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 154before or after the value is returned. This also means that modifying
c543c01b 155a variable twice in the same statement will lead to undefined behavior.
b033823e
A
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
c543c01b
TC
171 print ++(\$foo = "99"); # prints "100"
172 print ++(\$foo = "a0"); # prints "a1"
173 print ++(\$foo = "Az"); # prints "Ba"
174 print ++(\$foo = "zz"); # prints "aaa"
a0d0e21e 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 196
da2f94c5
FC
197Unary "-" performs arithmetic negation if the operand is numeric,
198including any string that looks like a number. If the operand is
199an identifier, a string consisting of a minus sign concatenated
200with the identifier is returned. Otherwise, if the string starts
201with a plus or minus, a string starting with the opposite sign is
202returned. One effect of these rules is that -bareword is equivalent
8705167b 203to the string "-bareword". If, however, the string begins with a
353c6505 204non-alphabetic character (excluding "+" or "-"), Perl will attempt to convert
06705523
SP
205the string to a numeric and the arithmetic negation is performed. If the
206string cannot be cleanly converted to a numeric, Perl will give the warning
207B<Argument "the string" isn't numeric in negation (-) at ...>.
d74e8afc 208X<-> X<negation, arithmetic>
a0d0e21e 209
972b05a9
JH
210Unary "~" performs bitwise negation, i.e., 1's complement. For
212L<Bitwise String Operators>.) Note that the width of the result is
213platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
214bits wide on a 64-bit platform, so if you are expecting a certain bit
f113cf86 215width, remember to use the "&" operator to mask off the excess bits.
d74e8afc 216X<~> X<negation, binary>
a0d0e21e 217
f113cf86
TC
218When complementing strings, if all characters have ordinal values under
219256, then their complements will, also. But if they do not, all
220characters will be in either 32- or 64-bit complements, depending on your
221architecture. So for example, C<~"\x{3B1}"> is C<"\x{FFFF_FC4E}"> on
22232-bit machines and C<"\x{FFFF_FFFF_FFFF_FC4E}"> on 64-bit machines.
223
a0d0e21e
LW
224Unary "+" has no effect whatsoever, even on strings. It is useful
225syntactically for separating a function name from a parenthesized expression
226that would otherwise be interpreted as the complete list of function
5ba421f6 227arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
d74e8afc 228X<+>
a0d0e21e 229
19799a22
GS
230Unary "\" creates a reference to whatever follows it. See L<perlreftut>
231and L<perlref>. Do not confuse this behavior with the behavior of
232backslash within a string, although both forms do convey the notion
233of protecting the next thing from interpolation.
d74e8afc 234X<\> X<reference> X<backslash>
a0d0e21e
LW
235
d74e8afc 237X<binding> X<operator, binding> X<=~> X<!~>
a0d0e21e 238
c07a80fd 239Binary "=~" binds a scalar expression to a pattern match. Certain operations
cb1a09d0
240search or modify the string \$_ by default. This operator makes that kind
241of operation work on some other string. The right argument is a search
TP
242pattern, substitution, or transliteration. The left argument is what is
243supposed to be searched, substituted, or transliterated instead of the default
f8bab1e9 244\$_. When used in scalar context, the return value generally indicates the
8ff32507
FC
245success of the operation. The exceptions are substitution (s///)
246and transliteration (y///) with the C</r> (non-destructive) option,
247which cause the B<r>eturn value to be the result of the substitution.
248Behavior in list context depends on the particular operator.
000c65fc
DG
249See L</"Regexp Quote-Like Operators"> for details and L<perlretut> for
250examples using these operators.
f8bab1e9
GS
251
252If the right argument is an expression rather than a search pattern,
2c268ad5 253substitution, or transliteration, it is interpreted as a search pattern at run
89d205f2
YO
254time. Note that this means that its contents will be interpolated twice, so
255
256 '\\' =~ q'\\';
257
258is not ok, as the regex engine will end up trying to compile the
259pattern C<\>, which it will consider a syntax error.
a0d0e21e
LW
260
261Binary "!~" is just like "=~" except the return value is negated in
262the logical sense.
263
8ff32507
FC
264Binary "!~" with a non-destructive substitution (s///r) or transliteration
265(y///r) is a syntax error.
4f4d7508 266
d74e8afc 268X<operator, multiplicative>
a0d0e21e
LW
269
270Binary "*" multiplies two numbers.
d74e8afc 271X<*>
a0d0e21e
LW
272
273Binary "/" divides two numbers.
d74e8afc 274X</> X<slash>
a0d0e21e 275
f7918450
KW
276Binary "%" is the modulo operator, which computes the division
277remainder of its first argument with respect to its second argument.
278Given integer
54310121 279operands C<\$a> and C<\$b>: If C<\$b> is positive, then C<\$a % \$b> is
f7918450 280C<\$a> minus the largest multiple of C<\$b> less than or equal to
54310121 281C<\$a>. If C<\$b> is negative, then C<\$a % \$b> is C<\$a> minus the
282smallest multiple of C<\$b> that is not less than C<\$a> (i.e. the
89b4f0ad 283result will be less than or equal to zero). If the operands
4848a83b
TS
284C<\$a> and C<\$b> are floating point values and the absolute value of
285C<\$b> (that is C<abs(\$b)>) is less than C<(UV_MAX + 1)>, only
286the integer portion of C<\$a> and C<\$b> will be used in the operation
287(Note: here C<UV_MAX> means the maximum of the unsigned integer type).
288If the absolute value of the right operand (C<abs(\$b)>) is greater than
289or equal to C<(UV_MAX + 1)>, "%" computes the floating-point remainder
290C<\$r> in the equation C<(\$r = \$a - \$i*\$b)> where C<\$i> is a certain
f7918450 291integer that makes C<\$r> have the same sign as the right operand
4848a83b
TS
292C<\$b> (B<not> as the left operand C<\$a> like C function C<fmod()>)
293and the absolute value less than that of C<\$b>.
0412d526 294Note that when C<use integer> is in scope, "%" gives you direct access
f7918450 295to the modulo operator as implemented by your C compiler. This
55d729e4
GS
296operator is not as well defined for negative operands, but it will
297execute faster.
f7918450 298X<%> X<remainder> X<modulo> X<mod>
55d729e4 299
62d10b70
GS
300Binary "x" is the repetition operator. In scalar context or if the left
301operand is not enclosed in parentheses, it returns a string consisting
302of the left operand repeated the number of times specified by the right
303operand. In list context, if the left operand is enclosed in
3585017f
YST
304parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
305If the right operand is zero or negative, it returns an empty string
306or an empty list, depending on the context.
d74e8afc 307X<x>
a0d0e21e
LW
308
309 print '-' x 80; # print row of dashes
310
311 print "\t" x (\$tab/8), ' ' x (\$tab%8); # tab over
312
313 @ones = (1) x 80; # a list of 80 1's
314 @ones = (5) x @ones; # set all elements to 5
315
316
a0d0e21e
LW
319
320Binary "+" returns the sum of two numbers.
d74e8afc 321X<+>
a0d0e21e
LW
322
323Binary "-" returns the difference of two numbers.
d74e8afc 324X<->
a0d0e21e
LW
325
326Binary "." concatenates two strings.
d74e8afc
ITB
327X<string, concatenation> X<concatenation>
328X<cat> X<concat> X<concatenate> X<.>
a0d0e21e
LW
329
d74e8afc
ITB
331X<shift operator> X<operator, shift> X<<< << >>>
332X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
333X<shl> X<shr> X<shift, right> X<shift, left>
a0d0e21e 334
55497cff 335Binary "<<" returns the value of its left argument shifted left by the
336number of bits specified by the right argument. Arguments should be
a0d0e21e 338
55497cff 339Binary ">>" returns the value of its left argument shifted right by
340the number of bits specified by the right argument. Arguments should
a0d0e21e 342
b16cf6df
JH
343Note that both "<<" and ">>" in Perl are implemented directly using
344"<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
345in force then signed C integers are used, else unsigned C integers are
346used. Either way, the implementation isn't going to generate results
347larger than the size of the integer type Perl was built with (32 bits
348or 64 bits).
