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
61
a0d0e21e 62=head2 Terms and List Operators (Leftward)
d74e8afc 63X<list operator> X<operator, list> X<term>
a0d0e21e 64
62c18ce2 65A TERM has the highest precedence in Perl. They include variables,
5f05dabc 66quote and quote-like operators, any expression in parentheses,
a0d0e21e
LW
67and any function whose arguments are parenthesized. Actually, there
68aren't really functions in this sense, just list operators and unary
69operators behaving as functions because you put parentheses around
70the arguments. These are all documented in L<perlfunc>.
71
72If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
73is followed by a left parenthesis as the next token, the operator and
74arguments within parentheses are taken to be of highest precedence,
75just like a normal function call.
76
77In the absence of parentheses, the precedence of list operators such as
78C<print>, C<sort>, or C<chmod> is either very high or very low depending on
54310121 79whether you are looking at the left side or the right side of the operator.
a0d0e21e
LW
80For example, in
81
82 @ary = (1, 3, sort 4, 2);
83 print @ary; # prints 1324
84
19799a22
GS
85the commas on the right of the sort are evaluated before the sort,
86but the commas on the left are evaluated after. In other words,
87list operators tend to gobble up all arguments that follow, and
a0d0e21e 88then act like a simple TERM with regard to the preceding expression.
19799a22 89Be careful with parentheses:
a0d0e21e
LW
90
91 # These evaluate exit before doing the print:
92 print(\$foo, exit); # Obviously not what you want.
93 print \$foo, exit; # Nor is this.
94
95 # These do the print before evaluating exit:
96 (print \$foo), exit; # This is what you want.
97 print(\$foo), exit; # Or this.
98 print (\$foo), exit; # Or even this.
99
100Also note that
101
102 print (\$foo & 255) + 1, "\n";
103
d042e63d
MS
104probably doesn't do what you expect at first glance. The parentheses
105enclose the argument list for C<print> which is evaluated (printing
106the result of C<\$foo & 255>). Then one is added to the return value
107of C<print> (usually 1). The result is something like this:
108
109 1 + 1, "\n"; # Obviously not what you meant.
110
111To do what you meant properly, you must write:
112
113 print((\$foo & 255) + 1, "\n");
114
115See L<Named Unary Operators> for more discussion of this.
a0d0e21e
LW
116
117Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
54310121 118well as subroutine and method calls, and the anonymous
a0d0e21e
LW
119constructors C<[]> and C<{}>.
120
2ae324a7 121See also L<Quote and Quote-like Operators> toward the end of this section,
da87341d 122as well as L</"I/O Operators">.
a0d0e21e
LW
123
d74e8afc 125X<arrow> X<dereference> X<< -> >>
a0d0e21e 126
35f2feb0 127"C<< -> >>" is an infix dereference operator, just as it is in C
19799a22
GS
128and C++. If the right side is either a C<[...]>, C<{...}>, or a
129C<(...)> subscript, then the left side must be either a hard or
130symbolic reference to an array, a hash, or a subroutine respectively.
131(Or technically speaking, a location capable of holding a hard
132reference, if it's an array or hash reference being used for
133assignment.) See L<perlreftut> and L<perlref>.
a0d0e21e 134
19799a22
GS
135Otherwise, the right side is a method name or a simple scalar
136variable containing either the method name or a subroutine reference,
137and the left side must be either an object (a blessed reference)
138or a class name (that is, a package name). See L<perlobj>.
a0d0e21e 139
d74e8afc 141X<increment> X<auto-increment> X<++> X<decrement> X<auto-decrement> X<-->
a0d0e21e 142
d042e63d
MS
143"++" and "--" work as in C. That is, if placed before a variable,
144they increment or decrement the variable by one before returning the
145value, and if placed after, increment or decrement after returning the
146value.
147
148 \$i = 0; \$j = 0;
149 print \$i++; # prints 0
150 print ++\$j; # prints 1
a0d0e21e 151
b033823e 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
d74e8afc 191X<unary operator> X<operator, unary>
a0d0e21e 192
5f05dabc 193Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
a0d0e21e 194precedence version of this.
d74e8afc 195X<!>
a0d0e21e
LW
196
197Unary "-" performs arithmetic negation if the operand is numeric. If
198the operand is an identifier, a string consisting of a minus sign
199concatenated with the identifier is returned. Otherwise, if the string
200starts with a plus or minus, a string starting with the opposite sign
bff5667c 201is returned. One effect of these rules is that -bareword is equivalent
8705167b 202to the string "-bareword". If, however, the string begins with a
353c6505 203non-alphabetic character (excluding "+" or "-"), Perl will attempt to convert
06705523
SP
204the string to a numeric and the arithmetic negation is performed. If the
205string cannot be cleanly converted to a numeric, Perl will give the warning
206B<Argument "the string" isn't numeric in negation (-) at ...>.
d74e8afc 207X<-> X<negation, arithmetic>
a0d0e21e 208
972b05a9
JH
209Unary "~" performs bitwise negation, i.e., 1's complement. For
211L<Bitwise String Operators>.) Note that the width of the result is
212platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
213bits wide on a 64-bit platform, so if you are expecting a certain bit
d042e63d 214width, remember to use the & operator to mask off the excess bits.
d74e8afc 215X<~> X<negation, binary>
a0d0e21e
LW
216
217Unary "+" has no effect whatsoever, even on strings. It is useful
218syntactically for separating a function name from a parenthesized expression
219that would otherwise be interpreted as the complete list of function
5ba421f6 220arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
d74e8afc 221X<+>
a0d0e21e 222
19799a22
GS
223Unary "\" creates a reference to whatever follows it. See L<perlreftut>
224and L<perlref>. Do not confuse this behavior with the behavior of
225backslash within a string, although both forms do convey the notion
226of protecting the next thing from interpolation.
d74e8afc 227X<\> X<reference> X<backslash>
a0d0e21e
LW
228
d74e8afc 230X<binding> X<operator, binding> X<=~> X<!~>
a0d0e21e 231
c07a80fd 232Binary "=~" binds a scalar expression to a pattern match. Certain operations
cb1a09d0
233search or modify the string \$_ by default. This operator makes that kind
234of operation work on some other string. The right argument is a search
TP
235pattern, substitution, or transliteration. The left argument is what is
236supposed to be searched, substituted, or transliterated instead of the default
f8bab1e9 237\$_. When used in scalar context, the return value generally indicates the
000c65fc
DG
238success of the operation. The exception is substitution with the C</r>
239(non-destructive) option, which causes the return value to be the result of
240the substition. Behavior in list context depends on the particular operator.
241See L</"Regexp Quote-Like Operators"> for details and L<perlretut> for
242examples using these operators.
f8bab1e9
GS
243
244If the right argument is an expression rather than a search pattern,
2c268ad5 245substitution, or transliteration, it is interpreted as a search pattern at run
89d205f2
YO
246time. Note that this means that its contents will be interpolated twice, so
247
248 '\\' =~ q'\\';
249
250is not ok, as the regex engine will end up trying to compile the
251pattern C<\>, which it will consider a syntax error.
a0d0e21e
LW
252
253Binary "!~" is just like "=~" except the return value is negated in
254the logical sense.
255
000c65fc 256Binary "!~" with a non-destructive substitution (s///r) is a syntax error.
4f4d7508 257
d74e8afc 259X<operator, multiplicative>
a0d0e21e
LW
260
261Binary "*" multiplies two numbers.
d74e8afc 262X<*>
a0d0e21e
LW
263
264Binary "/" divides two numbers.
d74e8afc 265X</> X<slash>
a0d0e21e 266
f7918450
KW
267Binary "%" is the modulo operator, which computes the division
268remainder of its first argument with respect to its second argument.
269Given integer
54310121 270operands C<\$a> and C<\$b>: If C<\$b> is positive, then C<\$a % \$b> is
f7918450 271C<\$a> minus the largest multiple of C<\$b> less than or equal to
54310121 272C<\$a>. If C<\$b> is negative, then C<\$a % \$b> is C<\$a> minus the
273smallest multiple of C<\$b> that is not less than C<\$a> (i.e. the
89b4f0ad 274result will be less than or equal to zero). If the operands
4848a83b
TS
275C<\$a> and C<\$b> are floating point values and the absolute value of
276C<\$b> (that is C<abs(\$b)>) is less than C<(UV_MAX + 1)>, only
277the integer portion of C<\$a> and C<\$b> will be used in the operation
278(Note: here C<UV_MAX> means the maximum of the unsigned integer type).
279If the absolute value of the right operand (C<abs(\$b)>) is greater than
280or equal to C<(UV_MAX + 1)>, "%" computes the floating-point remainder
281C<\$r> in the equation C<(\$r = \$a - \$i*\$b)> where C<\$i> is a certain
f7918450 282integer that makes C<\$r> have the same sign as the right operand
4848a83b
TS
283C<\$b> (B<not> as the left operand C<\$a> like C function C<fmod()>)
284and the absolute value less than that of C<\$b>.
0412d526 285Note that when C<use integer> is in scope, "%" gives you direct access
f7918450 286to the modulo operator as implemented by your C compiler. This
55d729e4
GS
287operator is not as well defined for negative operands, but it will
288execute faster.
f7918450 289X<%> X<remainder> X<modulo> X<mod>
55d729e4 290
62d10b70
GS
291Binary "x" is the repetition operator. In scalar context or if the left
292operand is not enclosed in parentheses, it returns a string consisting
293of the left operand repeated the number of times specified by the right
294operand. In list context, if the left operand is enclosed in
3585017f
YST
295parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
296If the right operand is zero or negative, it returns an empty string
297or an empty list, depending on the context.
d74e8afc 298X<x>
a0d0e21e
LW
299
300 print '-' x 80; # print row of dashes
301
302 print "\t" x (\$tab/8), ' ' x (\$tab%8); # tab over
303
304 @ones = (1) x 80; # a list of 80 1's
305 @ones = (5) x @ones; # set all elements to 5
306
307
a0d0e21e
LW
310
311Binary "+" returns the sum of two numbers.
d74e8afc 312X<+>
a0d0e21e
LW
313
314Binary "-" returns the difference of two numbers.
d74e8afc 315X<->
a0d0e21e
LW
316
317Binary "." concatenates two strings.
d74e8afc
ITB
318X<string, concatenation> X<concatenation>
319X<cat> X<concat> X<concatenate> X<.>
a0d0e21e
LW
320
d74e8afc
ITB
322X<shift operator> X<operator, shift> X<<< << >>>
323X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
324X<shl> X<shr> X<shift, right> X<shift, left>
a0d0e21e 325
55497cff 326Binary "<<" returns the value of its left argument shifted left by the
327number of bits specified by the right argument. Arguments should be
a0d0e21e 329
55497cff 330Binary ">>" returns the value of its left argument shifted right by
331the number of bits specified by the right argument. Arguments should
a0d0e21e 333
b16cf6df
JH
334Note that both "<<" and ">>" in Perl are implemented directly using
335"<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
336in force then signed C integers are used, else unsigned C integers are
337used. Either way, the implementation isn't going to generate results
338larger than the size of the integer type Perl was built with (32 bits
339or 64 bits).
340
341The result of overflowing the range of the integers is undefined
342because it is undefined also in C. In other words, using 32-bit
343integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
344of bits is also undefined.
345
d74e8afc 347X<operator, named unary>
a0d0e21e
LW
348
349The various named unary operators are treated as functions with one
568e6d8b 350argument, with optional parentheses.
a0d0e21e
LW
351
352If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
353is followed by a left parenthesis as the next token, the operator and
354arguments within parentheses are taken to be of highest precedence,
3981b0eb
JA
355just like a normal function call. For example,
356because named unary operators are higher precedence than ||:
a0d0e21e
LW
357
358 chdir \$foo || die; # (chdir \$foo) || die
359 chdir(\$foo) || die; # (chdir \$foo) || die
360 chdir (\$foo) || die; # (chdir \$foo) || die
361 chdir +(\$foo) || die; # (chdir \$foo) || die
362
3981b0eb 363but, because * is higher precedence than named operators:
a0d0e21e
LW
364
365 chdir \$foo * 20; # chdir (\$foo * 20)
366 chdir(\$foo) * 20; # (chdir \$foo) * 20
367 chdir (\$foo) * 20; # (chdir \$foo) * 20
368 chdir +(\$foo) * 20; # chdir (\$foo * 20)
369
370 rand 10 * 20; # rand (10 * 20)
371 rand(10) * 20; # (rand 10) * 20
372 rand (10) * 20; # (rand 10) * 20
373 rand +(10) * 20; # rand (10 * 20)
374
568e6d8b
RGS
375Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
376treated like named unary operators, but they don't follow this functional
377parenthesis rule. That means, for example, that C<-f(\$file).".bak"> is
378equivalent to C<-f "\$file.bak">.
d74e8afc 379X<-X> X<filetest> X<operator, filetest>
568e6d8b 380
a0d0e21e
LW
382
d74e8afc 384X<relational operator> X<operator, relational>
a0d0e21e 385
35f2feb0 386Binary "<" returns true if the left argument is numerically less than
a0d0e21e 387the right argument.
d74e8afc 388X<< < >>
a0d0e21e 389
35f2feb0 390Binary ">" returns true if the left argument is numerically greater
a0d0e21e 391than the right argument.
d74e8afc 392X<< > >>
a0d0e21e 393
35f2feb0 394Binary "<=" returns true if the left argument is numerically less than
a0d0e21e 395or equal to the right argument.
d74e8afc 396X<< <= >>
a0d0e21e 397
35f2feb0 398Binary ">=" returns true if the left argument is numerically greater
a0d0e21e 399than or equal to the right argument.
d74e8afc 400X<< >= >>
a0d0e21e
LW
401
402Binary "lt" returns true if the left argument is stringwise less than
403the right argument.
