right and
2372converts it to a finite automaton.
2373
2374Backslashed characters are either replaced with corresponding
2375literal strings (as with C<\{>), or else they generate special nodes
2376in the finite automaton (as with C<\b>). Characters special to the
2377RE engine (such as C<|>) generate corresponding nodes or groups of
2378nodes. C<(?#...)> comments are ignored. All the rest is either
2379converted to literal strings to match, or else is ignored (as is
2380whitespace and C<#>-style comments if C<//x> is present).
2381
2382Parsing of the bracketed character class construct, C<[...]>, is
2383rather different than the rule used for the rest of the pattern.
2384The terminator of this construct is found using the same rules as
2385for finding the terminator of a C<{}>-delimited construct, the only
2386exception being that C<]> immediately following C<[> is treated as
2387though preceded by a backslash. Similarly, the terminator of
2388C<(?{...})> is found using the same rules as for finding the
2389terminator of a C<{}>-delimited construct.
2390
2391It is possible to inspect both the string given to RE engine and the
2392resulting finite automaton. See the arguments C<debug>/C<debugcolor>
2393in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 2394switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
2395
2396=item Optimization of regular expressions
d74e8afc 2397X<regexp, optimization>
75e14d17 2398
7522fed5 2399This step is listed for completeness only. Since it does not change
75e14d17 2400semantics, details of this step are not documented and are subject
19799a22
GS
2401to change without notice. This step is performed over the finite
2402automaton that was generated during the previous pass.
2a94b7ce 2403
19799a22
GS
2404It is at this stage that C<split()> silently optimizes C</^/> to
2405mean C</^/m>.
75e14d17
IZ
2406
2407=back
2408
a0d0e21e 2409=head2 I/O Operators
d74e8afc
ITB
2410X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
2411X<< <> >> X<@ARGV>
a0d0e21e 2412
54310121 2413There are several I/O operators you should know about.
7b8d334a 2415A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
2416double-quote interpolation. It is then interpreted as an external
2417command, and the output of that command is the value of the
e9c56f9b
JH
2418backtick string, like in a shell. In scalar context, a single string
2419consisting of all output is returned. In list context, a list of
2420values is returned, one per line of output. (You can set C<\$/> to use
2421a different line terminator.) The command is executed each time the
2422pseudo-literal is evaluated. The status value of the command is
2423returned in C<\$?> (see L<perlvar> for the interpretation of C<\$?>).
2424Unlike in B<csh>, no translation is done on the return data--newlines
2425remain newlines. Unlike in any of the shells, single quotes do not
2426hide variable names in the command from interpretation. To pass a
2427literal dollar-sign through to the shell you need to hide it with a
2428backslash. The generalized form of backticks is C<qx//>. (Because
2429backticks always undergo shell expansion as well, see L<perlsec> for
2430security concerns.)
d74e8afc 2431X<qx> X<`> X<``> X<backtick> X<glob>
19799a22
GS
2432
2433In scalar context, evaluating a filehandle in angle brackets yields
2434the next line from that file (the newline, if any, included), or
2435C<undef> at end-of-file or on error. When C<\$/> is set to C<undef>
2436(sometimes known as file-slurp mode) and the file is empty, it
2437returns C<''> the first time, followed by C<undef> subsequently.
2438
2439Ordinarily you must assign the returned value to a variable, but
2440there is one situation where an automatic assignment happens. If
2441and only if the input symbol is the only thing inside the conditional
2442of a C<while> statement (even if disguised as a C<for(;;)> loop),
2443the value is automatically assigned to the global variable \$_,
2444destroying whatever was there previously. (This may seem like an
2445odd thing to you, but you'll use the construct in almost every Perl
17b829fa 2446script you write.) The \$_ variable is not implicitly localized.
19799a22
GS
2447You'll have to put a C<local \$_;> before the loop if you want that
2448to happen.