349
350The result of overflowing the range of the integers is undefined
351because it is undefined also in C. In other words, using 32-bit
352integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
353of bits is also undefined.
354
d74e8afc 356X<operator, named unary>
a0d0e21e
LW
357
358The various named unary operators are treated as functions with one
568e6d8b 359argument, with optional parentheses.
a0d0e21e
LW
360
361If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
362is followed by a left parenthesis as the next token, the operator and
363arguments within parentheses are taken to be of highest precedence,
3981b0eb
JA
364just like a normal function call. For example,
365because named unary operators are higher precedence than ||:
a0d0e21e
LW
366
367 chdir \$foo || die; # (chdir \$foo) || die
368 chdir(\$foo) || die; # (chdir \$foo) || die
369 chdir (\$foo) || die; # (chdir \$foo) || die
370 chdir +(\$foo) || die; # (chdir \$foo) || die
371
3981b0eb 372but, because * is higher precedence than named operators:
a0d0e21e
LW
373
374 chdir \$foo * 20; # chdir (\$foo * 20)
375 chdir(\$foo) * 20; # (chdir \$foo) * 20
376 chdir (\$foo) * 20; # (chdir \$foo) * 20
377 chdir +(\$foo) * 20; # chdir (\$foo * 20)
378
379 rand 10 * 20; # rand (10 * 20)
380 rand(10) * 20; # (rand 10) * 20
381 rand (10) * 20; # (rand 10) * 20
382 rand +(10) * 20; # rand (10 * 20)
383
568e6d8b
RGS
384Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
385treated like named unary operators, but they don't follow this functional
386parenthesis rule. That means, for example, that C<-f(\$file).".bak"> is
387equivalent to C<-f "\$file.bak">.
d74e8afc 388X<-X> X<filetest> X<operator, filetest>
568e6d8b 389
a0d0e21e
LW
391
d74e8afc 393X<relational operator> X<operator, relational>
a0d0e21e 394
35f2feb0 395Binary "<" returns true if the left argument is numerically less than
a0d0e21e 396the right argument.
d74e8afc 397X<< < >>
a0d0e21e 398
35f2feb0 399Binary ">" returns true if the left argument is numerically greater
a0d0e21e 400than the right argument.
d74e8afc 401X<< > >>
a0d0e21e 402
35f2feb0 403Binary "<=" returns true if the left argument is numerically less than
a0d0e21e 404or equal to the right argument.
d74e8afc 405X<< <= >>
a0d0e21e 406
35f2feb0 407Binary ">=" returns true if the left argument is numerically greater
a0d0e21e 408than or equal to the right argument.
d74e8afc 409X<< >= >>
a0d0e21e
LW
410
411Binary "lt" returns true if the left argument is stringwise less than
412the right argument.
d74e8afc 413X<< lt >>
a0d0e21e
LW
414
415Binary "gt" returns true if the left argument is stringwise greater
416than the right argument.
d74e8afc 417X<< gt >>
a0d0e21e
LW
418
419Binary "le" returns true if the left argument is stringwise less than
420or equal to the right argument.
d74e8afc 421X<< le >>
a0d0e21e
LW
422
423Binary "ge" returns true if the left argument is stringwise greater
424than or equal to the right argument.
d74e8afc 425X<< ge >>
a0d0e21e
LW
426
d74e8afc 428X<equality> X<equal> X<equals> X<operator, equality>
a0d0e21e
LW
429
430Binary "==" returns true if the left argument is numerically equal to
431the right argument.
d74e8afc 432X<==>
a0d0e21e
LW
433
434Binary "!=" returns true if the left argument is numerically not equal
435to the right argument.
d74e8afc 436X<!=>
a0d0e21e 437
35f2feb0 438Binary "<=>" returns -1, 0, or 1 depending on whether the left
6ee5d4e7 439argument is numerically less than, equal to, or greater than the right
7d3a9d88
NC
441values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
442"<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
443returns true, as does NaN != anything else. If your platform doesn't
444support NaNs then NaN is just a string with numeric value 0.
d74e8afc 445X<< <=> >> X<spaceship>
7d3a9d88 446
2b54f59f
YST
447 perl -le '\$a = "NaN"; print "No NaN support here" if \$a == \$a'
448 perl -le '\$a = "NaN"; print "NaN support here" if \$a != \$a'
a0d0e21e
LW
449
450Binary "eq" returns true if the left argument is stringwise equal to
451the right argument.
d74e8afc 452X<eq>
a0d0e21e
LW
453
454Binary "ne" returns true if the left argument is stringwise not equal
455to the right argument.
d74e8afc 456X<ne>
a0d0e21e 457
JH
458Binary "cmp" returns -1, 0, or 1 depending on whether the left
459argument is stringwise less than, equal to, or greater than the right
460argument.
d74e8afc 461X<cmp>
a0d0e21e 462
0d863452 463Binary "~~" does a smart match between its arguments. Smart matching
0f7107a0 464is described in L<perlsyn/"Smart matching in detail">.
0d863452
RH
465X<~~>
466
a034a98d
DD
467"lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
468by the current locale if C<use locale> is in effect. See L<perllocale>.
469
d74e8afc 471X<operator, bitwise, and> X<bitwise and> X<&>
a0d0e21e 472
2cdc098b 473Binary "&" returns its operands ANDed together bit by bit.
a0d0e21e 475
2cdc098b
MG
476Note that "&" has lower priority than relational operators, so for example
477the brackets are essential in a test like
478
479 print "Even\n" if (\$x & 1) == 0;
480
a0d0e21e 481=head2 Bitwise Or and Exclusive Or
d74e8afc
ITB
482X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
483X<bitwise xor> X<^>
a0d0e21e 484
2cdc098b 485Binary "|" returns its operands ORed together bit by bit.
a0d0e21e 487
2cdc098b 488Binary "^" returns its operands XORed together bit by bit.
a0d0e21e 490
2cdc098b
MG
491Note that "|" and "^" have lower priority than relational operators, so
492for example the brackets are essential in a test like
493
494 print "false\n" if (8 | 2) != 10;
495
d74e8afc 497X<&&> X<logical and> X<operator, logical, and>
a0d0e21e
LW
498
499Binary "&&" performs a short-circuit logical AND operation. That is,
500if the left operand is false, the right operand is not even evaluated.
501Scalar or list context propagates down to the right operand if it
502is evaluated.
503
d74e8afc 505X<||> X<operator, logical, or>
a0d0e21e
LW
506
507Binary "||" performs a short-circuit logical OR operation. That is,
508if the left operand is true, the right operand is not even evaluated.