d74e8afc 404X<< lt >>
a0d0e21e
LW
405
406Binary "gt" returns true if the left argument is stringwise greater
407than the right argument.
d74e8afc 408X<< gt >>
a0d0e21e
LW
409
410Binary "le" returns true if the left argument is stringwise less than
411or equal to the right argument.
d74e8afc 412X<< le >>
a0d0e21e
LW
413
414Binary "ge" returns true if the left argument is stringwise greater
415than or equal to the right argument.
d74e8afc 416X<< ge >>
a0d0e21e
LW
417
d74e8afc 419X<equality> X<equal> X<equals> X<operator, equality>
a0d0e21e
LW
420
421Binary "==" returns true if the left argument is numerically equal to
422the right argument.
d74e8afc 423X<==>
a0d0e21e
LW
424
425Binary "!=" returns true if the left argument is numerically not equal
426to the right argument.
d74e8afc 427X<!=>
a0d0e21e 428
35f2feb0 429Binary "<=>" returns -1, 0, or 1 depending on whether the left
6ee5d4e7 430argument is numerically less than, equal to, or greater than the right
7d3a9d88
NC
432values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
433"<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
434returns true, as does NaN != anything else. If your platform doesn't
435support NaNs then NaN is just a string with numeric value 0.
d74e8afc 436X<< <=> >> X<spaceship>
7d3a9d88 437
2b54f59f
YST
438 perl -le '\$a = "NaN"; print "No NaN support here" if \$a == \$a'
439 perl -le '\$a = "NaN"; print "NaN support here" if \$a != \$a'
a0d0e21e
LW
440
441Binary "eq" returns true if the left argument is stringwise equal to
442the right argument.
d74e8afc 443X<eq>
a0d0e21e
LW
444
445Binary "ne" returns true if the left argument is stringwise not equal
446to the right argument.
d74e8afc 447X<ne>
a0d0e21e 448
JH
449Binary "cmp" returns -1, 0, or 1 depending on whether the left
450argument is stringwise less than, equal to, or greater than the right
451argument.
d74e8afc 452X<cmp>
a0d0e21e 453
0d863452 454Binary "~~" does a smart match between its arguments. Smart matching
0f7107a0 455is described in L<perlsyn/"Smart matching in detail">.
0d863452
RH
456X<~~>
457
a034a98d
DD
458"lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
459by the current locale if C<use locale> is in effect. See L<perllocale>.
460
d74e8afc 462X<operator, bitwise, and> X<bitwise and> X<&>
a0d0e21e 463
2cdc098b 464Binary "&" returns its operands ANDed together bit by bit.
a0d0e21e 466
2cdc098b
MG
467Note that "&" has lower priority than relational operators, so for example
468the brackets are essential in a test like
469
470 print "Even\n" if (\$x & 1) == 0;
471
a0d0e21e 472=head2 Bitwise Or and Exclusive Or
d74e8afc
ITB
473X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
474X<bitwise xor> X<^>
a0d0e21e 475
2cdc098b 476Binary "|" returns its operands ORed together bit by bit.
a0d0e21e 478
2cdc098b 479Binary "^" returns its operands XORed together bit by bit.
a0d0e21e 481
2cdc098b
MG
482Note that "|" and "^" have lower priority than relational operators, so
483for example the brackets are essential in a test like
484
485 print "false\n" if (8 | 2) != 10;
486
d74e8afc 488X<&&> X<logical and> X<operator, logical, and>
a0d0e21e
LW
489
490Binary "&&" performs a short-circuit logical AND operation. That is,
491if the left operand is false, the right operand is not even evaluated.
492Scalar or list context propagates down to the right operand if it
493is evaluated.
494
d74e8afc 496X<||> X<operator, logical, or>
a0d0e21e
LW
497
498Binary "||" performs a short-circuit logical OR operation. That is,
499if the left operand is true, the right operand is not even evaluated.
500Scalar or list context propagates down to the right operand if it
501is evaluated.
502
d74e8afc 504X<//> X<operator, logical, defined-or>
c963b151
BD
505
506Although it has no direct equivalent in C, Perl's C<//> operator is related
89d205f2 507to its C-style or. In fact, it's exactly the same as C<||>, except that it
c963b151 508tests the left hand side's definedness instead of its truth. Thus, C<\$a // \$b>
89d205f2
YO
509is similar to C<defined(\$a) || \$b> (except that it returns the value of C<\$a>
510rather than the value of C<defined(\$a)>) and is exactly equivalent to
c963b151 511C<defined(\$a) ? \$a : \$b>. This is very useful for providing default values
89d205f2 512for variables. If you actually want to test if at least one of C<\$a> and
d042e63d 513C<\$b> is defined, use C<defined(\$a // \$b)>.
c963b151 514
d042e63d
MS
515The C<||>, C<//> and C<&&> operators return the last value evaluated
516(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
517portable way to find out the home directory might be:
a0d0e21e 518
c963b151
BD
519 \$home = \$ENV{'HOME'} // \$ENV{'LOGDIR'} //
520 (getpwuid(\$<))[7] // die "You're homeless!\n";
a0d0e21e 521
5a964f20
TC
522In particular, this means that you shouldn't use this
523for selecting between two aggregates for assignment:
524
525 @a = @b || @c; # this is wrong
526 @a = scalar(@b) || @c; # really meant this
527 @a = @b ? @b : @c; # this works fine, though
528
f23102e2
RGS
529As more readable alternatives to C<&&> and C<||> when used for
530control flow, Perl provides the C<and> and C<or> operators (see below).
531The short-circuit behavior is identical. The precedence of "and"
c963b151 532and "or" is much lower, however, so that you can safely use them after a
5a964f20 533list operator without the need for parentheses:
a0d0e21e
LW
534
536 or gripe(), next LINE;
537
538With the C-style operators that would have been written like this:
539
541 || (gripe(), next LINE);
542
eeb6a2c9 543Using "or" for assignment is unlikely to do what you want; see below.
5a964f20
TC
544
d74e8afc 546X<operator, range> X<range> X<..> X<...>
a0d0e21e
LW
547
548Binary ".." is the range operator, which is really two different
fb53bbb2 549operators depending on the context. In list context, it returns a
54ae734e 550list of values counting (up by ones) from the left value to the right
2cdbc966 551value. If the left value is greater than the right value then it
fb53bbb2 552returns the empty list. The range operator is useful for writing
54ae734e 553C<foreach (1..10)> loops and for doing slice operations on arrays. In
2cdbc966
JD
554the current implementation, no temporary array is created when the
555range operator is used as the expression in C<foreach> loops, but older
556versions of Perl might burn a lot of memory when you write something
557like this:
a0d0e21e
LW
558
559 for (1 .. 1_000_000) {
560 # code
54310121 561 }
a0d0e21e 562
8f0f46f8 563The range operator also works on strings, using the magical
564auto-increment, see below.
54ae734e 565
5a964f20 566In scalar context, ".." returns a boolean value. The operator is
8f0f46f8 567bistable, like a flip-flop, and emulates the line-range (comma)
568operator of B<sed>, B<awk>, and various editors. Each ".." operator
569maintains its own boolean state, even across calls to a subroutine
570that contains it. It is false as long as its left operand is false.
a0d0e21e
LW
571Once the left operand is true, the range operator stays true until the
572right operand is true, I<AFTER> which the range operator becomes false
8f0f46f8 573again. It doesn't become false till the next time the range operator
574is evaluated. It can test the right operand and become false on the
575same evaluation it became true (as in B<awk>), but it still returns
576true once. If you don't want it to test the right operand until the
577next evaluation, as in B<sed>, just use three dots ("...") instead of
19799a22
GS
578two. In all other regards, "..." behaves just like ".." does.
579
580The right operand is not evaluated while the operator is in the
581"false" state, and the left operand is not evaluated while the
582operator is in the "true" state. The precedence is a little lower
583than || and &&. The value returned is either the empty string for
8f0f46f8 584false, or a sequence number (beginning with 1) for true. The sequence
585number is reset for each range encountered. The final sequence number
586in a range has the string "E0" appended to it, which doesn't affect
587its numeric value, but gives you something to search for if you want
588to exclude the endpoint. You can exclude the beginning point by
589waiting for the sequence number to be greater than 1.
df5f8116
CW
590
591If either operand of scalar ".." is a constant expression,
592that operand is considered true if it is equal (C<==>) to the current
593input line number (the C<\$.> variable).
594
595To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
596but that is only an issue if you use a floating point expression; when
597implicitly using C<\$.> as described in the previous paragraph, the
598comparison is C<int(EXPR) == int(\$.)> which is only an issue when C<\$.>
599is set to a floating point value and you are not reading from a file.
600Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
601you want in scalar context because each of the operands are evaluated
602using their integer representation.
603
604Examples:
a0d0e21e
LW
605
606As a scalar operator:
607
df5f8116 608 if (101 .. 200) { print; } # print 2nd hundred lines, short for
950b09ed 609 # if (\$. == 101 .. \$. == 200) { print; }
9f10b797
RGS
610
611 next LINE if (1 .. /^\$/); # skip header lines, short for
f343f960 612 # next LINE if (\$. == 1 .. /^\$/);
9f10b797
RGS
613 # (typically in a loop labeled LINE)
614
615 s/^/> / if (/^\$/ .. eof()); # quote body
a0d0e21e 616
5a964f20
TC
617 # parse mail messages
618 while (<>) {
619 \$in_header = 1 .. /^\$/;
df5f8116
CW
620 \$in_body = /^\$/ .. eof;
f343f960 622 # do something
df5f8116 623 } else { # in body
f343f960 624 # do something else
df5f8116 625 }
5a964f20 626 } continue {
df5f8116 627 close ARGV if eof; # reset \$. each file
5a964f20
TC
628 }
629
acf31ca5
SF
630Here's a simple example to illustrate the difference between
631the two range operators:
632
633 @lines = (" - Foo",
634 "01 - Bar",
635 "1 - Baz",
636 " - Quux");
637
9f10b797
RGS
638 foreach (@lines) {
639 if (/0/ .. /1/) {
acf31ca5
SF
640 print "\$_\n";
641 }
642 }
643
9f10b797
RGS
644This program will print only the line containing "Bar". If
645the range operator is changed to C<...>, it will also print the
acf31ca5
SF
646"Baz" line.
647
648And now some examples as a list operator:
a0d0e21e
LW
649
650 for (101 .. 200) { print; } # print \$_ 100 times
3e3baf6d 651 @foo = @foo[0 .. \$#foo]; # an expensive no-op
a0d0e21e
LW
652 @foo = @foo[\$#foo-4 .. \$#foo]; # slice last 5 items
653
5a964f20 654The range operator (in list context) makes use of the magical
5f05dabc 655auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
656can say
657
658 @alphabet = ('A' .. 'Z');
659
54ae734e 660to get all normal letters of the English alphabet, or
a0d0e21e
LW
661
662 \$hexdigit = (0 .. 9, 'a' .. 'f')[\$num & 15];
663
664to get a hexadecimal digit, or
665
666 @z2 = ('01' .. '31'); print \$z2[\$mday];
667
ea4f5703
YST
668to get dates with leading zeros.
669
670If the final value specified is not in the sequence that the magical
671increment would produce, the sequence goes until the next value would
672be longer than the final value specified.
673
674If the initial value specified isn't part of a magical increment
675sequence (that is, a non-empty string matching "/^[a-zA-Z]*[0-9]*\z/"),
676only the initial value will be returned. So the following will only
677return an alpha:
678
679 use charnames 'greek';
680 my @greek_small = ("\N{alpha}" .. "\N{omega}");
681
682To get lower-case greek letters, use this instead:
683
950b09ed
KW
684 my @greek_small = map { chr } ( ord("\N{alpha}") ..
685 ord("\N{omega}") );
a0d0e21e 686
df5f8116
CW
687Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
688return two elements in list context.
689
690 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
691
d74e8afc 693X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
a0d0e21e
LW
694
695Ternary "?:" is the conditional operator, just as in C. It works much
696like an if-then-else. If the argument before the ? is true, the
697argument before the : is returned, otherwise the argument after the :
cb1a09d0
698is returned. For example:
699
54310121 700 printf "I have %d dog%s.\n", \$n,
cb1a09d0
701 (\$n == 1) ? '' : "s";
702
703Scalar or list context propagates downward into the 2nd
54310121 704or 3rd argument, whichever is selected.
cb1a09d0
705
706 \$a = \$ok ? \$b : \$c; # get a scalar
707 @a = \$ok ? @b : @c; # get an array
708 \$a = \$ok ? @b : @c; # oops, that's just a count!
709
710The operator may be assigned to if both the 2nd and 3rd arguments are
711legal lvalues (meaning that you can assign to them):
a0d0e21e
LW
712
713 (\$a_or_b ? \$a : \$b) = \$c;
714
5a964f20
TC
715Because this operator produces an assignable result, using assignments
716without parentheses will get you in trouble. For example, this:
717
718 \$a % 2 ? \$a += 10 : \$a += 2
719
720Really means this:
721
722 ((\$a % 2) ? (\$a += 10) : \$a) += 2
723
724Rather than this:
725
726 (\$a % 2) ? (\$a += 10) : (\$a += 2)
727
19799a22
GS
728That should probably be written more simply as:
729
730 \$a += (\$a % 2) ? 10 : 2;
731
d74e8afc 733X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
5ac3b81c 734X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
d74e8afc 735X<%=> X<^=> X<x=>
a0d0e21e
LW
736
737"=" is the ordinary assignment operator.
738
739Assignment operators work as in C. That is,
740
741 \$a += 2;
742
743is equivalent to
744
745 \$a = \$a + 2;
746
747although without duplicating any side effects that dereferencing the lvalue
54310121 748might trigger, such as from tie(). Other assignment operators work similarly.