2449
2450The following lines are equivalent:
a0d0e21e 2451
748a9306 2452 while (defined(\$_ = <STDIN>)) { print; }
7b8d334a 2453 while (\$_ = <STDIN>) { print; }
a0d0e21e
LW
2454 while (<STDIN>) { print; }
2455 for (;<STDIN>;) { print; }
748a9306 2456 print while defined(\$_ = <STDIN>);
7b8d334a 2457 print while (\$_ = <STDIN>);
a0d0e21e
LW
2458 print while <STDIN>;
2459
19799a22 2460This also behaves similarly, but avoids \$_ :
7b8d334a 2461
89d205f2 2462 while (my \$line = <STDIN>) { print \$line }
7b8d334a 2463
19799a22
GS
2464In these loop constructs, the assigned value (whether assignment
2465is automatic or explicit) is then tested to see whether it is
2466defined. The defined test avoids problems where line has a string
2467value that would be treated as false by Perl, for example a "" or
2468a "0" with no trailing newline. If you really mean for such values
2469to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
2470
2471 while ((\$_ = <STDIN>) ne '0') { ... }
2472 while (<STDIN>) { last unless \$_; ... }
2473
5ef4d93e 2474In other boolean contexts, C<< <filehandle> >> without an
2475explicit C<defined> test or comparison elicits a warning if the
9f1b1f2d 2476C<use warnings> pragma or the B<-w>
19799a22 2477command-line switch (the C<\$^W> variable) is in effect.
7b8d334a 2478
5f05dabc 2479The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
2480filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
2481in packages, where they would be interpreted as local identifiers
2482rather than global.) Additional filehandles may be created with
2483the open() function, amongst others. See L<perlopentut> and
2484L<perlfunc/open> for details on this.
d74e8afc 2485X<stdin> X<stdout> X<sterr>
a0d0e21e 2486
35f2feb0 2487If a <FILEHANDLE> is used in a context that is looking for
19799a22
GS
2488a list, a list comprising all input lines is returned, one line per
2489list element. It's easy to grow to a rather large data space this
2490way, so use with care.
a0d0e21e 2491
35f2feb0 2492<FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
35f2feb0
GS
2495The null filehandle <> is special: it can be used to emulate the
2496behavior of B<sed> and B<awk>. Input from <> comes either from
a0d0e21e 2497standard input, or from each file listed on the command line. Here's
35f2feb0 2498how it works: the first time <> is evaluated, the @ARGV array is
5a964f20 2499checked, and if it is empty, C<\$ARGV[0]> is set to "-", which when opened
a0d0e21e
LW
2500gives you standard input. The @ARGV array is then processed as a list
2501of filenames. The loop
2502
2503 while (<>) {
2504 ... # code for each line
2505 }
2506
2507is equivalent to the following Perl-like pseudo code:
2508
3e3baf6d 2509 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
2510 while (\$ARGV = shift) {
2511 open(ARGV, \$ARGV);
2512 while (<ARGV>) {
2513 ... # code for each line
2514 }
2515 }
2516
19799a22
GS
2517except that it isn't so cumbersome to say, and will actually work.
2518It really does shift the @ARGV array and put the current filename
2519into the \$ARGV variable. It also uses filehandle I<ARGV>
ac036724 2520internally. <> is just a synonym for <ARGV>, which
19799a22 2521is magical. (The pseudo code above doesn't work because it treats
35f2feb0 2522<ARGV> as non-magical.)
a0d0e21e 2523
48ab5743
ML
2524Since the null filehandle uses the two argument form of L<perlfunc/open>
2525it interprets special characters, so if you have a script like this:
2526
2527 while (<>) {
2528 print;
2529 }
2530
2531and call it with C<perl dangerous.pl 'rm -rfv *|'>, it actually opens a
2532pipe, executes the C<rm> command and reads C<rm>'s output from that pipe.
2533If you want all items in C<@ARGV> to be interpreted as file names, you
2534can use the module C<ARGV::readonly> from CPAN.
2535
35f2feb0 2536You can modify @ARGV before the first <> as long as the array ends up
a0d0e21e 2537containing the list of filenames you really want. Line numbers (C<\$.>)
19799a22
GS
2538continue as though the input were one big happy file. See the example
2539in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20 2540
89d205f2 2541If you want to set @ARGV to your own list of files, go right ahead.