509Scalar or list context propagates down to the right operand if it
510is evaluated.
511
d74e8afc 513X<//> X<operator, logical, defined-or>
c963b151
BD
514
515Although it has no direct equivalent in C, Perl's C<//> operator is related
89d205f2 516to its C-style or. In fact, it's exactly the same as C<||>, except that it
95bee9ba
A
517tests the left hand side's definedness instead of its truth. Thus,
518C<< EXPR1 // EXPR2 >> returns the value of C<< EXPR1 >> if it's defined,
519otherwise, the value of C<< EXPR2 >> is returned. (C<< EXPR1 >> is evaluated
520in scalar context, C<< EXPR2 >> in the context of C<< // >> itself). Usually,
521this is the same result as C<< defined(EXPR1) ? EXPR1 : EXPR2 >> (except that
522the ternary-operator form can be used as a lvalue, while C<< EXPR1 // EXPR2 >>
523cannot). This is very useful for
bdc7923b
RGS
524providing default values for variables. If you actually want to test if
525at least one of C<\$a> and C<\$b> is defined, use C<defined(\$a // \$b)>.
c963b151 526
d042e63d
MS
527The C<||>, C<//> and C<&&> operators return the last value evaluated
528(unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
529portable way to find out the home directory might be:
a0d0e21e 530
c543c01b
TC
531 \$home = \$ENV{HOME}
532 // \$ENV{LOGDIR}
533 // (getpwuid(\$<))[7]
534 // die "You're homeless!\n";
a0d0e21e 535
5a964f20
TC
536In particular, this means that you shouldn't use this
537for selecting between two aggregates for assignment:
538
539 @a = @b || @c; # this is wrong
540 @a = scalar(@b) || @c; # really meant this
541 @a = @b ? @b : @c; # this works fine, though
542
f23102e2
RGS
543As more readable alternatives to C<&&> and C<||> when used for
544control flow, Perl provides the C<and> and C<or> operators (see below).
545The short-circuit behavior is identical. The precedence of "and"
c963b151 546and "or" is much lower, however, so that you can safely use them after a
5a964f20 547list operator without the need for parentheses:
a0d0e21e
LW
548
550 or gripe(), next LINE;
551
552With the C-style operators that would have been written like this:
553
555 || (gripe(), next LINE);
556
eeb6a2c9 557Using "or" for assignment is unlikely to do what you want; see below.
5a964f20
TC
558
d74e8afc 560X<operator, range> X<range> X<..> X<...>
a0d0e21e
LW
561
562Binary ".." is the range operator, which is really two different
fb53bbb2 563operators depending on the context. In list context, it returns a
54ae734e 564list of values counting (up by ones) from the left value to the right
2cdbc966 565value. If the left value is greater than the right value then it
fb53bbb2 566returns the empty list. The range operator is useful for writing
54ae734e 567C<foreach (1..10)> loops and for doing slice operations on arrays. In
2cdbc966
JD
568the current implementation, no temporary array is created when the
569range operator is used as the expression in C<foreach> loops, but older
570versions of Perl might burn a lot of memory when you write something
571like this:
a0d0e21e
LW
572
573 for (1 .. 1_000_000) {
574 # code
54310121 575 }
a0d0e21e 576
8f0f46f8 577The range operator also works on strings, using the magical
578auto-increment, see below.
54ae734e 579
5a964f20 580In scalar context, ".." returns a boolean value. The operator is
8f0f46f8 581bistable, like a flip-flop, and emulates the line-range (comma)
582operator of B<sed>, B<awk>, and various editors. Each ".." operator
583maintains its own boolean state, even across calls to a subroutine
584that contains it. It is false as long as its left operand is false.
a0d0e21e
LW
585Once the left operand is true, the range operator stays true until the
586right operand is true, I<AFTER> which the range operator becomes false
8f0f46f8 587again. It doesn't become false till the next time the range operator
588is evaluated. It can test the right operand and become false on the
589same evaluation it became true (as in B<awk>), but it still returns
590true once. If you don't want it to test the right operand until the
591next evaluation, as in B<sed>, just use three dots ("...") instead of
19799a22
GS
592two. In all other regards, "..." behaves just like ".." does.
593
594The right operand is not evaluated while the operator is in the
595"false" state, and the left operand is not evaluated while the
596operator is in the "true" state. The precedence is a little lower
597than || and &&. The value returned is either the empty string for
8f0f46f8 598false, or a sequence number (beginning with 1) for true. The sequence
599number is reset for each range encountered. The final sequence number
600in a range has the string "E0" appended to it, which doesn't affect
601its numeric value, but gives you something to search for if you want
602to exclude the endpoint. You can exclude the beginning point by
603waiting for the sequence number to be greater than 1.
df5f8116
CW
604
605If either operand of scalar ".." is a constant expression,
606that operand is considered true if it is equal (C<==>) to the current
607input line number (the C<\$.> variable).
608
609To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
610but that is only an issue if you use a floating point expression; when
611implicitly using C<\$.> as described in the previous paragraph, the
612comparison is C<int(EXPR) == int(\$.)> which is only an issue when C<\$.>
613is set to a floating point value and you are not reading from a file.
614Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
615you want in scalar context because each of the operands are evaluated
616using their integer representation.
617
618Examples:
a0d0e21e
LW
619
620As a scalar operator:
621
df5f8116 622 if (101 .. 200) { print; } # print 2nd hundred lines, short for
950b09ed 623 # if (\$. == 101 .. \$. == 200) { print; }
9f10b797
RGS
624
625 next LINE if (1 .. /^\$/); # skip header lines, short for
f343f960 626 # next LINE if (\$. == 1 .. /^\$/);
9f10b797
RGS
627 # (typically in a loop labeled LINE)
628
629 s/^/> / if (/^\$/ .. eof()); # quote body
a0d0e21e 630
5a964f20
TC
631 # parse mail messages
632 while (<>) {
633 \$in_header = 1 .. /^\$/;
df5f8116
CW
634 \$in_body = /^\$/ .. eof;
f343f960 636 # do something
df5f8116 637 } else { # in body
f343f960 638 # do something else
df5f8116 639 }
5a964f20 640 } continue {
df5f8116 641 close ARGV if eof; # reset \$. each file
5a964f20
TC
642 }
643
acf31ca5
SF
644Here's a simple example to illustrate the difference between
645the two range operators:
646
647 @lines = (" - Foo",
648 "01 - Bar",
649 "1 - Baz",
650 " - Quux");
651
9f10b797
RGS
652 foreach (@lines) {
653 if (/0/ .. /1/) {
acf31ca5
SF
654 print "\$_\n";
655 }
656 }
657
9f10b797
RGS
658This program will print only the line containing "Bar". If
659the range operator is changed to C<...>, it will also print the
acf31ca5
SF
660"Baz" line.