749The following are recognized:
a0d0e21e
LW
750
751 **= += *= &= <<= &&=
9f10b797
RGS
752 -= /= |= >>= ||=
753 .= %= ^= //=
754 x=
a0d0e21e 755
19799a22 756Although these are grouped by family, they all have the precedence
a0d0e21e
LW
757of assignment.
758
b350dd2f
GS
759Unlike in C, the scalar assignment operator produces a valid lvalue.
760Modifying an assignment is equivalent to doing the assignment and
761then modifying the variable that was assigned to. This is useful
762for modifying a copy of something, like this:
a0d0e21e
LW
763
764 (\$tmp = \$global) =~ tr [A-Z] [a-z];
765
766Likewise,
767
768 (\$a += 2) *= 3;
769
770is equivalent to
771
772 \$a += 2;
773 \$a *= 3;
774
b350dd2f
GS
775Similarly, a list assignment in list context produces the list of
776lvalues assigned to, and a list assignment in scalar context returns
777the number of elements produced by the expression on the right hand
778side of the assignment.
779
d74e8afc 781X<comma> X<operator, comma> X<,>
a0d0e21e 782
5a964f20 783Binary "," is the comma operator. In scalar context it evaluates
a0d0e21e
LW
784its left argument, throws that value away, then evaluates its right
785argument and returns that value. This is just like C's comma operator.
786
5a964f20 787In list context, it's just the list argument separator, and inserts
ed5c6d31
PJ
788both its arguments into the list. These arguments are also evaluated
789from left to right.
a0d0e21e 790
344f2c40
IG
791The C<< => >> operator is a synonym for the comma except that it causes
792its left operand to be interpreted as a string if it begins with a letter
793or underscore and is composed only of letters, digits and underscores.
794This includes operands that might otherwise be interpreted as operators,
795constants, single number v-strings or function calls. If in doubt about
796this behaviour, the left operand can be quoted explicitly.
797
798Otherwise, the C<< => >> operator behaves exactly as the comma operator
799or list argument separator, according to context.
800
801For example:
a44e5664
MS
802
803 use constant FOO => "something";
804
805 my %h = ( FOO => 23 );
806
807is equivalent to:
808
809 my %h = ("FOO", 23);
810
811It is I<NOT>:
812
813 my %h = ("something", 23);
814
719b43e8
RGS
815The C<< => >> operator is helpful in documenting the correspondence
816between keys and values in hashes, and other paired elements in lists.
748a9306 817
a44e5664
MS
818 %hash = ( \$key => \$value );
820
678ae90b
RGS
be25f609 823
e8163f9b 824The yada yada operator (noted C<...>) is a placeholder for code. Perl
825parses it without error, but when you try to execute a yada yada, it
826throws an exception with the text C<Unimplemented>:
827
828 sub unimplemented { ... }
829
830 eval { unimplemented() };
831 if( \$@ eq 'Unimplemented' ) {
833 }
834
835You can only use the yada yada to stand in for a complete statement.
837
838 { ... }
839
840 sub foo { ... }
841
842 ...;
843
844 eval { ... };
845
846 sub foo {
847 my( \$self ) = shift;
848
849 ...;
850 }
851
852 do { my \$n; ...; print 'Hurrah!' };
853
854The yada yada cannot stand in for an expression that is part of a
855larger statement since the C<...> is also the three-dot version of the
856range operator (see L<Range Operators>). These examples of the yada
858
859 print ...;
860
861 open my(\$fh), '>', '/dev/passwd' or ...;
862
863 if( \$condition && ... ) { print "Hello\n" };
864
865There are some cases where Perl can't immediately tell the difference
866between an expression and a statement. For instance, the syntax for a
867block and an anonymous hash reference constructor look the same unless
868there's something in the braces that give Perl a hint. The yada yada
869is a syntax error if Perl doesn't guess that the C<{ ... }> is a
870block. In that case, it doesn't think the C<...> is the yada yada
871because it's expecting an expression instead of a statement:
872
873 my @transformed = map { ... } @input; # syntax error
874
875You can use a C<;> inside your block to denote that the C<{ ... }> is
876a block and not a hash reference constructor. Now the yada yada works:
877
878 my @transformed = map {; ... } @input; # ; disambiguates
879
880 my @transformed = map { ...; } @input; # ; disambiguates
be25f609 881
d74e8afc 883X<operator, list, rightward> X<list operator>
a0d0e21e
LW
884
885On the right side of a list operator, it has very low precedence,
886such that it controls all comma-separated expressions found there.
887The only operators with lower precedence are the logical operators
888"and", "or", and "not", which may be used to evaluate calls to list
889operators without the need for extra parentheses:
890
891 open HANDLE, "filename"
892 or die "Can't open: \$!\n";
893
5ba421f6 894See also discussion of list operators in L<Terms and List Operators (Leftward)>.
a0d0e21e
LW
895
d74e8afc 897X<operator, logical, not> X<not>
a0d0e21e
LW
898
899Unary "not" returns the logical negation of the expression to its right.
900It's the equivalent of "!" except for the very low precedence.
901
d74e8afc 903X<operator, logical, and> X<and>
a0d0e21e
LW
904
905Binary "and" returns the logical conjunction of the two surrounding
906expressions. It's equivalent to && except for the very low
5f05dabc 907precedence. This means that it short-circuits: i.e., the right
a0d0e21e
LW
908expression is evaluated only if the left expression is true.
909
c963b151 910=head2 Logical or, Defined or, and Exclusive Or
f23102e2 911X<operator, logical, or> X<operator, logical, xor>
d74e8afc 912X<operator, logical, defined or> X<operator, logical, exclusive or>
f23102e2 913X<or> X<xor>
a0d0e21e
LW
914
915Binary "or" returns the logical disjunction of the two surrounding
5a964f20
TC
916expressions. It's equivalent to || except for the very low precedence.
917This makes it useful for control flow
918
919 print FH \$data or die "Can't write to FH: \$!";
920
921This means that it short-circuits: i.e., the right expression is evaluated
922only if the left expression is false. Due to its precedence, you should
923probably avoid using this for assignment, only for control flow.
924
925 \$a = \$b or \$c; # bug: this is wrong
926 (\$a = \$b) or \$c; # really means this
927 \$a = \$b || \$c; # better written this way
928
19799a22 929However, when it's a list-context assignment and you're trying to use
5a964f20
TC
930"||" for control flow, you probably need "or" so that the assignment
931takes higher precedence.
932
933 @info = stat(\$file) || die; # oops, scalar sense of stat!
934 @info = stat(\$file) or die; # better, now @info gets its due
935
c963b151
BD
936Then again, you could always use parentheses.
937
a0d0e21e
LW
938Binary "xor" returns the exclusive-OR of the two surrounding expressions.
939It cannot short circuit, of course.
940
941=head2 C Operators Missing From Perl
d74e8afc
ITB
942X<operator, missing from perl> X<&> X<*>
943X<typecasting> X<(TYPE)>
a0d0e21e
LW
944
945Here is what C has that Perl doesn't:
946
947=over 8
948
949=item unary &
950
951Address-of operator. (But see the "\" operator for taking a reference.)
952
953=item unary *
954
54310121 955Dereference-address operator. (Perl's prefix dereferencing
a0d0e21e
LW
956operators are typed: \$, @, %, and &.)
957
958=item (TYPE)
959
19799a22 960Type-casting operator.
a0d0e21e
LW
961
962=back
963
5f05dabc 964=head2 Quote and Quote-like Operators
89d205f2 965X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
d74e8afc
ITB
966X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
967X<escape sequence> X<escape>
968
a0d0e21e
LW
969
970While we usually think of quotes as literal values, in Perl they
971function as operators, providing various kinds of interpolating and
972pattern matching capabilities. Perl provides customary quote characters
973for these behaviors, but also provides a way for you to choose your
974quote character for any of them. In the following table, a C<{}> represents
9f10b797 975any pair of delimiters you choose.
a0d0e21e 976
TP
977 Customary Generic Meaning Interpolates
978 '' q{} Literal no
979 "" qq{} Literal yes
af9219ee 980 `` qx{} Command yes*
2c268ad5 981 qw{} Word list no
af9219ee
MG
982 // m{} Pattern match yes*
983 qr{} Pattern yes*
984 s{}{} Substitution yes*
2c268ad5 985 tr{}{} Transliteration no (but see below)
7e3b091d 986 <<EOF here-doc yes*
a0d0e21e 987
af9219ee
MG
988 * unless the delimiter is ''.
989
87275199
GS
990Non-bracketing delimiters use the same character fore and aft, but the four
991sorts of brackets (round, angle, square, curly) will all nest, which means
9f10b797 992that
87275199 993
9f10b797 994 q{foo{bar}baz}
35f2feb0 995
9f10b797 996is the same as
87275199
GS
997
998 'foo{bar}baz'
999
1000Note, however, that this does not always work for quoting Perl code:
1001
1002 \$s = q{ if(\$a eq "}") ... }; # WRONG
1003
83df6a1d
JH
1004is a syntax error. The C<Text::Balanced> module (from CPAN, and
1005starting from Perl 5.8 part of the standard distribution) is able
1006to do this properly.
87275199 1007
19799a22 1008There can be whitespace between the operator and the quoting
fb73857a 1009characters, except when C<#> is being used as the quoting character.
19799a22
GS
1010C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
1011operator C<q> followed by a comment. Its argument will be taken
1012from the next line. This allows you to write:
fb73857a 1013
1014 s {foo} # Replace foo
1015 {bar} # with bar.
1016
904501ec
MG
1017The following escape sequences are available in constructs that interpolate
1018and in transliterations.
5691ca5f
KW
1019X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N> X<\N{}>
1020
1021 Sequence Note Description
40687185
KW
1022 \t tab (HT, TAB)
1023 \n newline (NL)
1024 \r return (CR)
1025 \f form feed (FF)
1026 \b backspace (BS)
1027 \a alarm (bell) (BEL)
1028 \e escape (ESC)
1029 \x{263a} [1] hex char (example: SMILEY)
1030 \x1b [2] narrow hex char (example: ESC)
1031 \N{name} [3] named Unicode character
1032 \N{U+263D} [4] Unicode character (example: FIRST QUARTER MOON)
1033 \c[ [5] control char (example: chr(27))
1034 \033 [6] octal char (example: ESC)
5691ca5f
KW
1035
1036=over 4
1037
1038=item [1]
1039
40687185
KW
1040The result is the character whose ordinal is the hexadecimal number between the
1041braces. If something other than a hexadecimal digit is encountered, it and
1042everything following it up to the closing brace are discarded, and if warnings
1043are enabled, a warning is raised. The leading digits that are hex then
1044comprise the entire number. If the first thing after the opening brace is not
1045a hex digit, the generated character is the NULL character. C<\x{}> is the
1046NULL character with no warning given.
1047
1048=item [2]
1049
1050The result is the character whose ordinal is the given two-digit hexadecimal
1051number. But, if I<H> is a hex digit and I<G> is not, then C<\xI<HG>...> is the
1052same as C<\x0I<HG>...>, and C<\xI<G>...> is the same thing as C<\x00I<G>...>.
1053In both cases, the result is two characters, and if warnings are enabled, a
1054misleading warning message is raised that I<G> is ignored, when in fact it is
1055used. Note that in the second case, the first character currently is a NULL.
1056
1057=item [3]
1058
1059For documentation of C<\N{name}>, see L<charnames>.
1060
1061=item [4]
1062
1063C<\N{U+I<wide hex char>}> means the Unicode character whose Unicode ordinal
1064number is I<wide hex char>.
1065
1066=item [5]
1067
5691ca5f
KW
1068The character following C<\c> is mapped to some other character as shown in the
1069table:
1070
1071 Sequence Value
1072 \c@ chr(0)
1073 \cA chr(1)
1074 \ca chr(1)
1075 \cB chr(2)
1076 \cb chr(2)
1077 ...
1078 \cZ chr(26)
1079 \cz chr(26)
1080 \c[ chr(27)
1081 \c] chr(29)
1082 \c^ chr(30)
1083 \c? chr(127)
1084
1085Also, C<\c\I<X>> yields C< chr(28) . "I<X>"> for any I<X>, but cannot come at the
1086end of a string, because the backslash would be parsed as escaping the end
1087quote.
1088
1089On ASCII platforms, the resulting characters from the list above are the
1090complete set of ASCII controls. This isn't the case on EBCDIC platforms; see
1091L<perlebcdic/OPERATOR DIFFERENCES> for the complete list of what these
1092sequences mean on both ASCII and EBCDIC platforms.
1093
1094Use of any other character following the "c" besides those listed above is
6ae9f32a
KW
1095discouraged, and may become deprecated or forbidden. What happens for those
1096other characters currently though, is that the value is derived by inverting
1097the 7th bit (0x40).
5691ca5f
KW
1098
1099To get platform independent controls, you can use C<\N{...}>.
1100
40687185
KW
1101=item [6]
1102
1103The result is the character whose ordinal is the given three digit octal
1104number. Some contexts allow 2 or even 1 digit, but any usage without exactly
1105three digits, the first being a zero, may give unintended results. (For
1106example, see L<perlrebackslash/Octal escapes>.) It is best therefore to use
1107this construct only for ordinals C<\077> and below, remembering to pad to the
1108left with zeros to make three digits. For larger ordinals, it's best to
1109convert to some other construct, such as to hex and use C<\x{}> instead.
1110
1111A backslash followed by a non-octal digit in a bracketed character class
1112(C<[\8]> or C<[\9]>) will be interpreted as a NULL character and the digit.