5a964f20
TC
2542This sets @ARGV to all plain text files if no @ARGV was given:
2543
2544 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 2545
5a964f20
TC
2546You can even set them to pipe commands. For example, this automatically
2547filters compressed arguments through B<gzip>:
2548
2549 @ARGV = map { /\.(gz|Z)\$/ ? "gzip -dc < \$_ |" : \$_ } @ARGV;
2550
2551If you want to pass switches into your script, you can use one of the
a0d0e21e
LW
2552Getopts modules or put a loop on the front like this:
2553
2554 while (\$_ = \$ARGV[0], /^-/) {
2555 shift;
2556 last if /^--\$/;
2557 if (/^-D(.*)/) { \$debug = \$1 }
2558 if (/^-v/) { \$verbose++ }
5a964f20 2559 # ... # other switches
a0d0e21e 2560 }
5a964f20 2561
a0d0e21e 2562 while (<>) {
5a964f20 2563 # ... # code for each line
a0d0e21e
LW
2564 }
2565
89d205f2
YO
2566The <> symbol will return C<undef> for end-of-file only once.
2567If you call it again after this, it will assume you are processing another
19799a22 2568@ARGV list, and if you haven't set @ARGV, will read input from STDIN.
a0d0e21e 2569
b159ebd3 2570If what the angle brackets contain is a simple scalar variable (e.g.,
35f2feb0 2571<\$foo>), then that variable contains the name of the
19799a22
GS
2572filehandle to input from, or its typeglob, or a reference to the
2573same. For example:
cb1a09d0
2574
2575 \$fh = \*STDIN;
2576 \$line = <\$fh>;
a0d0e21e 2577
5a964f20
TC
2578If what's within the angle brackets is neither a filehandle nor a simple
2579scalar variable containing a filehandle name, typeglob, or typeglob
2580reference, it is interpreted as a filename pattern to be globbed, and
2581either a list of filenames or the next filename in the list is returned,
19799a22 2582depending on context. This distinction is determined on syntactic
35f2feb0
GS
2583grounds alone. That means C<< <\$x> >> is always a readline() from
2584an indirect handle, but C<< <\$hash{key}> >> is always a glob().
5a964f20 2585That's because \$x is a simple scalar variable, but C<\$hash{key}> is
ef191992
YST
2586not--it's a hash element. Even C<< <\$x > >> (note the extra space)
2587is treated as C<glob("\$x ")>, not C<readline(\$x)>.
5a964f20
TC
2588
2589One level of double-quote interpretation is done first, but you can't
35f2feb0 2590say C<< <\$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
2591in the previous paragraph. (In older versions of Perl, programmers
2592would insert curly brackets to force interpretation as a filename glob:
35f2feb0 2593C<< <\${foo}> >>. These days, it's considered cleaner to call the
5a964f20 2594internal function directly as C<glob(\$foo)>, which is probably the right
19799a22 2595way to have done it in the first place.) For example:
a0d0e21e
LW
2596
2597 while (<*.c>) {
2598 chmod 0644, \$_;
2599 }
2600
3a4b19e4 2601is roughly equivalent to:
a0d0e21e
LW
2602
2603 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
2604 while (<FOO>) {
5b3eff12 2605 chomp;
a0d0e21e
LW
2606 chmod 0644, \$_;
2607 }
2608
3a4b19e4
GS
2609except that the globbing is actually done internally using the standard
2610C<File::Glob> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
2611
2612 chmod 0644, <*.c>;
2613
19799a22
GS
2614A (file)glob evaluates its (embedded) argument only when it is
2615starting a new list. All values must be read before it will start
2616over. In list context, this isn't important because you automatically
2617get them all anyway. However, in scalar context the operator returns
069e01df 2618the next value each time it's called, or C<undef> when the list has
19799a22
GS
2619run out. As with filehandle reads, an automatic C<defined> is
2620generated when the glob occurs in the test part of a C<while>,
2621because legal glob returns (e.g. a file called F<0>) would otherwise
2622terminate the loop. Again, C<undef> is returned only once. So if
2623you're expecting a single value from a glob, it is much better to
2624say
4633a7c4
LW
2625
2626 (\$file) = <blurch*>;
2627
2628than
2629
2630 \$file = <blurch*>;
2631
2632because the latter will alternate between returning a filename and
19799a22 2633returning false.