661
662And now some examples as a list operator:
a0d0e21e
LW
663
664 for (101 .. 200) { print; } # print \$_ 100 times
3e3baf6d 665 @foo = @foo[0 .. \$#foo]; # an expensive no-op
a0d0e21e
LW
666 @foo = @foo[\$#foo-4 .. \$#foo]; # slice last 5 items
667
5a964f20 668The range operator (in list context) makes use of the magical
5f05dabc 669auto-increment algorithm if the operands are strings. You
a0d0e21e
LW
670can say
671
c543c01b 672 @alphabet = ("A" .. "Z");
a0d0e21e 673
54ae734e 674to get all normal letters of the English alphabet, or
a0d0e21e 675
c543c01b 676 \$hexdigit = (0 .. 9, "a" .. "f")[\$num & 15];
a0d0e21e
LW
677
678to get a hexadecimal digit, or
679
c543c01b 680 @z2 = ("01" .. "31"); print \$z2[\$mday];
a0d0e21e 681
ea4f5703
YST
682to get dates with leading zeros.
683
684If the final value specified is not in the sequence that the magical
685increment would produce, the sequence goes until the next value would
686be longer than the final value specified.
687
688If the initial value specified isn't part of a magical increment
c543c01b 689sequence (that is, a non-empty string matching C</^[a-zA-Z]*[0-9]*\z/>),
ea4f5703
YST
690only the initial value will be returned. So the following will only
691return an alpha:
692
c543c01b 693 use charnames "greek";
ea4f5703
YST
694 my @greek_small = ("\N{alpha}" .. "\N{omega}");
695
c543c01b
TC
696To get the 25 traditional lowercase Greek letters, including both sigmas,
ea4f5703 698
c543c01b
TC
699 use charnames "greek";
700 my @greek_small = map { chr }
701 ord "\N{alpha}" .. ord "\N{omega}";
702
703However, because there are I<many> other lowercase Greek characters than
704just those, to match lowercase Greek characters in a regular expression,
705you would use the pattern C</(?:(?=\p{Greek})\p{Lower})+/>.
a0d0e21e 706
df5f8116
CW
707Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
708return two elements in list context.
709
710 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
711
d74e8afc 713X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
a0d0e21e
LW
714
715Ternary "?:" is the conditional operator, just as in C. It works much
716like an if-then-else. If the argument before the ? is true, the
717argument before the : is returned, otherwise the argument after the :
cb1a09d0
718is returned. For example:
719
54310121 720 printf "I have %d dog%s.\n", \$n,
c543c01b 721 (\$n == 1) ? "" : "s";
cb1a09d0
722
723Scalar or list context propagates downward into the 2nd
54310121 724or 3rd argument, whichever is selected.
cb1a09d0
725
726 \$a = \$ok ? \$b : \$c; # get a scalar
727 @a = \$ok ? @b : @c; # get an array
728 \$a = \$ok ? @b : @c; # oops, that's just a count!
729
730The operator may be assigned to if both the 2nd and 3rd arguments are
731legal lvalues (meaning that you can assign to them):
a0d0e21e
LW
732
733 (\$a_or_b ? \$a : \$b) = \$c;
734
5a964f20
TC
735Because this operator produces an assignable result, using assignments
736without parentheses will get you in trouble. For example, this:
737
738 \$a % 2 ? \$a += 10 : \$a += 2
739
740Really means this:
741
742 ((\$a % 2) ? (\$a += 10) : \$a) += 2
743
744Rather than this:
745
746 (\$a % 2) ? (\$a += 10) : (\$a += 2)
747
19799a22
GS
748That should probably be written more simply as:
749
750 \$a += (\$a % 2) ? 10 : 2;
751
d74e8afc 753X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
5ac3b81c 754X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
d74e8afc 755X<%=> X<^=> X<x=>
a0d0e21e
LW
756
757"=" is the ordinary assignment operator.
758
759Assignment operators work as in C. That is,
760
761 \$a += 2;
762
763is equivalent to
764
765 \$a = \$a + 2;
766
767although without duplicating any side effects that dereferencing the lvalue
54310121 768might trigger, such as from tie(). Other assignment operators work similarly.
769The following are recognized:
a0d0e21e
LW
770
771 **= += *= &= <<= &&=
9f10b797
RGS
772 -= /= |= >>= ||=
773 .= %= ^= //=
774 x=
a0d0e21e 775
19799a22 776Although these are grouped by family, they all have the precedence
a0d0e21e
LW
777of assignment.
778
b350dd2f
GS
779Unlike in C, the scalar assignment operator produces a valid lvalue.
780Modifying an assignment is equivalent to doing the assignment and
781then modifying the variable that was assigned to. This is useful
782for modifying a copy of something, like this:
a0d0e21e 783
c543c01b 784 (\$tmp = \$global) =~ tr [0-9] [a-j];
a0d0e21e
LW
785
786Likewise,
787
788 (\$a += 2) *= 3;
789
790is equivalent to
791
792 \$a += 2;
793 \$a *= 3;
794
b350dd2f
GS
795Similarly, a list assignment in list context produces the list of
796lvalues assigned to, and a list assignment in scalar context returns
797the number of elements produced by the expression on the right hand
798side of the assignment.
799
c543c01b
TC
802X<triple-dot operator>
803
804The triple-dot operator, C<...>, sometimes called the "whatever operator", the
806code. Perl parses it without error, but when you try to execute a whatever,
807it throws an exception with the text C<Unimplemented>:
808
809 sub unimplemented { ... }
810
811 eval { unimplemented() };
812 if (\$@ eq "Unimplemented" ) {
813 say "Oh look, an exception--whatever.";
814 }
815
816You can only use the triple-dot operator to stand in for a complete statement.
817These examples of the triple-dot work:
818
819 { ... }
820
821 sub foo { ... }
822
823 ...;
824
825 eval { ... };
826
827 sub foo {
828 my (\$self) = shift;
829 ...;
830 }
831
832 do {
833 my \$variable;
834 ...;
835 say "Hurrah!";
836 } while \$cheering;
837
839part of a larger statement since the C<...> is also the three-dot version
840of the binary range operator (see L<Range Operators>). These examples of
841the whatever operator are still syntax errors:
842
843 print ...;
844
845 open(PASSWD, ">", "/dev/passwd") or ...;
846
847 if (\$condition && ...) { say "Hello" }
848
849There are some cases where Perl can't immediately tell the difference
850between an expression and a statement. For instance, the syntax for a
851block and an anonymous hash reference constructor look the same unless
852there's something in the braces that give Perl a hint. The whatever
853is a syntax error if Perl doesn't guess that the C<{ ... }> is a
854block. In that case, it doesn't think the C<...> is the whatever
855because it's expecting an expression instead of a statement:
856
857 my @transformed = map { ... } @input; # syntax error
858
859You can use a C<;> inside your block to denote that the C<{ ... }> is
860a block and not a hash reference constructor. Now the whatever works:
861
862 my @transformed = map {; ... } @input; # ; disambiguates
863
864 my @transformed = map { ...; } @input; # ; disambiguates
865
d74e8afc 867X<comma> X<operator, comma> X<,>
a0d0e21e 868
5a964f20 869Binary "," is the comma operator. In scalar context it evaluates
a0d0e21e
LW
870its left argument, throws that value away, then evaluates its right
871argument and returns that value. This is just like C's comma operator.
872
5a964f20 873In list context, it's just the list argument separator, and inserts
ed5c6d31
PJ
874both its arguments into the list. These arguments are also evaluated
875from left to right.
a0d0e21e 876
4e1988c6
FC
877The C<< => >> operator is a synonym for the comma except that it causes a
878word on its left to be interpreted as a string if it begins with a letter
344f2c40
IG
879or underscore and is composed only of letters, digits and underscores.