1113Having fewer than 3 digits may lead to a misleading warning message that says
1114that what follows is ignored. For example, C<"\128"> in the ASCII character set
1115is equivalent to the two characters C<"\n8">, but the warning C<Illegal octal
1116digit '8' ignored> will be thrown. To avoid this warning, make sure to pad
1117your octal number with C<0>s: C<"\0128">.
5691ca5f
KW
1118
1119=back
4c77eaa2 1120
e526e8bb
KW
1121B<NOTE>: Unlike C and other languages, Perl has no C<\v> escape sequence for
1122the vertical tab (VT - ASCII 11), but you may use C<\ck> or C<\x0b>. (C<\v>
1123does have meaning in regular expression patterns in Perl, see L<perlre>.)
1124
1125The following escape sequences are available in constructs that interpolate,
904501ec 1126but not in transliterations.
d74e8afc 1127X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q>
904501ec 1128
a0d0e21e
LW
1129 \l lowercase next char
1130 \u uppercase next char
1131 \L lowercase till \E
1132 \U uppercase till \E
1133 \E end case modification
1d2dff63 1134 \Q quote non-word characters till \E
a0d0e21e 1135
95cc3e0c
JH
1136If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
1137C<\u> and C<\U> is taken from the current locale. See L<perllocale>.
1138If Unicode (for example, C<\N{}> or wide hex characters of 0x100 or
1139beyond) is being used, the case map used by C<\l>, C<\L>, C<\u> and
e526e8bb 1140C<\U> is as defined by Unicode.
a034a98d 1141
5a964f20
TC
1142All systems use the virtual C<"\n"> to represent a line terminator,
1143called a "newline". There is no such thing as an unvarying, physical
19799a22 1144newline character. It is only an illusion that the operating system,
5a964f20
TC
1145device drivers, C libraries, and Perl all conspire to preserve. Not all
1146systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
1147on a Mac, these are reversed, and on systems without line terminator,
1148printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
1149you mean a "newline" for your system, but use the literal ASCII when you
1150need an exact character. For example, most networking protocols expect
2a380090 1151and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
1152and although they often accept just C<"\012">, they seldom tolerate just
1153C<"\015">. If you get in the habit of using C<"\n"> for networking,
1154you may be burned some day.
d74e8afc
ITB
1155X<newline> X<line terminator> X<eol> X<end of line>
1156X<\n> X<\r> X<\r\n>
5a964f20 1157
904501ec
MG
1158For constructs that do interpolate, variables beginning with "C<\$>"
1159or "C<@>" are interpolated. Subscripted variables such as C<\$a[3]> or
A
1160C<< \$href->{key}[0] >> are also interpolated, as are array and hash slices.
1161But method calls such as C<< \$obj->meth >> are not.
af9219ee
MG
1162
1163Interpolating an array or slice interpolates the elements in order,
1164separated by the value of C<\$">, so is equivalent to interpolating
6deea57f
TS
1165C<join \$", @array>. "Punctuation" arrays such as C<@*> are only
1166interpolated if the name is enclosed in braces C<@{*}>, but special
1167arrays C<@_>, C<@+>, and C<@-> are interpolated, even without braces.
af9219ee 1168
89d205f2
YO
1169You cannot include a literal C<\$> or C<@> within a C<\Q> sequence.
1170An unescaped C<\$> or C<@> interpolates the corresponding variable,
1d2dff63 1171while escaping will cause the literal string C<\\$> to be inserted.
89d205f2 1172You'll need to write something like C<m/\Quser\E\@\Qhost/>.
1d2dff63 1173
a0d0e21e
LW
1174Patterns are subject to an additional level of interpretation as a
1175regular expression. This is done as a second pass, after variables are
1176interpolated, so that regular expressions may be incorporated into the
1177pattern from the variables. If this is not what you want, use C<\Q> to
1178interpolate a variable literally.
1179
19799a22
GS
1180Apart from the behavior described above, Perl does not expand
1181multiple levels of interpolation. In particular, contrary to the
1182expectations of shell programmers, back-quotes do I<NOT> interpolate
1183within double quotes, nor do single quotes impede evaluation of
1184variables when used within double quotes.
a0d0e21e 1185
d74e8afc 1187X<operator, regexp>
cb1a09d0 1188
5f05dabc 1189Here are the quote-like operators that apply to pattern
cb1a09d0
1190matching and related activities.
1191
a0d0e21e
LW
1192=over 8
1193
87e95b7f 1194=item qr/STRING/msixpo
01c6f5f4 1195X<qr> X</i> X</m> X</o> X</s> X</x> X</p>
a0d0e21e 1196
87e95b7f
YO
1197This operator quotes (and possibly compiles) its I<STRING> as a regular
1198expression. I<STRING> is interpolated the same way as I<PATTERN>
1199in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1200is done. Returns a Perl value which may be used instead of the
64c5a566 1201corresponding C</STRING/msixpo> expression. The returned value is a
85dd5c8b 1202normalized version of the original pattern. It magically differs from
64c5a566 1203a string containing the same characters: C<ref(qr/x/)> returns "Regexp",
85dd5c8b 1204even though dereferencing the result returns undef.
a0d0e21e 1205
87e95b7f
YO
1206For example,
1207
1208 \$rex = qr/my.STRING/is;
85dd5c8b 1209 print \$rex; # prints (?si-xm:my.STRING)
87e95b7f
YO
1210 s/\$rex/foo/;
1211
1212is equivalent to
1213
1214 s/my.STRING/foo/is;
1215
1216The result may be used as a subpattern in a match:
1217
1218 \$re = qr/\$pattern/;
1219 \$string =~ /foo\${re}bar/; # can be interpolated in other patterns
1220 \$string =~ \$re; # or used standalone
1221 \$string =~ /\$re/; # or this way
1222
1223Since Perl may compile the pattern at the moment of execution of qr()
1224operator, using qr() may have speed advantages in some situations,
1225notably if the result of qr() is used standalone:
1226
1227 sub match {
1228 my \$patterns = shift;
1229 my @compiled = map qr/\$_/i, @\$patterns;
1230 grep {
1231 my \$success = 0;
1232 foreach my \$pat (@compiled) {
1233 \$success = 1, last if /\$pat/;
1234 }
1235 \$success;
1236 } @_;
5a964f20
TC
1237 }
1238
87e95b7f
YO
1239Precompilation of the pattern into an internal representation at
1240the moment of qr() avoids a need to recompile the pattern every
1241time a match C</\$pat/> is attempted. (Perl has many other internal
1242optimizations, but none would be triggered in the above example if
1243we did not use qr() operator.)
1244
1245Options are:
1246
1247 m Treat string as multiple lines.
1248 s Treat string as single line. (Make . match a newline)
1249 i Do case-insensitive pattern matching.
1250 x Use extended regular expressions.
1251 p When matching preserve a copy of the matched string so
1252 that \${^PREMATCH}, \${^MATCH}, \${^POSTMATCH} will be defined.
1253 o Compile pattern only once.
1254
1255If a precompiled pattern is embedded in a larger pattern then the effect
1256of 'msixp' will be propagated appropriately. The effect of the 'o'
1257modifier has is not propagated, being restricted to those patterns
1258explicitly using it.
1259
1260See L<perlre> for additional information on valid syntax for STRING, and
1261for a detailed look at the semantics of regular expressions.
a0d0e21e 1262
87e95b7f 1263=item m/PATTERN/msixpogc
89d205f2
YO
1264X<m> X<operator, match>
1265X<regexp, options> X<regexp> X<regex, options> X<regex>
01c6f5f4 1266X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c>
a0d0e21e 1267
87e95b7f 1268=item /PATTERN/msixpogc
a0d0e21e 1269
5a964f20 1270Searches a string for a pattern match, and in scalar context returns
19799a22
GS
1271true if it succeeds, false if it fails. If no string is specified
1272via the C<=~> or C<!~> operator, the \$_ string is searched. (The
1273string specified with C<=~> need not be an lvalue--it may be the
1274result of an expression evaluation, but remember the C<=~> binds
1276discussion of additional considerations that apply when C<use locale>
1277is in effect.
a0d0e21e 1278
01c6f5f4
RGS
1279Options are as described in C<qr//>; in addition, the following match
1280process modifiers are available:
a0d0e21e 1281
950b09ed
KW
1282 g Match globally, i.e., find all occurrences.
1283 c Do not reset search position on a failed match when /g is in effect.
a0d0e21e
LW
1284
1285If "/" is the delimiter then the initial C<m> is optional. With the C<m>
ed02a3bf 1286you can use any pair of non-whitespace characters
19799a22
GS
1287as delimiters. This is particularly useful for matching path names
1288that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
7bac28a0 1289the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
19799a22 1290If "'" is the delimiter, no interpolation is performed on the PATTERN.
ed02a3bf
DN
1291When using a character valid in an identifier, whitespace is required
1292after the C<m>.
a0d0e21e
LW
1293
1294PATTERN may contain variables, which will be interpolated (and the
f70b4f9c 1295pattern recompiled) every time the pattern search is evaluated, except
1f247705
GS
1296for when the delimiter is a single quote. (Note that C<\$(>, C<\$)>, and
1297C<\$|> are not interpolated because they look like end-of-string tests.)
f70b4f9c
AB
1298If you want such a pattern to be compiled only once, add a C</o> after
1299the trailing delimiter. This avoids expensive run-time recompilations,
1300and is useful when the value you are interpolating won't change over
1301the life of the script. However, mentioning C</o> constitutes a promise
1302that you won't change the variables in the pattern. If you change them,
a0d0e21e 1304
e9d89077
DN
1305=item The empty pattern //
1306
5a964f20 1307If the PATTERN evaluates to the empty string, the last
d65afb4b
HS
1308I<successfully> matched regular expression is used instead. In this
1309case, only the C<g> and C<c> flags on the empty pattern is honoured -
1310the other flags are taken from the original pattern. If no match has
1311previously succeeded, this will (silently) act instead as a genuine
1312empty pattern (which will always match).
a0d0e21e 1313
89d205f2
YO
1314Note that it's possible to confuse Perl into thinking C<//> (the empty
1315regex) is really C<//> (the defined-or operator). Perl is usually pretty
1317C<\$a///> (is that C<(\$a) / (//)> or C<\$a // />?) and C<print \$fh //>
1318(C<print \$fh(//> or C<print(\$fh //>?). In all of these examples, Perl
1319will assume you meant defined-or. If you meant the empty regex, just
1320use parentheses or spaces to disambiguate, or even prefix the empty
c963b151
BD
1321regex with an C<m> (so C<//> becomes C<m//>).
1322
e9d89077
DN
1323=item Matching in list context
1324
19799a22 1325If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1326list consisting of the subexpressions matched by the parentheses in the
f7e33566
GS
1327pattern, i.e., (C<\$1>, C<\$2>, C<\$3>...). (Note that here C<\$1> etc. are
1328also set, and that this differs from Perl 4's behavior.) When there are
1329no parentheses in the pattern, the return value is the list C<(1)> for
1330success. With or without parentheses, an empty list is returned upon
1331failure.
a0d0e21e
LW
1332
1333Examples:
1334
1335 open(TTY, '/dev/tty');
1336 <TTY> =~ /^y/i && foo(); # do foo if desired
1337
1338 if (/Version: *([0-9.]*)/) { \$version = \$1; }
1339
1340 next if m#^/usr/spool/uucp#;
1341
1342 # poor man's grep
1343 \$arg = shift;
1344 while (<>) {
1345 print if /\$arg/o; # compile only once
1346 }
1347
1348 if ((\$F1, \$F2, \$Etc) = (\$foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
1349
1350This last example splits \$foo into the first two words and the
5f05dabc 1351remainder of the line, and assigns those three fields to \$F1, \$F2, and
1352\$Etc. The conditional is true if any variables were assigned, i.e., if
a0d0e21e
LW
1353the pattern matched.
1354
19799a22
GS
1355The C</g> modifier specifies global pattern matching--that is,
1356matching as many times as possible within the string. How it behaves
1357depends on the context. In list context, it returns a list of the
1358substrings matched by any capturing parentheses in the regular
1359expression. If there are no parentheses, it returns a list of all
1360the matched strings, as if there were parentheses around the whole
1361pattern.
a0d0e21e 1362
7e86de3e 1363In scalar context, each execution of C<m//g> finds the next match,
19799a22 1364returning true if it matches, and false if there is no further match.
7e86de3e
MG
1365The position after the last match can be read or set using the pos()
1366function; see L<perlfunc/pos>. A failed match normally resets the
1367search position to the beginning of the string, but you can avoid that
1368by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
1369string also resets the search position.
c90c0ff4 1370
e9d89077
DN
1371=item \G assertion
1372
c90c0ff4 1373You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
1374zero-width assertion that matches the exact position where the previous
5d43e42d
DC
1375C<m//g>, if any, left off. Without the C</g> modifier, the C<\G> assertion
1376still anchors at pos(), but the match is of course only attempted once.