4633a7c4 2634
b159ebd3 2635If you're trying to do variable interpolation, it's definitely better
4633a7c4 2636to use the glob() function, because the older notation can cause people
e37d713d 2637to become confused with the indirect filehandle notation.
4633a7c4
LW
2638
2639 @files = glob("\$dir/*.[ch]");
2640 @files = glob(\$files[\$i]);
2641
a0d0e21e 2642=head2 Constant Folding
d74e8afc 2643X<constant folding> X<folding>
a0d0e21e
LW
2644
2645Like C, Perl does a certain amount of expression evaluation at
19799a22 2646compile time whenever it determines that all arguments to an
a0d0e21e
LW
2647operator are static and have no side effects. In particular, string
2648concatenation happens at compile time between literals that don't do
19799a22 2649variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
2650compile time. You can say
2651
2652 'Now is the time for all' . "\n" .
2653 'good men to come to.'
2654
54310121 2655and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
2656you say
2657
2658 foreach \$file (@filenames) {
5a964f20 2659 if (-s \$file > 5 + 100 * 2**16) { }
54310121 2660 }
a0d0e21e 2661
19799a22
GS
2662the compiler will precompute the number which that expression
2663represents so that the interpreter won't have to.
a0d0e21e 2664
d74e8afc 2666X<no-op> X<nop>
fd1abbef
DN
2667
2668Perl doesn't officially have a no-op operator, but the bare constants
2669C<0> and C<1> are special-cased to not produce a warning in a void
2670context, so you can for example safely do
2671
2672 1 while foo();
2673
d74e8afc 2675X<operator, bitwise, string>
TP
2676
2677Bitstrings of any size may be manipulated by the bitwise operators
2678(C<~ | & ^>).
2679
19799a22
GS
2680If the operands to a binary bitwise op are strings of different
2681sizes, B<|> and B<^> ops act as though the shorter operand had
2682additional zero bits on the right, while the B<&> op acts as though
2683the longer operand were truncated to the length of the shorter.
2684The granularity for such extension or truncation is one or more
2685bytes.
89d205f2 2687 # ASCII-based examples
TP
2688 print "j p \n" ^ " a h"; # prints "JAPH\n"
2689 print "JA" | " ph\n"; # prints "japh\n"
2690 print "japh\nJunk" & '_____'; # prints "JAPH\n";
2691 print 'p N\$' ^ " E<H\n"; # prints "Perl\n";
2692
19799a22 2693If you are intending to manipulate bitstrings, be certain that
2c268ad5 2694you're supplying bitstrings: If an operand is a number, that will imply
19799a22 2695a B<numeric> bitwise operation. You may explicitly show which type of
TP
2696operation you intend by using C<""> or C<0+>, as in the examples below.
2697
4358a253
SS
2698 \$foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
2699 \$foo = '150' | 105; # yields 255
TP
2700 \$foo = 150 | '105'; # yields 255
2701 \$foo = '150' | '105'; # yields string '155' (under ASCII)
2702
2703 \$baz = 0+\$foo & 0+\$bar; # both ops explicitly numeric
2704 \$biz = "\$foo" ^ "\$bar"; # both ops explicitly stringy
a0d0e21e 2705
1ae175c8
GS
2706See L<perlfunc/vec> for information on how to manipulate individual bits
2707in a bit vector.
2708
55497cff 2709=head2 Integer Arithmetic
d74e8afc 2710X<integer>
a0d0e21e 2711
19799a22 2712By default, Perl assumes that it must do most of its arithmetic in
a0d0e21e
LW
2713floating point. But by saying
2714
2715 use integer;
2716
3eab78e3
CW
2717you may tell the compiler to use integer operations
2718(see L<integer> for a detailed explanation) from here to the end of
2719the enclosing BLOCK. An inner BLOCK may countermand this by saying
a0d0e21e
LW
2720
2721 no integer;
2722
19799a22 2723which lasts until the end of that BLOCK. Note that this doesn't
3eab78e3
CW
2724mean everything is an integer, merely that Perl will use integer
2725operations for arithmetic, comparison, and bitwise operators. For
2726example, even under C<use integer>, if you take the C<sqrt(2)>, you'll
2727still get C<1.4142135623731> or so.