880This includes operands that might otherwise be interpreted as operators,
881constants, single number v-strings or function calls. If in doubt about
c543c01b 882this behavior, the left operand can be quoted explicitly.
344f2c40
IG
883
884Otherwise, the C<< => >> operator behaves exactly as the comma operator
885or list argument separator, according to context.
886
887For example:
a44e5664
MS
888
889 use constant FOO => "something";
890
891 my %h = ( FOO => 23 );
892
893is equivalent to:
894
895 my %h = ("FOO", 23);
896
897It is I<NOT>:
898
899 my %h = ("something", 23);
900
719b43e8
RGS
901The C<< => >> operator is helpful in documenting the correspondence
902between keys and values in hashes, and other paired elements in lists.
748a9306 903
a12b8f3c
FC
904 %hash = ( \$key => \$value );
a44e5664 906
4e1988c6
FC
907The special quoting behavior ignores precedence, and hence may apply to
908I<part> of the left operand:
909
910 print time.shift => "bbb";
911
912That example prints something like "1314363215shiftbbb", because the
913C<< => >> implicitly quotes the C<shift> immediately on its left, ignoring
914the fact that C<time.shift> is the entire left operand.
915
d74e8afc 917X<operator, list, rightward> X<list operator>
a0d0e21e 918
c543c01b 919On the right side of a list operator, the comma has very low precedence,
a0d0e21e
LW
920such that it controls all comma-separated expressions found there.
921The only operators with lower precedence are the logical operators
922"and", "or", and "not", which may be used to evaluate calls to list
923operators without the need for extra parentheses:
924
c543c01b
TC
925 open HANDLE, "< \$file"
926 or die "Can't open \$file: \$!\n";
a0d0e21e 927
5ba421f6 928See also discussion of list operators in L<Terms and List Operators (Leftward)>.
a0d0e21e
LW
929
d74e8afc 931X<operator, logical, not> X<not>
a0d0e21e
LW
932
933Unary "not" returns the logical negation of the expression to its right.
934It's the equivalent of "!" except for the very low precedence.
935
d74e8afc 937X<operator, logical, and> X<and>
a0d0e21e
LW
938
939Binary "and" returns the logical conjunction of the two surrounding
c543c01b
TC
940expressions. It's equivalent to C<&&> except for the very low
941precedence. This means that it short-circuits: the right
a0d0e21e
LW
942expression is evaluated only if the left expression is true.
943
59ab9d6e 944=head2 Logical or and Exclusive Or
f23102e2 945X<operator, logical, or> X<operator, logical, xor>
59ab9d6e 946X<operator, logical, exclusive or>
f23102e2 947X<or> X<xor>
a0d0e21e
LW
948
949Binary "or" returns the logical disjunction of the two surrounding
c543c01b
TC
950expressions. It's equivalent to C<||> except for the very low precedence.
951This makes it useful for control flow:
5a964f20
TC
952
953 print FH \$data or die "Can't write to FH: \$!";
954
c543c01b
TC
955This means that it short-circuits: the right expression is evaluated
956only if the left expression is false. Due to its precedence, you must
957be careful to avoid using it as replacement for the C<||> operator.
958It usually works out better for flow control than in assignments:
5a964f20
TC
959
960 \$a = \$b or \$c; # bug: this is wrong
961 (\$a = \$b) or \$c; # really means this
962 \$a = \$b || \$c; # better written this way
963
19799a22 964However, when it's a list-context assignment and you're trying to use
c543c01b 965C<||> for control flow, you probably need "or" so that the assignment
5a964f20
TC
966takes higher precedence.
967
968 @info = stat(\$file) || die; # oops, scalar sense of stat!
969 @info = stat(\$file) or die; # better, now @info gets its due
970
c963b151
BD
971Then again, you could always use parentheses.
972
a0d0e21e 973Binary "xor" returns the exclusive-OR of the two surrounding expressions.
c543c01b 974It cannot short-circuit (of course).
a0d0e21e 975
59ab9d6e
MB
976There is no low precedence operator for defined-OR.
977
a0d0e21e 978=head2 C Operators Missing From Perl
d74e8afc
ITB
979X<operator, missing from perl> X<&> X<*>
980X<typecasting> X<(TYPE)>
a0d0e21e
LW
981
982Here is what C has that Perl doesn't:
983
984=over 8
985
986=item unary &
987
988Address-of operator. (But see the "\" operator for taking a reference.)
989
990=item unary *
991
54310121 992Dereference-address operator. (Perl's prefix dereferencing
a0d0e21e
LW
993operators are typed: \$, @, %, and &.)
994
995=item (TYPE)
996
19799a22 997Type-casting operator.
a0d0e21e
LW
998
999=back
1000
5f05dabc 1001=head2 Quote and Quote-like Operators
89d205f2 1002X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
d74e8afc
ITB
1003X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
1004X<escape sequence> X<escape>
1005
a0d0e21e
LW
1006While we usually think of quotes as literal values, in Perl they
1007function as operators, providing various kinds of interpolating and
1008pattern matching capabilities. Perl provides customary quote characters
1009for these behaviors, but also provides a way for you to choose your
1010quote character for any of them. In the following table, a C<{}> represents
9f10b797 1011any pair of delimiters you choose.
a0d0e21e 1012
TP
1013 Customary Generic Meaning Interpolates
1014 '' q{} Literal no
1015 "" qq{} Literal yes
af9219ee 1016 `` qx{} Command yes*
2c268ad5 1017 qw{} Word list no
af9219ee
MG
1018 // m{} Pattern match yes*
1019 qr{} Pattern yes*
1020 s{}{} Substitution yes*
2c268ad5 1021 tr{}{} Transliteration no (but see below)
c543c01b 1022 y{}{} Transliteration no (but see below)
7e3b091d 1023 <<EOF here-doc yes*
a0d0e21e 1024
af9219ee
MG
1025 * unless the delimiter is ''.
1026
87275199 1027Non-bracketing delimiters use the same character fore and aft, but the four
c543c01b 1028sorts of ASCII brackets (round, angle, square, curly) all nest, which means
9f10b797 1029that
87275199 1030
c543c01b 1031 q{foo{bar}baz}
35f2feb0 1032
9f10b797 1033is the same as
87275199 1034
c543c01b 1035 'foo{bar}baz'
87275199
GS
1036
1037Note, however, that this does not always work for quoting Perl code:
1038
c543c01b 1039 \$s = q{ if(\$a eq "}") ... }; # WRONG
87275199 1040
c543c01b
TC
1041is a syntax error. The C<Text::Balanced> module (standard as of v5.8,
1042and from CPAN before then) is able to do this properly.
87275199 1043
19799a22 1044There can be whitespace between the operator and the quoting
fb73857a 1045characters, except when C<#> is being used as the quoting character.
19799a22
GS
1046C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
1047operator C<q> followed by a comment. Its argument will be taken
1048from the next line. This allows you to write:
fb73857a 1049
1050 s {foo} # Replace foo
1051 {bar} # with bar.