1377Using C<\G> without C</g> on a target string that has not previously had a
1378C</g> match applied to it is the same as using the C<\A> assertion to match
fe4b3f22
RGS
1379the beginning of the string. Note also that, currently, C<\G> is only
1380properly supported when anchored at the very beginning of the pattern.
c90c0ff4 1381
1382Examples:
a0d0e21e
LW
1383
1384 # list context
1385 (\$one,\$five,\$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1386
1387 # scalar context
5d43e42d 1388 \$/ = "";
19799a22
GS
1389 while (defined(\$paragraph = <>)) {
1390 while (\$paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
1391 \$sentences++;
a0d0e21e
LW
1392 }
1393 }
1394 print "\$sentences\n";
1395
c90c0ff4 1396 # using m//gc with \G
137443ea 1397 \$_ = "ppooqppqq";
44a8e56a 1398 while (\$i++ < 2) {
1399 print "1: '";
c90c0ff4 1400 print \$1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1401 print "2: '";
c90c0ff4 1402 print \$1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1403 print "3: '";
c90c0ff4 1404 print \$1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1405 }
5d43e42d 1406 print "Final: '\$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a 1407
1408The last example should print:
1409
1410 1: 'oo', pos=4
137443ea 1411 2: 'q', pos=5
44a8e56a 1412 3: 'pp', pos=7
1413 1: '', pos=7
137443ea 1414 2: 'q', pos=8
1415 3: '', pos=8
5d43e42d
DC
1416 Final: 'q', pos=8
1417
1418Notice that the final match matched C<q> instead of C<p>, which a match
1419without the C<\G> anchor would have done. Also note that the final match
ac036724 1420did not update C<pos>. C<pos> is only updated on a C</g> match. If the
5d43e42d
DC
1421final match did indeed match C<p>, it's a good bet that you're running an
1422older (pre-5.6.0) Perl.
44a8e56a 1423
c90c0ff4 1424A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1425combine several regexps like this to process a string part-by-part,
c90c0ff4 1426doing different actions depending on which regexp matched. Each
1427regexp tries to match where the previous one leaves off.
e7ea3e70 1428
3fe9a6f1 1429 \$_ = <<'EOL';
950b09ed 1430 \$url = URI::URL->new( "http://example.com/" ); die if \$url eq "xXx";
3fe9a6f1 1431 EOL
1432 LOOP:
e7ea3e70 1433 {
950b09ed
KW
1434 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
1435 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
1436 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
1437 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
1438 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
1439 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
1440 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
1441 print ". That's all!\n";
e7ea3e70
IZ
1442 }
1443
1444Here is the output (split into several lines):
1445
1446 line-noise lowercase line-noise lowercase UPPERCASE line-noise
1447 UPPERCASE line-noise lowercase line-noise lowercase line-noise
1448 lowercase lowercase line-noise lowercase lowercase line-noise
1449 MiXeD line-noise. That's all!
44a8e56a 1450
87e95b7f
YO
1451=item ?PATTERN?
1452X<?>
1453
1454This is just like the C</pattern/> search, except that it matches only
1455once between calls to the reset() operator. This is a useful
1456optimization when you want to see only the first occurrence of
1457something in each file of a set of files, for instance. Only C<??>
1458patterns local to the current package are reset.
1459
1460 while (<>) {
1461 if (?^\$?) {
1462 # blank line between header and body
1463 }
1464 } continue {
1465 reset if eof; # clear ?? status for next file
1466 }
1467
1468This usage is vaguely deprecated, which means it just might possibly
1469be removed in some distant future version of Perl, perhaps somewhere
1470around the year 2168.
1471
4f4d7508 1472=item s/PATTERN/REPLACEMENT/msixpogcer
87e95b7f 1473X<substitute> X<substitution> X<replace> X<regexp, replace>
4f4d7508 1474X<regexp, substitute> X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c> X</e> X</r>
87e95b7f
YO
1475
1476Searches a string for a pattern, and if found, replaces that pattern
1477with the replacement text and returns the number of substitutions
1478made. Otherwise it returns false (specifically, the empty string).
1479
4f4d7508
DC
1480If the C</r> (non-destructive) option is used then it will perform the
1481substitution on a copy of the string and return the copy whether or not a
1482substitution occurred. The original string will always remain unchanged in
c8dbf8cd 1483this case. The copy will always be a plain string, even if the input is an
4f4d7508
DC
1484object or a tied variable.
1485
87e95b7f
YO
1486If no string is specified via the C<=~> or C<!~> operator, the C<\$_>
1487variable is searched and modified. (The string specified with C<=~> must
1488be scalar variable, an array element, a hash element, or an assignment
1489to one of those, i.e., an lvalue.)
1490
1491If the delimiter chosen is a single quote, no interpolation is
1492done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1493PATTERN contains a \$ that looks like a variable rather than an
1494end-of-string test, the variable will be interpolated into the pattern
1495at run-time. If you want the pattern compiled only once the first time
1496the variable is interpolated, use the C</o> option. If the pattern
1497evaluates to the empty string, the last successfully executed regular
1498expression is used instead. See L<perlre> for further explanation on these.
1499See L<perllocale> for discussion of additional considerations that apply
1500when C<use locale> is in effect.
1501
1502Options are as with m// with the addition of the following replacement
1503specific options:
1504
1505 e Evaluate the right side as an expression.
4f4d7508
DC
1506 ee Evaluate the right side as a string then eval the result.
1507 r Return substitution and leave the original string untouched.
87e95b7f 1508
ed02a3bf
DN
1509Any non-whitespace delimiter may replace the slashes. Add space after
1510the C<s> when using a character allowed in identifiers. If single quotes
1511are used, no interpretation is done on the replacement string (the C</e>
1512modifier overrides this, however). Unlike Perl 4, Perl 5 treats backticks
1513as normal delimiters; the replacement text is not evaluated as a command.
1514If the PATTERN is delimited by bracketing quotes, the REPLACEMENT has
1515its own pair of quotes, which may or may not be bracketing quotes, e.g.,
87e95b7f
YO
1516C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
1517replacement portion to be treated as a full-fledged Perl expression
1518and evaluated right then and there. It is, however, syntax checked at
1519compile-time. A second C<e> modifier will cause the replacement portion
1520to be C<eval>ed before being run as a Perl expression.
1521
1522Examples:
1523
1524 s/\bgreen\b/mauve/g; # don't change wintergreen
1525
1526 \$path =~ s|/usr/bin|/usr/local/bin|;
1527
1529
1530 (\$foo = \$bar) =~ s/this/that/; # copy first, then change
4f4d7508
DC
1531 (\$foo = "\$bar") =~ s/this/that/; # convert to string, copy, then change
1532 \$foo = \$bar =~ s/this/that/r; # Same as above using /r
1533 \$foo = \$bar =~ s/this/that/r
1534 =~ s/that/the other/r; # Chained substitutes using /r
1535 @foo = map { s/this/that/r } @bar # /r is very useful in maps
87e95b7f
YO
1536
1537 \$count = (\$paragraph =~ s/Mister\b/Mr./g); # get change-count
1538
1539 \$_ = 'abc123xyz';
1540 s/\d+/\$&*2/e; # yields 'abc246xyz'
1541 s/\d+/sprintf("%5d",\$&)/e; # yields 'abc 246xyz'
1542 s/\w/\$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1543
1544 s/%(.)/\$percent{\$1}/g; # change percent escapes; no /e
1545 s/%(.)/\$percent{\$1} || \$&/ge; # expr now, so /e
1546 s/^=(\w+)/pod(\$1)/ge; # use function call
1547
4f4d7508
DC
1548 \$_ = 'abc123xyz';
1549 \$a = s/abc/def/r; # \$a is 'def123xyz' and
1550 # \$_ remains 'abc123xyz'.
1551
87e95b7f
YO
1552 # expand variables in \$_, but dynamics only, using
1553 # symbolic dereferencing
1554 s/\\$(\w+)/\${\$1}/g;
1555
1556 # Add one to the value of any numbers in the string
1557 s/(\d+)/1 + \$1/eg;
1558
1559 # This will expand any embedded scalar variable
1560 # (including lexicals) in \$_ : First \$1 is interpolated
1561 # to the variable name, and then evaluated
1562 s/(\\$\w+)/\$1/eeg;
1563
1564 # Delete (most) C comments.
1565 \$program =~ s {
1566 /\* # Match the opening delimiter.
1567 .*? # Match a minimal number of characters.
1568 \*/ # Match the closing delimiter.
1569 } []gsx;
1570
1571 s/^\s*(.*?)\s*\$/\$1/; # trim whitespace in \$_, expensively
1572
1573 for (\$variable) { # trim whitespace in \$variable, cheap
1574 s/^\s+//;
1575 s/\s+\$//;
1576 }
1577
1578 s/([^ ]*) *([^ ]*)/\$2 \$1/; # reverse 1st two fields
1579
1580Note the use of \$ instead of \ in the last example. Unlike
1581B<sed>, we use the \<I<digit>> form in only the left hand side.
1582Anywhere else it's \$<I<digit>>.
1583
1584Occasionally, you can't use just a C</g> to get all the changes
1585to occur that you might want. Here are two common cases:
1586
1587 # put commas in the right places in an integer
1588 1 while s/(\d)(\d\d\d)(?!\d)/\$1,\$2/g;
1589
1590 # expand tabs to 8-column spacing
1591 1 while s/\t+/' ' x (length(\$&)*8 - length(\$`)%8)/e;
1592
1593=back
1594
1596X<operator, quote-like>
1597
01c6f5f4
RGS
1598=over 4
1599
a0d0e21e 1600=item q/STRING/
5d44bfff 1601X<q> X<quote, single> X<'> X<''>
a0d0e21e 1602
5d44bfff 1603=item 'STRING'
a0d0e21e 1604
19799a22 1605A single-quoted, literal string. A backslash represents a backslash
68dc0745 1606unless followed by the delimiter or another backslash, in which case
1607the delimiter or backslash is interpolated.
a0d0e21e
LW
1608
1609 \$foo = q!I said, "You said, 'She said it.'"!;
1610 \$bar = q('This is it.');
68dc0745 1611 \$baz = '\n'; # a two-character string
a0d0e21e
LW
1612
1613=item qq/STRING/
d74e8afc 1614X<qq> X<quote, double> X<"> X<"">
a0d0e21e
LW
1615
1616=item "STRING"
1617
1618A double-quoted, interpolated string.
1619
1620 \$_ .= qq
1621 (*** The previous line contains the naughty word "\$1".\n)
19799a22 1622 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 1623 \$baz = "\n"; # a one-character string
a0d0e21e
LW
1624
1625=item qx/STRING/
d74e8afc 1626X<qx> X<`> X<``> X<backtick>
a0d0e21e
LW
1627
1628=item `STRING`
1629
43dd4d21
JH
1630A string which is (possibly) interpolated and then executed as a
1631system command with C</bin/sh> or its equivalent. Shell wildcards,
1632pipes, and redirections will be honored. The collected standard
1633output of the command is returned; standard error is unaffected. In
1634scalar context, it comes back as a single (potentially multi-line)
1635string, or undef if the command failed. In list context, returns a
1636list of lines (however you've defined lines with \$/ or
1637\$INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
5a964f20
TC
1638
1639Because backticks do not affect standard error, use shell file descriptor
1640syntax (assuming the shell supports this) if you care to address this.
1641To capture a command's STDERR and STDOUT together:
a0d0e21e 1642
5a964f20
TC
1643 \$output = `cmd 2>&1`;
1644
1645To capture a command's STDOUT but discard its STDERR:
1646
1647 \$output = `cmd 2>/dev/null`;
1648
1649To capture a command's STDERR but discard its STDOUT (ordering is
1650important here):
1651
1652 \$output = `cmd 2>&1 1>/dev/null`;
1653
1654To exchange a command's STDOUT and STDERR in order to capture the STDERR
1655but leave its STDOUT to come out the old STDERR:
1656
1657 \$output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1658
1659To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
1660to redirect them separately to files, and then read from those files
1661when the program is done:
5a964f20 1662
2359510d 1663 system("program args 1>program.stdout 2>program.stderr");
5a964f20 1664
30398227
SP
1665The STDIN filehandle used by the command is inherited from Perl's STDIN.
1666For example:
1667
1668 open BLAM, "blam" || die "Can't open: \$!";
1669 open STDIN, "<&BLAM";
1670 print `sort`;
1671
1672will print the sorted contents of the file "blam".
1673
5a964f20
TC
1674Using single-quote as a delimiter protects the command from Perl's
1675double-quote interpolation, passing it on to the shell instead:
1676
1677 \$perl_info = qx(ps \$\$); # that's Perl's \$\$
1678 \$shell_info = qx'ps \$\$'; # that's the new shell's \$\$
1679
19799a22 1680How that string gets evaluated is entirely subject to the command
5a964f20
TC
1681interpreter on your system. On most platforms, you will have to protect
1682shell metacharacters if you want them treated literally. This is in
1683practice difficult to do, as it's unclear how to escape which characters.
1684See L<perlsec> for a clean and safe example of a manual fork() and exec()
1685to emulate backticks safely.
a0d0e21e 1686
bb32b41a
GS
1687On some platforms (notably DOS-like ones), the shell may not be
1688capable of dealing with multiline commands, so putting newlines in
1689the string may not get you what you want. You may be able to evaluate
1690multiple commands in a single line by separating them with the command
1691separator character, if your shell supports that (e.g. C<;> on many Unix
1692shells; C<&> on the Windows NT C<cmd> shell).
1693
0f897271
GS
1694Beginning with v5.6.0, Perl will attempt to flush all files opened for
1695output before starting the child process, but this may not be supported
1696on some platforms (see L<perlport>). To be safe, you may need to set
1697C<\$|> (\$AUTOFLUSH in English) or call the C<autoflush()> method of
1698C<IO::Handle> on any open handles.
1699
bb32b41a
GS
1700Beware that some command shells may place restrictions on the length
1701of the command line. You must ensure your strings don't exceed this
1702limit after any necessary interpolations. See the platform-specific
1704
5a964f20
TC
1705Using this operator can lead to programs that are difficult to port,
1706because the shell commands called vary between systems, and may in
1707fact not be present at all. As one example, the C<type> command under
1708the POSIX shell is very different from the C<type> command under DOS.
1709That doesn't mean you should go out of your way to avoid backticks
1710when they're the right way to get something done. Perl was made to be
1711a glue language, and one of the things it glues together is commands.