19799a22
GS
2728
2729Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
89d205f2 2730and ">>") always produce integral results. (But see also
13a2d996 2731L<Bitwise String Operators>.) However, C<use integer> still has meaning for
19799a22
GS
2732them. By default, their results are interpreted as unsigned integers, but
2733if C<use integer> is in effect, their results are interpreted
2734as signed integers. For example, C<~0> usually evaluates to a large
0be96356 2735integral value. However, C<use integer; ~0> is C<-1> on two's-complement
19799a22 2736machines.
68dc0745 2737
d74e8afc 2739X<floating-point> X<floating point> X<float> X<real>
68dc0745 2740
2741While C<use integer> provides integer-only arithmetic, there is no
19799a22
GS
2742analogous mechanism to provide automatic rounding or truncation to a
2743certain number of decimal places. For rounding to a certain number
2744of digits, sprintf() or printf() is usually the easiest route.
2745See L<perlfaq4>.
68dc0745 2746
5a964f20
TC
2747Floating-point numbers are only approximations to what a mathematician
2748would call real numbers. There are infinitely more reals than floats,
2749so some corners must be cut. For example:
2750
2751 printf "%.20g\n", 123456789123456789;
2752 # produces 123456789123456784
2753
8548cb57
RGS
2754Testing for exact floating-point equality or inequality is not a
2755good idea. Here's a (relatively expensive) work-around to compare
5a964f20
TC
2756whether two floating-point numbers are equal to a particular number of
2757decimal places. See Knuth, volume II, for a more robust treatment of
2758this topic.
2759
2760 sub fp_equal {
2761 my (\$X, \$Y, \$POINTS) = @_;
2762 my (\$tX, \$tY);
2763 \$tX = sprintf("%.\${POINTS}g", \$X);
2764 \$tY = sprintf("%.\${POINTS}g", \$Y);
2765 return \$tX eq \$tY;
2766 }
2767
68dc0745 2768The POSIX module (part of the standard perl distribution) implements
19799a22
GS
2769ceil(), floor(), and other mathematical and trigonometric functions.
2770The Math::Complex module (part of the standard perl distribution)
2771defines mathematical functions that work on both the reals and the
2772imaginary numbers. Math::Complex not as efficient as POSIX, but
68dc0745 2773POSIX can't work with complex numbers.
2774
2775Rounding in financial applications can have serious implications, and
2776the rounding method used should be specified precisely. In these
2777cases, it probably pays not to trust whichever system rounding is
2778being used by Perl, but to instead implement the rounding function you
2779need yourself.
5a964f20
TC
2780
d74e8afc 2782X<number, arbitrary precision>
5a964f20
TC
2783
2784The standard Math::BigInt and Math::BigFloat modules provide
19799a22 2785variable-precision arithmetic and overloaded operators, although
cd5c4fce 2786they're currently pretty slow. At the cost of some space and
19799a22
GS
2787considerable speed, they avoid the normal pitfalls associated with
2788limited-precision representations.
5a964f20
TC
2789
2790 use Math::BigInt;
2791 \$x = Math::BigInt->new('123456789123456789');
2792 print \$x * \$x;
2793
2794 # prints +15241578780673678515622620750190521
19799a22 2795
cd5c4fce
T
2796There are several modules that let you calculate with (bound only by
2797memory and cpu-time) unlimited or fixed precision. There are also
2798some non-standard modules that provide faster implementations via
2799external C libraries.
2800
2801Here is a short, but incomplete summary:
2802
950b09ed
KW
2803 Math::Fraction big, unlimited fractions like 9973 / 12967
2804 Math::String treat string sequences like numbers
2805 Math::FixedPrecision calculate with a fixed precision
2806 Math::Currency for currency calculations
2807 Bit::Vector manipulate bit vectors fast (uses C)
2808 Math::BigIntFast Bit::Vector wrapper for big numbers
2809 Math::Pari provides access to the Pari C library
2810 Math::BigInteger uses an external C library
2811 Math::Cephes uses external Cephes C library (no big numbers)
2812 Math::Cephes::Fraction fractions via the Cephes library
2813 Math::GMP another one using an external C library
cd5c4fce
T
2814
2815Choose wisely.
16070b82
GS
2816
2817=cut