1052
c543c01b
TC
1053The following escape sequences are available in constructs that interpolate,
1054and in transliterations:
5691ca5f 1055X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N> X<\N{}>
04341565 1056X<\o{}>
5691ca5f 1057
2c4c1ff2
KW
1058 Sequence Note Description
1059 \t tab (HT, TAB)
1060 \n newline (NL)
1061 \r return (CR)
1062 \f form feed (FF)
1063 \b backspace (BS)
1064 \a alarm (bell) (BEL)
1065 \e escape (ESC)
c543c01b 1066 \x{263A} [1,8] hex char (example: SMILEY)
2c4c1ff2 1067 \x1b [2,8] restricted range hex char (example: ESC)
fb121860 1068 \N{name} [3] named Unicode character or character sequence
2c4c1ff2
KW
1069 \N{U+263D} [4,8] Unicode character (example: FIRST QUARTER MOON)
1070 \c[ [5] control char (example: chr(27))
1071 \o{23072} [6,8] octal char (example: SMILEY)
1072 \033 [7,8] restricted range octal char (example: ESC)
5691ca5f
KW
1073
1074=over 4
1075
1076=item [1]
1077
2c4c1ff2
KW
1078The result is the character specified by the hexadecimal number between
1079the braces. See L</[8]> below for details on which character.
96448467
DG
1080
1081Only hexadecimal digits are valid between the braces. If an invalid
1082character is encountered, a warning will be issued and the invalid
1083character and all subsequent characters (valid or invalid) within the
1085
1086If there are no valid digits between the braces, the generated character is
1087the NULL character (C<\x{00}>). However, an explicit empty brace (C<\x{}>)
c543c01b 1088will not cause a warning (currently).
40687185
KW
1089
1090=item [2]
1091
2c4c1ff2
KW
1092The result is the character specified by the hexadecimal number in the range
10930x00 to 0xFF. See L</[8]> below for details on which character.
96448467
DG
1094
1095Only hexadecimal digits are valid following C<\x>. When C<\x> is followed
2c4c1ff2 1096by fewer than two valid digits, any valid digits will be zero-padded. This
c543c01b 1097means that C<\x7> will be interpreted as C<\x07>, and a lone <\x> will be
2c4c1ff2 1098interpreted as C<\x00>. Except at the end of a string, having fewer than
c543c01b 1099two valid digits will result in a warning. Note that although the warning
96448467
DG
1100says the illegal character is ignored, it is only ignored as part of the
1101escape and will still be used as the subsequent character in the string.
1102For example:
1103
1104 Original Result Warns?
1105 "\x7" "\x07" no
1106 "\x" "\x00" no
1107 "\x7q" "\x07q" yes
1108 "\xq" "\x00q" yes
1109
40687185
KW
1110=item [3]
1111
fb121860 1112The result is the Unicode character or character sequence given by I<name>.
2c4c1ff2 1113See L<charnames>.
40687185
KW
1114
1115=item [4]
1116
2c4c1ff2
KW
1117C<\N{U+I<hexadecimal number>}> means the Unicode character whose Unicode code
40687185
KW
1119
1120=item [5]
1121
5691ca5f
KW
1122The character following C<\c> is mapped to some other character as shown in the
1123table:
1124
1125 Sequence Value
1126 \c@ chr(0)
1127 \cA chr(1)
1128 \ca chr(1)
1129 \cB chr(2)
1130 \cb chr(2)
1131 ...
1132 \cZ chr(26)
1133 \cz chr(26)
1134 \c[ chr(27)
1135 \c] chr(29)
1136 \c^ chr(30)
1137 \c? chr(127)
1138
d813941f
KW
1139In other words, it's the character whose code point has had 64 xor'd with
1140its uppercase. C<\c?> is DELETE because C<ord("@") ^ 64> is 127, and
1141C<\c@> is NULL because the ord of "@" is 64, so xor'ing 64 itself produces 0.
1142
5691ca5f
KW
1143Also, C<\c\I<X>> yields C< chr(28) . "I<X>"> for any I<X>, but cannot come at the
1144end of a string, because the backslash would be parsed as escaping the end
1145quote.
1146
1147On ASCII platforms, the resulting characters from the list above are the
1148complete set of ASCII controls. This isn't the case on EBCDIC platforms; see
1149L<perlebcdic/OPERATOR DIFFERENCES> for the complete list of what these
1150sequences mean on both ASCII and EBCDIC platforms.
1151
1152Use of any other character following the "c" besides those listed above is
17a3df4c
KW
1153discouraged, and some are deprecated with the intention of removing
1154those in Perl 5.16. What happens for any of these
d813941f
KW
1155other characters currently though, is that the value is derived by xor'ing
1156with the seventh bit, which is 64.
5691ca5f
KW
1157
1158To get platform independent controls, you can use C<\N{...}>.
1159
40687185
KW
1160=item [6]
1161
2c4c1ff2
KW
1162The result is the character specified by the octal number between the braces.
1163See L</[8]> below for details on which character.
04341565
DG
1164
1165If a character that isn't an octal digit is encountered, a warning is raised,
1166and the value is based on the octal digits before it, discarding it and all
1167following characters up to the closing brace. It is a fatal error if there are
1168no octal digits at all.
1169
1170=item [7]
1171
c543c01b 1172The result is the character specified by the three-digit octal number in the
2c4c1ff2
KW
1173range 000 to 777 (but best to not use above 077, see next paragraph). See
1174L</[8]> below for details on which character.
1175
1176Some contexts allow 2 or even 1 digit, but any usage without exactly
40687185 1177three digits, the first being a zero, may give unintended results. (For
5db3e519
FC
1178example, in a regular expression it may be confused with a backreference;
1179see L<perlrebackslash/Octal escapes>.) Starting in Perl 5.14, you may
c543c01b 1180use C<\o{}> instead, which avoids all these problems. Otherwise, it is best to
04341565
DG
1181use this construct only for ordinals C<\077> and below, remembering to pad to
1182the left with zeros to make three digits. For larger ordinals, either use
d90d5a38 1183C<\o{}> , or convert to something else, such as to hex and use C<\x{}>
40687185 1185
40687185
KW
1186Having fewer than 3 digits may lead to a misleading warning message that says
1187that what follows is ignored. For example, C<"\128"> in the ASCII character set
1188is equivalent to the two characters C<"\n8">, but the warning C<Illegal octal
5db3e519
FC
1189digit '8' ignored> will be thrown. If C<"\n8"> is what you want, you can
5691ca5f 1191
2c4c1ff2
KW
1192=item [8]
1193
c543c01b 1194Several constructs above specify a character by a number. That number
2c4c1ff2 1195gives the character's position in the character set encoding (indexed from 0).
c543c01b 1196This is called synonymously its ordinal, code position, or code point. Perl
2c4c1ff2
KW
1197works on platforms that have a native encoding currently of either ASCII/Latin1
1198or EBCDIC, each of which allow specification of 256 characters. In general, if
1199the number is 255 (0xFF, 0377) or below, Perl interprets this in the platform's
1200native encoding. If the number is 256 (0x100, 0400) or above, Perl interprets
c543c01b 1201it as a Unicode code point and the result is the corresponding Unicode
2c4c1ff2
KW
1202character. For example C<\x{50}> and C<\o{120}> both are the number 80 in
1203decimal, which is less than 256, so the number is interpreted in the native
1204character set encoding. In ASCII the character in the 80th position (indexed
1205from 0) is the letter "P", and in EBCDIC it is the ampersand symbol "&".
1206C<\x{100}> and C<\o{400}> are both 256 in decimal, so the number is interpreted
1207as a Unicode code point no matter what the native encoding is. The name of the
1208character in the 100th position (indexed by 0) in Unicode is
1209C<LATIN CAPITAL LETTER A WITH MACRON>.