1712Just understand what you're getting yourself into.
bb32b41a 1713
da87341d 1714See L</"I/O Operators"> for more discussion.
a0d0e21e 1715
945c54fd 1716=item qw/STRING/
d74e8afc 1717X<qw> X<quote, list> X<quote, words>
945c54fd
JH
1718
1719Evaluates to a list of the words extracted out of STRING, using embedded
1720whitespace as the word delimiters. It can be understood as being roughly
1721equivalent to:
1722
1723 split(' ', q/STRING/);
1724
efb1e162
CW
1725the differences being that it generates a real list at compile time, and
1726in scalar context it returns the last element in the list. So
945c54fd
JH
1727this expression:
1728
1729 qw(foo bar baz)
1730
1731is semantically equivalent to the list:
1732
1733 'foo', 'bar', 'baz'
1734
1735Some frequently seen examples:
1736
1737 use POSIX qw( setlocale localeconv )
1738 @EXPORT = qw( foo bar baz );
1739
1740A common mistake is to try to separate the words with comma or to
1741put comments into a multi-line C<qw>-string. For this reason, the
89d205f2 1742C<use warnings> pragma and the B<-w> switch (that is, the C<\$^W> variable)
945c54fd
JH
1743produces warnings if the STRING contains the "," or the "#" character.
1744
a0d0e21e 1745
6940069f 1746=item tr/SEARCHLIST/REPLACEMENTLIST/cds
d74e8afc 1747X<tr> X<y> X<transliterate> X</c> X</d> X</s>
a0d0e21e 1748
6940069f 1749=item y/SEARCHLIST/REPLACEMENTLIST/cds
a0d0e21e 1750
2c268ad5 1751Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
1752with the corresponding character in the replacement list. It returns
1753the number of characters replaced or deleted. If no string is
2c268ad5 1754specified via the =~ or !~ operator, the \$_ string is transliterated. (The
54310121 1755string specified with =~ must be a scalar variable, an array element, a
1756hash element, or an assignment to one of those, i.e., an lvalue.)
89d205f2 1758A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
2c268ad5 1759does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121 1760For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1761SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1762its own pair of quotes, which may or may not be bracketing quotes,
a0d0e21e 1764
cc255d5f 1765Note that C<tr> does B<not> do regular expression character classes
e0c83546 1766such as C<\d> or C<[:lower:]>. The C<tr> operator is not equivalent to
cc255d5f
JH
1767the tr(1) utility. If you want to map strings between lower/upper
1768cases, see L<perlfunc/lc> and L<perlfunc/uc>, and in general consider
1769using the C<s> operator if you need regular expressions.
1770
JH
1771Note also that the whole range idea is rather unportable between
1772character sets--and even within character sets they may cause results
1773you probably didn't expect. A sound principle is to use only ranges
1774that begin from and end at either alphabets of equal case (a-e, A-E),
1775or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1776character sets in full.
1777
a0d0e21e
LW
1778Options:
1779
1780 c Complement the SEARCHLIST.
1781 d Delete found but unreplaced characters.
1782 s Squash duplicate replaced characters.
1783
19799a22
GS
1784If the C</c> modifier is specified, the SEARCHLIST character set
1785is complemented. If the C</d> modifier is specified, any characters
1787(Note that this is slightly more flexible than the behavior of some
1788B<tr> programs, which delete anything they find in the SEARCHLIST,
1789period.) If the C</s> modifier is specified, sequences of characters
1790that were transliterated to the same character are squashed down
1791to a single instance of the character.
a0d0e21e
LW
1792
1793If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1794exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1795than the SEARCHLIST, the final character is replicated till it is long
5a964f20 1796enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
a0d0e21e
LW
1797This latter is useful for counting characters in a class or for
1798squashing character sequences in a class.
1799
1800Examples:
1801
1802 \$ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1803
1804 \$cnt = tr/*/*/; # count the stars in \$_
1805
1806 \$cnt = \$sky =~ tr/*/*/; # count the stars in \$sky
1807
1808 \$cnt = tr/0-9//; # count the digits in \$_
1809
1810 tr/a-zA-Z//s; # bookkeeper -> bokeper
1811
1812 (\$HOST = \$host) =~ tr/a-z/A-Z/;
1813
1814 tr/a-zA-Z/ /cs; # change non-alphas to single space
1815
1816 tr [\200-\377]
1817 [\000-\177]; # delete 8th bit
1818
19799a22
GS
1819If multiple transliterations are given for a character, only the
1820first one is used:
748a9306
LW
1821
1822 tr/AAA/XYZ/
1823
2c268ad5 1824will transliterate any A to X.
748a9306 1825
19799a22 1826Because the transliteration table is built at compile time, neither
a0d0e21e 1827the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
19799a22
GS
1828interpolation. That means that if you want to use variables, you
1829must use an eval():
a0d0e21e
LW
1830
1831 eval "tr/\$oldlist/\$newlist/";
1832 die \$@ if \$@;
1833
1834 eval "tr/\$oldlist/\$newlist/, 1" or die \$@;
1835
7e3b091d 1836=item <<EOF
d74e8afc 1837X<here-doc> X<heredoc> X<here-document> X<<< << >>>
7e3b091d
DA
1838
1839A line-oriented form of quoting is based on the shell "here-document"
1840syntax. Following a C<< << >> you specify a string to terminate
1841the quoted material, and all lines following the current line down to
89d205f2
YO
1842the terminating string are the value of the item.
1843
1844The terminating string may be either an identifier (a word), or some
1845quoted text. An unquoted identifier works like double quotes.
1846There may not be a space between the C<< << >> and the identifier,
1847unless the identifier is explicitly quoted. (If you put a space it
1848will be treated as a null identifier, which is valid, and matches the
1849first empty line.) The terminating string must appear by itself
1850(unquoted and with no surrounding whitespace) on the terminating line.
1851
1852If the terminating string is quoted, the type of quotes used determine
1853the treatment of the text.
1854
1855=over 4
1856
1857=item Double Quotes
1858
1859Double quotes indicate that the text will be interpolated using exactly
1860the same rules as normal double quoted strings.
7e3b091d
DA
1861
1862 print <<EOF;
1863 The price is \$Price.
1864 EOF
1865
1866 print << "EOF"; # same as above
1867 The price is \$Price.
1868 EOF
1869
89d205f2
YO
1870
1871=item Single Quotes
1872
1873Single quotes indicate the text is to be treated literally with no
1874interpolation of its content. This is similar to single quoted
1875strings except that backslashes have no special meaning, with C<\\>
1876being treated as two backslashes and not one as they would in every
1877other quoting construct.
1878
1879This is the only form of quoting in perl where there is no need
1880to worry about escaping content, something that code generators
1881can and do make good use of.
1882
1883=item Backticks
1884
1885The content of the here doc is treated just as it would be if the
1886string were embedded in backticks. Thus the content is interpolated
1887as though it were double quoted and then executed via the shell, with
1888the results of the execution returned.
1889
1890 print << `EOC`; # execute command and get results
7e3b091d 1891 echo hi there
7e3b091d
DA
1892 EOC
1893
89d205f2
YO
1894=back
1895
1896It is possible to stack multiple here-docs in a row:
1897
7e3b091d
DA
1898 print <<"foo", <<"bar"; # you can stack them
1899 I said foo.
1900 foo
1901 I said bar.
1902 bar
1903
1904 myfunc(<< "THIS", 23, <<'THAT');
1905 Here's a line
1906 or two.
1907 THIS
1908 and here's another.
1909 THAT
1910
1911Just don't forget that you have to put a semicolon on the end
1912to finish the statement, as Perl doesn't know you're not going to
1913try to do this:
1914
1915 print <<ABC
1916 179231
1917 ABC
1918 + 20;
1919
872d7e53
TS
1920If you want to remove the line terminator from your here-docs,
1921use C<chomp()>.
1922
1923 chomp(\$string = <<'END');
1924 This is a string.
1925 END
1926
1927If you want your here-docs to be indented with the rest of the code,
1928you'll need to remove leading whitespace from each line manually:
7e3b091d
DA
1929
1930 (\$quote = <<'FINIS') =~ s/^\s+//gm;
89d205f2 1931 The Road goes ever on and on,
7e3b091d
DA
1932 down from the door where it began.
1933 FINIS
1934
1935If you use a here-doc within a delimited construct, such as in C<s///eg>,
1936the quoted material must come on the lines following the final delimiter.
1938
1939 s/this/<<E . 'that'
1940 the other
1941 E
1942 . 'more '/eg;
1943
1944you have to write
1945
89d205f2
YO
1946 s/this/<<E . 'that'
1947 . 'more '/eg;
1948 the other
1949 E
7e3b091d
DA
1950
1951If the terminating identifier is on the last line of the program, you
1952must be sure there is a newline after it; otherwise, Perl will give the
1953warning B<Can't find string terminator "END" anywhere before EOF...>.
1954
89d205f2 1955Additionally, the quoting rules for the end of string identifier are not
ac036724 1956related to Perl's quoting rules. C<q()>, C<qq()>, and the like are not
89d205f2
YO
1957supported in place of C<''> and C<"">, and the only interpolation is for
1958backslashing the quoting character:
7e3b091d
DA
1959
1960 print << "abc\"def";
1961 testing...
1962 abc"def
1963
1964Finally, quoted strings cannot span multiple lines. The general rule is
1965that the identifier must be a string literal. Stick with that, and you
1966should be safe.
1967
a0d0e21e
LW
1968=back
1969
75e14d17 1970=head2 Gory details of parsing quoted constructs
d74e8afc 1971X<quote, gory details>
75e14d17 1972
19799a22
GS
1973When presented with something that might have several different
1974interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
1975principle to pick the most probable interpretation. This strategy
1976is so successful that Perl programmers often do not suspect the
1977ambivalence of what they write. But from time to time, Perl's
1978notions differ substantially from what the author honestly meant.
1979
1980This section hopes to clarify how Perl handles quoted constructs.
1981Although the most common reason to learn this is to unravel labyrinthine
1982regular expressions, because the initial steps of parsing are the
1983same for all quoting operators, they are all discussed together.
1984
1985The most important Perl parsing rule is the first one discussed
1986below: when processing a quoted construct, Perl first finds the end
1987of that construct, then interprets its contents. If you understand
1988this rule, you may skip the rest of this section on the first
1990expectations much less frequently than this first one.
1991
1992Some passes discussed below are performed concurrently, but because
1993their results are the same, we consider them individually. For different
1994quoting constructs, Perl performs different numbers of passes, from
6deea57f 1995one to four, but these passes are always performed in the same order.
75e14d17 1996
13a2d996 1997=over 4
75e14d17
IZ
1998
1999=item Finding the end
2000
6deea57f
TS
2001The first pass is finding the end of the quoted construct, where
2002the information about the delimiters is used in parsing.
2003During this search, text between the starting and ending delimiters
2004is copied to a safe location. The text copied gets delimiter-independent.
2005
2006If the construct is a here-doc, the ending delimiter is a line
2007that has a terminating string as the content. Therefore C<<<EOF> is
2008terminated by C<EOF> immediately followed by C<"\n"> and starting
2009from the first column of the terminating line.
2010When searching for the terminating line of a here-doc, nothing
2011is skipped. In other words, lines after the here-doc syntax
2012are compared with the terminating string line by line.
2013
2014For the constructs except here-docs, single characters are used as starting
2015and ending delimiters. If the starting delimiter is an opening punctuation
2016(that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
2017corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
2018If the starting delimiter is an unpaired character like C</> or a closing
2019punctuation, the ending delimiter is same as the starting delimiter.
2020Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
2021C<qq[]> and C<qq]]> constructs.
2022
2023When searching for single-character delimiters, escaped delimiters
2024and C<\\> are skipped. For example, while searching for terminating C</>,
2025combinations of C<\\> and C<\/> are skipped. If the delimiters are
2026bracketing, nested pairs are also skipped. For example, while searching
2027for closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
2028and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
2029However, when backslashes are used as the delimiters (like C<qq\\> and
2030C<tr\\\>), nothing is skipped.
2031During the search for the end, backslashes that escape delimiters
2032are removed (exactly speaking, they are not copied to the safe location).
75e14d17 2033
19799a22
GS
2034For constructs with three-part delimiters (C<s///>, C<y///>, and
2035C<tr///>), the search is repeated once more.
6deea57f
TS
2036If the first delimiter is not an opening punctuation, three delimiters must
2037be same such as C<s!!!> and C<tr)))>, in which case the second delimiter
2038terminates the left part and starts the right part at once.
2039If the left part is delimited by bracketing punctuations (that is C<()>,
2040C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
2041delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespaces
2042and comments are allowed between both parts, though the comment must follow
2043at least one whitespace; otherwise a character expected as the start of
2044the comment may be regarded as the starting delimiter of the right part.
75e14d17 2045
19799a22
GS
2046During this search no attention is paid to the semantics of the construct.
2047Thus:
75e14d17
IZ
2048
2049 "\$hash{"\$foo/\$bar"}"
2050
2a94b7ce 2051or:
75e14d17 2052
89d205f2 2053 m/
2a94b7ce 2054 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
2055 /x
2056
19799a22
GS
2057do not form legal quoted expressions. The quoted part ends on the
2058first C<"> and C</>, and the rest happens to be a syntax error.
2059Because the slash that terminated C<m//> was followed by a C<SPACE>,
2060the example above is not C<m//x>, but rather C<m//> with no C</x>
2061modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 2062
89d205f2
YO
2063Also no attention is paid to C<\c\> (multichar control char syntax) during
2064this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
2065of C<\/>, and the following C</> is not recognized as a delimiter.
0d594e51
TS
2066Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
2067
75e14d17 2068=item Interpolation
d74e8afc 2069X<interpolation>
75e14d17 2070
19799a22 2071The next step is interpolation in the text obtained, which is now
89d205f2 2072delimiter-independent. There are multiple cases.
75e14d17 2073
13a2d996 2074=over 4
75e14d17 2075
89d205f2 2076=item C<<<'EOF'>
75e14d17
IZ
2077
2078No interpolation is performed.
6deea57f
TS
2079Note that the combination C<\\> is left intact, since escaped delimiters
2080are not available for here-docs.
75e14d17 2081
6deea57f 2082=item C<m''>, the pattern of C<s'''>
89d205f2 2083
6deea57f
TS
2084No interpolation is performed at this stage.
2085Any backslashed sequences including C<\\> are treated at the stage
2086to L</"parsing regular expressions">.
89d205f2 2087
6deea57f 2088=item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
75e14d17 2089
89d205f2 2090The only interpolation is removal of C<\> from pairs of C<\\>.
6deea57f
TS
2091Therefore C<-> in C<tr'''> and C<y'''> is treated literally
2092as a hyphen and no character range is available.
2093C<\1> in the replacement of C<s'''> does not work as C<\$1>.
89d205f2
YO
2094
2095=item C<tr///>, C<y///>
2096
6deea57f
TS
2097No variable interpolation occurs. String modifying combinations for
2098case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
2099The other escape sequences such as C<\200> and C<\t> and backslashed
2100characters such as C<\\> and C<\-> are converted to appropriate literals.
89d205f2
YO
2101The character C<-> is treated specially and therefore C<\-> is treated
2102as a literal C<->.
75e14d17 2103
89d205f2 2104=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
75e14d17 2105
19799a22
GS
2106C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
2107converted to corresponding Perl constructs. Thus, C<"\$foo\Qbaz\$bar">
2108is converted to C<\$foo . (quotemeta("baz" . \$bar))> internally.
6deea57f
TS
2109The other escape sequences such as C<\200> and C<\t> and backslashed
2110characters such as C<\\> and C<\-> are replaced with appropriate
2111expansions.
2a94b7ce 2112
19799a22
GS
2113Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
2114is interpolated in the usual way. Something like C<"\Q\\E"> has
2115no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
2116result is the same as for C<"\\\\E">. As a general rule, backslashes
2117between C<\Q> and C<\E> may lead to counterintuitive results. So,
2118C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
2119as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
2120
2121 \$str = '\t';
2122 return "\Q\$str";
2123
2124may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
2125
19799a22 2126Interpolated scalars and arrays are converted internally to the C<join> and
92d29cee 2127C<.> catenation operations. Thus, C<"\$foo XXX '@arr'"> becomes:
75e14d17 2128
19799a22 2129 \$foo . " XXX '" . (join \$", @arr) . "'";
75e14d17 2130
19799a22 2131All operations above are performed simultaneously, left to right.
75e14d17 2132
19799a22
GS
2133Because the result of C<"\Q STRING \E"> has all metacharacters
2134quoted, there is no way to insert a literal C<\$> or C<@> inside a
2135C<\Q\E> pair. If protected by C<\>, C<\$> will be quoted to became
2136C<"\\\\$">; if not, it is interpreted as the start of an interpolated
2137scalar.
75e14d17 2138
19799a22 2139Note also that the interpolation code needs to make a decision on
89d205f2 2140where the interpolated scalar ends. For instance, whether
35f2feb0 2141C<< "a \$b -> {c}" >> really means:
75e14d17
IZ
2142
2143 "a " . \$b . " -> {c}";
2144
2a94b7ce 2145or:
75e14d17
IZ
2146
2147 "a " . \$b -> {c};
2148
19799a22
GS
2149Most of the time, the longest possible text that does not include
2150spaces between components and which contains matching braces or
2151brackets. because the outcome may be determined by voting based
2152on heuristic estimators, the result is not strictly predictable.
2153Fortunately, it's usually correct for ambiguous cases.
75e14d17 2154
6deea57f 2155=item the replacement of C<s///>
75e14d17 2156
19799a22 2157Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
6deea57f
TS
2158happens as with C<qq//> constructs.
2159
2160It is at this step that C<\1> is begrudgingly converted to C<\$1> in
2161the replacement text of C<s///>, in order to correct the incorrigible
2162I<sed> hackers who haven't picked up the saner idiom yet. A warning
2163is emitted if the C<use warnings> pragma or the B<-w> command-line flag
2164(that is, the C<\$^W> variable) was set.
2165
2166=item C<RE> in C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
2167
cc74c5bd
TS
2168Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\E>,
2169and interpolation happens (almost) as with C<qq//> constructs.
2170
5d03b57c
KW
2171Processing of C<\N{...}> is also done here, and compiled into an intermediate
2172form for the regex compiler. (This is because, as mentioned below, the regex
2173compilation may be done at execution time, and C<\N{...}> is a compile-time
2174construct.)
2175
cc74c5bd
TS
2176However any other combinations of C<\> followed by a character
2177are not substituted but only skipped, in order to parse them
2178as regular expressions at the following step.
6deea57f 2179As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
1749ea0d 2180treated as an array symbol (for example C<@foo>),
6deea57f 2181even though the same text in C<qq//> gives interpolation of C<\c@>.
6deea57f
TS
2182
2183Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
19799a22
GS
2184a C<#>-comment in a C<//x>-regular expression, no processing is
2185performed whatsoever. This is the first step at which the presence
2186of the C<//x> modifier is relevant.
2187
1749ea0d
TS
2188Interpolation in patterns has several quirks: C<\$|>, C<\$(>, C<\$)>, C<@+>
2189and C<@-> are not interpolated, and constructs C<\$var[SOMETHING]> are
2190voted (by several different estimators) to be either an array element
2191or C<\$var> followed by an RE alternative. This is where the notation
19799a22
GS
2192C<\${arr[\$bar]}> comes handy: C</\${arr[0-9]}/> is interpreted as
2193array element C<-9>, not as a regular expression from the variable
2194C<\$arr> followed by a digit, which would be the interpretation of
2195C</\$arr[0-9]/>. Since voting among different estimators may occur,
2196the result is not predictable.
2197
19799a22
GS
2198The lack of processing of C<\\> creates specific restrictions on
2199the post-processed text. If the delimiter is C</>, one cannot get
2200the combination C<\/> into the result of this step. C</> will
2201finish the regular expression, C<\/> will be stripped to C</> on
2202the previous step, and C<\\/> will be left as is. Because C</> is
2203equivalent to C<\/> inside a regular expression, this does not
2204matter unless the delimiter happens to be character special to the
2205RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
2206alphanumeric char, as in:
2a94b7ce
IZ
2207
2208 m m ^ a \s* b mmx;
2209
19799a22 2210In the RE above, which is intentionally obfuscated for illustration, the
6deea57f 2211delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
89d205f2 2212RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
19799a22
GS
2213reason you're encouraged to restrict your delimiters to non-alphanumeric,
2214non-whitespace choices.
75e14d17
IZ
2215
2216=back
2217
19799a22 2218This step is the last one for all constructs except regular expressions,
75e14d17
IZ
2219which are processed further.
2220
6deea57f
TS
2221=item parsing regular expressions
2222X<regexp, parse>
75e14d17 2223
19799a22 2224Previous steps were performed during the compilation of Perl code,
ac036724 2225but this one happens at run time, although it may be optimized to
19799a22 2226be calculated at compile time if appropriate. After preprocessing
6deea57f 2227described above, and possibly after evaluation if concatenation,
19799a22
GS
2228joining, casing translation, or metaquoting are involved, the
2229resulting I<string> is passed to the RE engine for compilation.
2230
2231Whatever happens in the RE engine might be better discussed in L<perlre>,
2232but for the sake of continuity, we shall do so here.
2233
2234This is another step where the presence of the C<//x> modifier is
2235relevant. The RE engine scans the string from left to right and
2236converts it to a finite automaton.
2237
2238Backslashed characters are either replaced with corresponding
2239literal strings (as with C<\{>), or else they generate special nodes
2240in the finite automaton (as with C<\b>). Characters special to the
2241RE engine (such as C<|>) generate corresponding nodes or groups of
2242nodes. C<(?#...)> comments are ignored. All the rest is either
2243converted to literal strings to match, or else is ignored (as is
2244whitespace and C<#>-style comments if C<//x> is present).
2245
2246Parsing of the bracketed character class construct, C<[...]>, is
2247rather different than the rule used for the rest of the pattern.
2248The terminator of this construct is found using the same rules as
2249for finding the terminator of a C<{}>-delimited construct, the only
2250exception being that C<]> immediately following C<[> is treated as
2251though preceded by a backslash. Similarly, the terminator of
2252C<(?{...})> is found using the same rules as for finding the
2253terminator of a C<{}>-delimited construct.
2254
2255It is possible to inspect both the string given to RE engine and the
2256resulting finite automaton. See the arguments C<debug>/C<debugcolor>
2257in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 2258switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
2259
2260=item Optimization of regular expressions
d74e8afc 2261X<regexp, optimization>
75e14d17 2262
7522fed5 2263This step is listed for completeness only. Since it does not change
75e14d17 2264semantics, details of this step are not documented and are subject
19799a22
GS
2265to change without notice. This step is performed over the finite
2266automaton that was generated during the previous pass.
2a94b7ce 2267
19799a22
GS
2268It is at this stage that C<split()> silently optimizes C</^/> to
2269mean C</^/m>.
75e14d17
IZ
2270
2271=back
2272
d74e8afc
ITB
2274X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
2275X<< <> >> X<@ARGV>
a0d0e21e 2276
54310121 2277There are several I/O operators you should know about.
7b8d334a 2279A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
2280double-quote interpolation. It is then interpreted as an external
2281command, and the output of that command is the value of the
e9c56f9b
JH
2282backtick string, like in a shell. In scalar context, a single string
2283consisting of all output is returned. In list context, a list of
2284values is returned, one per line of output. (You can set C<\$/> to use
2285a different line terminator.) The command is executed each time the
2286pseudo-literal is evaluated. The status value of the command is
2287returned in C<\$?> (see L<perlvar> for the interpretation of C<\$?>).
2288Unlike in B<csh>, no translation is done on the return data--newlines
2289remain newlines. Unlike in any of the shells, single quotes do not
2290hide variable names in the command from interpretation. To pass a
2291literal dollar-sign through to the shell you need to hide it with a
2292backslash. The generalized form of backticks is C<qx//>. (Because
2293backticks always undergo shell expansion as well, see L<perlsec> for
2294security concerns.)
d74e8afc 2295X<qx> X<`> X<``> X<backtick> X<glob>
19799a22
GS
2296
2297In scalar context, evaluating a filehandle in angle brackets yields
2298the next line from that file (the newline, if any, included), or
2299C<undef> at end-of-file or on error. When C<\$/> is set to C<undef>
2300(sometimes known as file-slurp mode) and the file is empty, it
2301returns C<''> the first time, followed by C<undef> subsequently.
2302
2303Ordinarily you must assign the returned value to a variable, but
2304there is one situation where an automatic assignment happens. If
2305and only if the input symbol is the only thing inside the conditional
2306of a C<while> statement (even if disguised as a C<for(;;)> loop),
2307the value is automatically assigned to the global variable \$_,
2308destroying whatever was there previously. (This may seem like an
2309odd thing to you, but you'll use the construct in almost every Perl
17b829fa 2310script you write.) The \$_ variable is not implicitly localized.
19799a22
GS
2311You'll have to put a C<local \$_;> before the loop if you want that
2312to happen.
2313
2314The following lines are equivalent:
a0d0e21e 2315
748a9306 2316 while (defined(\$_ = <STDIN>)) { print; }
7b8d334a 2317 while (\$_ = <STDIN>) { print; }
a0d0e21e
LW
2318 while (<STDIN>) { print; }
2319 for (;<STDIN>;) { print; }
748a9306 2320 print while defined(\$_ = <STDIN>);
7b8d334a 2321 print while (\$_ = <STDIN>);
a0d0e21e
LW
2322 print while <STDIN>;
2323
19799a22 2324This also behaves similarly, but avoids \$_ :
7b8d334a 2325
89d205f2 2326 while (my \$line = <STDIN>) { print \$line }
7b8d334a 2327
19799a22
GS
2328In these loop constructs, the assigned value (whether assignment
2329is automatic or explicit) is then tested to see whether it is
2330defined. The defined test avoids problems where line has a string
2331value that would be treated as false by Perl, for example a "" or
2332a "0" with no trailing newline. If you really mean for such values
2333to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
2334
2335 while ((\$_ = <STDIN>) ne '0') { ... }
2336 while (<STDIN>) { last unless \$_; ... }
2337
5ef4d93e 2338In other boolean contexts, C<< <filehandle> >> without an
2339explicit C<defined> test or comparison elicits a warning if the
9f1b1f2d 2340C<use warnings> pragma or the B<-w>
19799a22 2341command-line switch (the C<\$^W> variable) is in effect.
7b8d334a 2342
5f05dabc 2343The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
2344filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
2345in packages, where they would be interpreted as local identifiers
2346rather than global.) Additional filehandles may be created with
2347the open() function, amongst others. See L<perlopentut> and
2348L<perlfunc/open> for details on this.
d74e8afc 2349X<stdin> X<stdout> X<sterr>
a0d0e21e 2350
35f2feb0 2351If a <FILEHANDLE> is used in a context that is looking for
19799a22
GS
2352a list, a list comprising all input lines is returned, one line per
2353list element. It's easy to grow to a rather large data space this
2354way, so use with care.
a0d0e21e 2355
35f2feb0 2356<FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
35f2feb0
GS
2359The null filehandle <> is special: it can be used to emulate the
2360behavior of B<sed> and B<awk>. Input from <> comes either from
a0d0e21e 2361standard input, or from each file listed on the command line. Here's
35f2feb0 2362how it works: the first time <> is evaluated, the @ARGV array is
5a964f20 2363checked, and if it is empty, C<\$ARGV[0]> is set to "-", which when opened
a0d0e21e
LW
2364gives you standard input. The @ARGV array is then processed as a list
2365of filenames. The loop
2366
2367 while (<>) {
2368 ... # code for each line
2369 }
2370
2371is equivalent to the following Perl-like pseudo code:
2372
3e3baf6d 2373 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
2374 while (\$ARGV = shift) {
2375 open(ARGV, \$ARGV);
2376 while (<ARGV>) {
2377 ... # code for each line
2378 }
2379 }
2380
19799a22
GS
2381except that it isn't so cumbersome to say, and will actually work.