1210
1211There are a couple of exceptions to the above rule. C<\N{U+I<hex number>}> is
1212always interpreted as a Unicode code point, so that C<\N{U+0050}> is "P" even
1213on EBCDIC platforms. And if L<C<S<use encoding>>|encoding> is in effect, the
1214number is considered to be in that encoding, and is translated from that into
1215the platform's native encoding if there is a corresponding native character;
1216otherwise to Unicode.
1217
5691ca5f 1218=back
4c77eaa2 1219
e526e8bb
KW
1220B<NOTE>: Unlike C and other languages, Perl has no C<\v> escape sequence for
1221the vertical tab (VT - ASCII 11), but you may use C<\ck> or C<\x0b>. (C<\v>
1222does have meaning in regular expression patterns in Perl, see L<perlre>.)
1223
1224The following escape sequences are available in constructs that interpolate,
904501ec 1225but not in transliterations.
d74e8afc 1226X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q>
904501ec 1227
c543c01b
TC
1228 \l lowercase next character only
1229 \u titlecase (not uppercase!) next character only
e4d34742
EB
1230 \L lowercase all characters till \E or end of string
1231 \U uppercase all characters till \E or end of string
1232 \Q quote non-word characters till \E or end of string
7e31b643 1233 \E end either case modification or quoted section
c543c01b
TC
1234 (whichever was last seen)
1235
1236C<\L>, C<\U>, and C<\Q> can stack, in which case you need one
1237C<\E> for each. For example:
1238
1239 say "This \Qquoting \ubusiness \Uhere isn't quite\E done yet,\E is it?";
1240 This quoting\ Business\ HERE\ ISN\'T\ QUITE\ done\ yet\, is it?
a0d0e21e 1241
95cc3e0c 1242If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
c543c01b 1243C<\u>, and C<\U> is taken from the current locale. See L<perllocale>.
b6538e4f 1244If Unicode (for example, C<\N{}> or code points of 0x100 or
c543c01b
TC
1245beyond) is being used, the case map used by C<\l>, C<\L>, C<\u>, and
1246C<\U> is as defined by Unicode. That means that case-mapping
1247a single character can sometimes produce several characters.
a034a98d 1248
5a964f20
TC
1249All systems use the virtual C<"\n"> to represent a line terminator,
1250called a "newline". There is no such thing as an unvarying, physical
19799a22 1251newline character. It is only an illusion that the operating system,
5a964f20
TC
1252device drivers, C libraries, and Perl all conspire to preserve. Not all
1253systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
c543c01b
TC
1254on the ancient Macs (pre-MacOS X) of yesteryear, these used to be reversed,
1255and on systems without line terminator,
1256printing C<"\n"> might emit no actual data. In general, use C<"\n"> when
5a964f20
TC
1257you mean a "newline" for your system, but use the literal ASCII when you
1258need an exact character. For example, most networking protocols expect
2a380090 1259and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
5a964f20
TC
1260and although they often accept just C<"\012">, they seldom tolerate just
1261C<"\015">. If you get in the habit of using C<"\n"> for networking,
1262you may be burned some day.
d74e8afc
ITB
1263X<newline> X<line terminator> X<eol> X<end of line>
1264X<\n> X<\r> X<\r\n>
5a964f20 1265
904501ec
MG
1266For constructs that do interpolate, variables beginning with "C<\$>"
1267or "C<@>" are interpolated. Subscripted variables such as C<\$a[3]> or
A
1268C<< \$href->{key}[0] >> are also interpolated, as are array and hash slices.
1269But method calls such as C<< \$obj->meth >> are not.
af9219ee
MG
1270
1271Interpolating an array or slice interpolates the elements in order,
1272separated by the value of C<\$">, so is equivalent to interpolating
c543c01b
TC
1273C<join \$", @array>. "Punctuation" arrays such as C<@*> are usually
1274interpolated only if the name is enclosed in braces C<@{*}>, but the
1275arrays C<@_>, C<@+>, and C<@-> are interpolated even without braces.
af9219ee 1276
bc7b91c6
EB
1277For double-quoted strings, the quoting from C<\Q> is applied after
1278interpolation and escapes are processed.
1279
1280 "abc\Qfoo\tbar\$s\Exyz"
1281
1282is equivalent to
1283
1284 "abc" . quotemeta("foo\tbar\$s") . "xyz"
1285
1286For the pattern of regex operators (C<qr//>, C<m//> and C<s///>),
1287the quoting from C<\Q> is applied after interpolation is processed,
1288but before escapes are processed. This allows the pattern to match
1289literally (except for C<\$> and C<@>). For example, the following matches:
1290
1291 '\s\t' =~ /\Q\s\t/
1292
1293Because C<\$> or C<@> trigger interpolation, you'll need to use something
1294like C</\Quser\E\@\Qhost/> to match them literally.
1d2dff63 1295
a0d0e21e
LW
1296Patterns are subject to an additional level of interpretation as a
1297regular expression. This is done as a second pass, after variables are
1298interpolated, so that regular expressions may be incorporated into the
1299pattern from the variables. If this is not what you want, use C<\Q> to
1300interpolate a variable literally.
1301
19799a22
GS
1302Apart from the behavior described above, Perl does not expand
1303multiple levels of interpolation. In particular, contrary to the
1304expectations of shell programmers, back-quotes do I<NOT> interpolate
1305within double quotes, nor do single quotes impede evaluation of
1306variables when used within double quotes.
a0d0e21e 1307
d74e8afc 1309X<operator, regexp>
cb1a09d0 1310
5f05dabc 1311Here are the quote-like operators that apply to pattern
cb1a09d0
1312matching and related activities.
1313
a0d0e21e
LW
1314=over 8
1315
b6fa137b 1316=item qr/STRING/msixpodual
01c6f5f4 1317X<qr> X</i> X</m> X</o> X</s> X</x> X</p>
a0d0e21e 1318
87e95b7f
YO
1319This operator quotes (and possibly compiles) its I<STRING> as a regular
1320expression. I<STRING> is interpolated the same way as I<PATTERN>
1321in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1322is done. Returns a Perl value which may be used instead of the
f6050459 1323corresponding C</STRING/msixpodual> expression. The returned value is a
85dd5c8b 1324normalized version of the original pattern. It magically differs from
1c8ee595
CO
1325a string containing the same characters: C<ref(qr/x/)> returns "Regexp";
1326however, dereferencing it is not well defined (you currently get the
1327normalized version of the original pattern, but this may change).
1328
a0d0e21e 1329
87e95b7f
YO
1330For example,
1331
1332 \$rex = qr/my.STRING/is;
85dd5c8b 1333 print \$rex; # prints (?si-xm:my.STRING)
87e95b7f
YO
1334 s/\$rex/foo/;
1335
1336is equivalent to
1337
1338 s/my.STRING/foo/is;
1339
1340The result may be used as a subpattern in a match:
1341
1342 \$re = qr/\$pattern/;
1343 \$string =~ /foo\${re}bar/; # can be interpolated in other patterns
1344 \$string =~ \$re; # or used standalone
1345 \$string =~ /\$re/; # or this way
1346
f6050459 1347Since Perl may compile the pattern at the moment of execution of the qr()
87e95b7f
YO
1348operator, using qr() may have speed advantages in some situations,
1349notably if the result of qr() is used standalone:
1350
1351 sub match {
1352 my \$patterns = shift;
1353 my @compiled = map qr/\$_/i, @\$patterns;
1354 grep {
1355 my \$success = 0;
1356 foreach my \$pat (@compiled) {
1357 \$success = 1, last if /\$pat/;
1358 }
1359 \$success;
1360 } @_;
5a964f20
TC
1361 }
1362
87e95b7f
YO
1363Precompilation of the pattern into an internal representation at
1364the moment of qr() avoids a need to recompile the pattern every
1365time a match C</\$pat/> is attempted. (Perl has many other internal
1366optimizations, but none would be triggered in the above example if
1367we did not use qr() operator.)