2382It really does shift the @ARGV array and put the current filename
2383into the \$ARGV variable. It also uses filehandle I<ARGV>
ac036724 2384internally. <> is just a synonym for <ARGV>, which
19799a22 2385is magical. (The pseudo code above doesn't work because it treats
35f2feb0 2386<ARGV> as non-magical.)
a0d0e21e 2387
48ab5743
ML
2388Since the null filehandle uses the two argument form of L<perlfunc/open>
2389it interprets special characters, so if you have a script like this:
2390
2391 while (<>) {
2392 print;
2393 }
2394
2395and call it with C<perl dangerous.pl 'rm -rfv *|'>, it actually opens a
2396pipe, executes the C<rm> command and reads C<rm>'s output from that pipe.
2397If you want all items in C<@ARGV> to be interpreted as file names, you
2398can use the module C<ARGV::readonly> from CPAN.
2399
35f2feb0 2400You can modify @ARGV before the first <> as long as the array ends up
a0d0e21e 2401containing the list of filenames you really want. Line numbers (C<\$.>)
19799a22
GS
2402continue as though the input were one big happy file. See the example
2403in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20 2404
89d205f2 2405If you want to set @ARGV to your own list of files, go right ahead.
5a964f20
TC
2406This sets @ARGV to all plain text files if no @ARGV was given:
2407
2408 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 2409
5a964f20
TC
2410You can even set them to pipe commands. For example, this automatically
2411filters compressed arguments through B<gzip>:
2412
2413 @ARGV = map { /\.(gz|Z)\$/ ? "gzip -dc < \$_ |" : \$_ } @ARGV;
2414
2415If you want to pass switches into your script, you can use one of the
a0d0e21e
LW
2416Getopts modules or put a loop on the front like this:
2417
2418 while (\$_ = \$ARGV[0], /^-/) {
2419 shift;
2420 last if /^--\$/;
2421 if (/^-D(.*)/) { \$debug = \$1 }
2422 if (/^-v/) { \$verbose++ }
5a964f20 2423 # ... # other switches
a0d0e21e 2424 }
5a964f20 2425
a0d0e21e 2426 while (<>) {
5a964f20 2427 # ... # code for each line
a0d0e21e
LW
2428 }
2429
89d205f2
YO
2430The <> symbol will return C<undef> for end-of-file only once.
2431If you call it again after this, it will assume you are processing another
19799a22 2432@ARGV list, and if you haven't set @ARGV, will read input from STDIN.
a0d0e21e 2433
b159ebd3 2434If what the angle brackets contain is a simple scalar variable (e.g.,
35f2feb0 2435<\$foo>), then that variable contains the name of the
19799a22
GS
2436filehandle to input from, or its typeglob, or a reference to the
2437same. For example:
cb1a09d0
2438
2439 \$fh = \*STDIN;
2440 \$line = <\$fh>;
a0d0e21e 2441
5a964f20
TC
2442If what's within the angle brackets is neither a filehandle nor a simple
2443scalar variable containing a filehandle name, typeglob, or typeglob
2444reference, it is interpreted as a filename pattern to be globbed, and
2445either a list of filenames or the next filename in the list is returned,
19799a22 2446depending on context. This distinction is determined on syntactic
35f2feb0
GS
2447grounds alone. That means C<< <\$x> >> is always a readline() from
2448an indirect handle, but C<< <\$hash{key}> >> is always a glob().
5a964f20 2449That's because \$x is a simple scalar variable, but C<\$hash{key}> is
ef191992
YST
2450not--it's a hash element. Even C<< <\$x > >> (note the extra space)
2451is treated as C<glob("\$x ")>, not C<readline(\$x)>.
5a964f20
TC
2452
2453One level of double-quote interpretation is done first, but you can't
35f2feb0 2454say C<< <\$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
2455in the previous paragraph. (In older versions of Perl, programmers
2456would insert curly brackets to force interpretation as a filename glob:
35f2feb0 2457C<< <\${foo}> >>. These days, it's considered cleaner to call the
5a964f20 2458internal function directly as C<glob(\$foo)>, which is probably the right
19799a22 2459way to have done it in the first place.) For example:
a0d0e21e
LW
2460
2461 while (<*.c>) {
2462 chmod 0644, \$_;
2463 }
2464
3a4b19e4 2465is roughly equivalent to:
a0d0e21e
LW
2466
2467 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
2468 while (<FOO>) {
5b3eff12 2469 chomp;
a0d0e21e
LW
2470 chmod 0644, \$_;
2471 }
2472
3a4b19e4
GS
2473except that the globbing is actually done internally using the standard
2474C<File::Glob> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
2475
2476 chmod 0644, <*.c>;
2477
19799a22
GS
2478A (file)glob evaluates its (embedded) argument only when it is
2479starting a new list. All values must be read before it will start
2480over. In list context, this isn't important because you automatically
2481get them all anyway. However, in scalar context the operator returns
069e01df 2482the next value each time it's called, or C<undef> when the list has
19799a22
GS
2483run out. As with filehandle reads, an automatic C<defined> is
2484generated when the glob occurs in the test part of a C<while>,
2485because legal glob returns (e.g. a file called F<0>) would otherwise
2486terminate the loop. Again, C<undef> is returned only once. So if
2487you're expecting a single value from a glob, it is much better to
2488say
4633a7c4
LW
2489
2490 (\$file) = <blurch*>;
2491
2492than
2493
2494 \$file = <blurch*>;
2495
2496because the latter will alternate between returning a filename and
19799a22 2497returning false.
4633a7c4 2498
b159ebd3 2499If you're trying to do variable interpolation, it's definitely better
4633a7c4 2500to use the glob() function, because the older notation can cause people
e37d713d 2501to become confused with the indirect filehandle notation.
4633a7c4
LW
2502
2503 @files = glob("\$dir/*.[ch]");
2504 @files = glob(\$files[\$i]);
2505
d74e8afc 2507X<constant folding> X<folding>
a0d0e21e
LW
2508
2509Like C, Perl does a certain amount of expression evaluation at
19799a22 2510compile time whenever it determines that all arguments to an
a0d0e21e
LW
2511operator are static and have no side effects. In particular, string
2512concatenation happens at compile time between literals that don't do
19799a22 2513variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
2514compile time. You can say
2515
2516 'Now is the time for all' . "\n" .
2517 'good men to come to.'
2518
54310121 2519and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
2520you say
2521
2522 foreach \$file (@filenames) {
5a964f20 2523 if (-s \$file > 5 + 100 * 2**16) { }
54310121 2524 }
a0d0e21e 2525
19799a22
GS
2526the compiler will precompute the number which that expression
2527represents so that the interpreter won't have to.
a0d0e21e 2528
d74e8afc 2530X<no-op> X<nop>
fd1abbef
DN
2531
2532Perl doesn't officially have a no-op operator, but the bare constants
2533C<0> and C<1> are special-cased to not produce a warning in a void
2534context, so you can for example safely do
2535
2536 1 while foo();
2537
d74e8afc 2539X<operator, bitwise, string>
TP
2540
2541Bitstrings of any size may be manipulated by the bitwise operators
2542(C<~ | & ^>).
2543
19799a22
GS
2544If the operands to a binary bitwise op are strings of different
2545sizes, B<|> and B<^> ops act as though the shorter operand had
2546additional zero bits on the right, while the B<&> op acts as though
2547the longer operand were truncated to the length of the shorter.
2548The granularity for such extension or truncation is one or more
2549bytes.
89d205f2 2551 # ASCII-based examples
TP
2552 print "j p \n" ^ " a h"; # prints "JAPH\n"
2553 print "JA" | " ph\n"; # prints "japh\n"
2554 print "japh\nJunk" & '_____'; # prints "JAPH\n";
2555 print 'p N\$' ^ " E<H\n"; # prints "Perl\n";
2556
19799a22 2557If you are intending to manipulate bitstrings, be certain that
2c268ad5 2558you're supplying bitstrings: If an operand is a number, that will imply
19799a22 2559a B<numeric> bitwise operation. You may explicitly show which type of
TP
2560operation you intend by using C<""> or C<0+>, as in the examples below.
2561
4358a253
SS
2562 \$foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
2563 \$foo = '150' | 105; # yields 255
TP
2564 \$foo = 150 | '105'; # yields 255
2565 \$foo = '150' | '105'; # yields string '155' (under ASCII)
2566
2567 \$baz = 0+\$foo & 0+\$bar; # both ops explicitly numeric
2568 \$biz = "\$foo" ^ "\$bar"; # both ops explicitly stringy
a0d0e21e 2569
1ae175c8
GS
2570See L<perlfunc/vec> for information on how to manipulate individual bits
2571in a bit vector.
2572
d74e8afc 2574X<integer>
a0d0e21e 2575
19799a22 2576By default, Perl assumes that it must do most of its arithmetic in
a0d0e21e
LW
2577floating point. But by saying
2578
2579 use integer;
2580
2581you may tell the compiler that it's okay to use integer operations
19799a22
GS
2582(if it feels like it) from here to the end of the enclosing BLOCK.
2583An inner BLOCK may countermand this by saying
a0d0e21e
LW
2584
2585 no integer;
2586
19799a22
GS
2587which lasts until the end of that BLOCK. Note that this doesn't
2588mean everything is only an integer, merely that Perl may use integer
2589operations if it is so inclined. For example, even under C<use
2590integer>, if you take the C<sqrt(2)>, you'll still get C<1.4142135623731>
2591or so.
2592
2593Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
13a2d996 2595L<Bitwise String Operators>.) However, C<use integer> still has meaning for
19799a22
GS
2596them. By default, their results are interpreted as unsigned integers, but
2597if C<use integer> is in effect, their results are interpreted
2598as signed integers. For example, C<~0> usually evaluates to a large
0be96356 2599integral value. However, C<use integer; ~0> is C<-1> on two's-complement
19799a22 2600machines.
68dc0745 2601
d74e8afc 2603X<floating-point> X<floating point> X<float> X<real>
68dc0745 2604
2605While C<use integer> provides integer-only arithmetic, there is no
19799a22
GS
2606analogous mechanism to provide automatic rounding or truncation to a
2607certain number of decimal places. For rounding to a certain number
2608of digits, sprintf() or printf() is usually the easiest route.
2609See L<perlfaq4>.
68dc0745 2610
5a964f20
TC
2611Floating-point numbers are only approximations to what a mathematician
2612would call real numbers. There are infinitely more reals than floats,
2613so some corners must be cut. For example:
2614
2615 printf "%.20g\n", 123456789123456789;
2616 # produces 123456789123456784
2617
8548cb57
RGS
2618Testing for exact floating-point equality or inequality is not a
2619good idea. Here's a (relatively expensive) work-around to compare
5a964f20
TC
2620whether two floating-point numbers are equal to a particular number of
2621decimal places. See Knuth, volume II, for a more robust treatment of
2622this topic.
2623
2624 sub fp_equal {
2625 my (\$X, \$Y, \$POINTS) = @_;
2626 my (\$tX, \$tY);
2627 \$tX = sprintf("%.\${POINTS}g", \$X);
2628 \$tY = sprintf("%.\${POINTS}g", \$Y);
2629 return \$tX eq \$tY;
2630 }
2631
68dc0745 2632The POSIX module (part of the standard perl distribution) implements
19799a22
GS
2633ceil(), floor(), and other mathematical and trigonometric functions.
2634The Math::Complex module (part of the standard perl distribution)
2635defines mathematical functions that work on both the reals and the
2636imaginary numbers. Math::Complex not as efficient as POSIX, but
68dc0745 2637POSIX can't work with complex numbers.
2638
2639Rounding in financial applications can have serious implications, and
2640the rounding method used should be specified precisely. In these
2641cases, it probably pays not to trust whichever system rounding is
2642being used by Perl, but to instead implement the rounding function you
2643need yourself.
5a964f20
TC
2644
d74e8afc 2646X<number, arbitrary precision>
5a964f20
TC
2647
2648The standard Math::BigInt and Math::BigFloat modules provide
19799a22 2649variable-precision arithmetic and overloaded operators, although
cd5c4fce 2650they're currently pretty slow. At the cost of some space and
19799a22
GS
2651considerable speed, they avoid the normal pitfalls associated with
2652limited-precision representations.
5a964f20
TC
2653
2654 use Math::BigInt;
2655 \$x = Math::BigInt->new('123456789123456789');
2656 print \$x * \$x;
2657
2658 # prints +15241578780673678515622620750190521
19799a22 2659
cd5c4fce
T
2660There are several modules that let you calculate with (bound only by
2661memory and cpu-time) unlimited or fixed precision. There are also
2662some non-standard modules that provide faster implementations via
2663external C libraries.
2664
2665Here is a short, but incomplete summary:
2666
950b09ed
KW
2667 Math::Fraction big, unlimited fractions like 9973 / 12967
2668 Math::String treat string sequences like numbers
2669 Math::FixedPrecision calculate with a fixed precision
2670 Math::Currency for currency calculations
2671 Bit::Vector manipulate bit vectors fast (uses C)
2672 Math::BigIntFast Bit::Vector wrapper for big numbers