1368
765fa144 1369Options (specified by the following modifiers) are:
87e95b7f
YO
1370
1371 m Treat string as multiple lines.
1372 s Treat string as single line. (Make . match a newline)
1373 i Do case-insensitive pattern matching.
1374 x Use extended regular expressions.
1375 p When matching preserve a copy of the matched string so
1376 that \${^PREMATCH}, \${^MATCH}, \${^POSTMATCH} will be defined.
1377 o Compile pattern only once.
b5c53fdc
KW
1378 a ASCII-restrict: Use ASCII for \d, \s, \w; specifying two a's
1379 further restricts /i matching so that no ASCII character will
1380 match a non-ASCII one
18509dec
KW
1381 l Use the locale
1382 u Use Unicode rules
b6fa137b 1383 d Use Unicode or native charset, as in 5.12 and earlier
87e95b7f
YO
1384
1385If a precompiled pattern is embedded in a larger pattern then the effect
c543c01b 1386of "msixpluad" will be propagated appropriately. The effect the "o"
87e95b7f
YO
1387modifier has is not propagated, being restricted to those patterns
1388explicitly using it.
1389
b6fa137b 1390The last four modifiers listed above, added in Perl 5.14,
18509dec
KW
1391control the character set semantics, but C</a> is the only one you are likely
1392to want to specify explicitly; the other three are selected
1393automatically by various pragmas.
da392a17 1394
87e95b7f 1395See L<perlre> for additional information on valid syntax for STRING, and
5e2aa8f5 1396for a detailed look at the semantics of regular expressions. In
f6050459
KW
1397particular, all the modifiers execpt C</o> are further explained in
1398L<perlre/Modifiers>. C</o> is described in the next section.
a0d0e21e 1399
b6fa137b 1400=item m/PATTERN/msixpodualgc
89d205f2
YO
1401X<m> X<operator, match>
1402X<regexp, options> X<regexp> X<regex, options> X<regex>
01c6f5f4 1403X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c>
a0d0e21e 1404
b6fa137b 1405=item /PATTERN/msixpodualgc
a0d0e21e 1406
5a964f20 1407Searches a string for a pattern match, and in scalar context returns
19799a22
GS
1408true if it succeeds, false if it fails. If no string is specified
1409via the C<=~> or C<!~> operator, the \$_ string is searched. (The
1410string specified with C<=~> need not be an lvalue--it may be the
1411result of an expression evaluation, but remember the C<=~> binds
a0d0e21e 1413
f6050459 1414Options are as described in C<qr//> above; in addition, the following match
01c6f5f4 1415process modifiers are available:
a0d0e21e 1416
950b09ed
KW
1417 g Match globally, i.e., find all occurrences.
1418 c Do not reset search position on a failed match when /g is in effect.
a0d0e21e 1419
725a61d7 1420If "/" is the delimiter then the initial C<m> is optional. With the C<m>
c543c01b 1421you can use any pair of non-whitespace (ASCII) characters
725a61d7
Z
1422as delimiters. This is particularly useful for matching path names
1423that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
1424the delimiter, then a match-only-once rule applies,
1425described in C<m?PATTERN?> below.
19799a22 1426If "'" is the delimiter, no interpolation is performed on the PATTERN.
ed02a3bf
DN
1427When using a character valid in an identifier, whitespace is required
1428after the C<m>.
a0d0e21e 1429
532c9e80
KW
1430PATTERN may contain variables, which will be interpolated
1431every time the pattern search is evaluated, except
1f247705
GS
1432for when the delimiter is a single quote. (Note that C<\$(>, C<\$)>, and
1433C<\$|> are not interpolated because they look like end-of-string tests.)
532c9e80
KW
1434Perl will not recompile the pattern unless an interpolated
1435variable that it contains changes. You can force Perl to skip the
1436test and never recompile by adding a C</o> (which stands for "once")
1437after the trailing delimiter.
1438Once upon a time, Perl would recompile regular expressions
1439unnecessarily, and this modifier was useful to tell it not to do so, in the
1440interests of speed. But now, the only reasons to use C</o> are either:
1441
1442=over
1443
1444=item 1
1445
1446The variables are thousands of characters long and you know that they
1447don't change, and you need to wring out the last little bit of speed by
1448having Perl skip testing for that. (There is a maintenance penalty for
1449doing this, as mentioning C</o> constitutes a promise that you won't
18509dec 1450change the variables in the pattern. If you do change them, Perl won't
532c9e80
KW
1451even notice.)
1452
1453=item 2
1454
1455you want the pattern to use the initial values of the variables
1456regardless of whether they change or not. (But there are saner ways
1457of accomplishing this than using C</o>.)
1458
1459=back
a0d0e21e 1460
18509dec
KW
1461The bottom line is that using C</o> is almost never a good idea.
1462
e9d89077
DN
1463=item The empty pattern //
1464
5a964f20 1465If the PATTERN evaluates to the empty string, the last
d65afb4b 1466I<successfully> matched regular expression is used instead. In this
c543c01b 1467case, only the C<g> and C<c> flags on the empty pattern are honored;
d65afb4b
HS
1468the other flags are taken from the original pattern. If no match has
1469previously succeeded, this will (silently) act instead as a genuine
1470empty pattern (which will always match).
a0d0e21e 1471
89d205f2
YO
1472Note that it's possible to confuse Perl into thinking C<//> (the empty
1473regex) is really C<//> (the defined-or operator). Perl is usually pretty
1475C<\$a///> (is that C<(\$a) / (//)> or C<\$a // />?) and C<print \$fh //>
1476(C<print \$fh(//> or C<print(\$fh //>?). In all of these examples, Perl
1477will assume you meant defined-or. If you meant the empty regex, just
1478use parentheses or spaces to disambiguate, or even prefix the empty
c963b151
BD
1479regex with an C<m> (so C<//> becomes C<m//>).
1480
e9d89077
DN
1481=item Matching in list context
1482
19799a22 1483If the C</g> option is not used, C<m//> in list context returns a
a0d0e21e 1484list consisting of the subexpressions matched by the parentheses in the
f7e33566
GS
1485pattern, i.e., (C<\$1>, C<\$2>, C<\$3>...). (Note that here C<\$1> etc. are
1486also set, and that this differs from Perl 4's behavior.) When there are
1487no parentheses in the pattern, the return value is the list C<(1)> for
1488success. With or without parentheses, an empty list is returned upon
1489failure.
a0d0e21e
LW
1490
1491Examples:
1492
c543c01b
TC
1493 open(TTY, "+>/dev/tty")
1494 || die "can't access /dev/tty: \$!";
1495
a0d0e21e
LW
1496 <TTY> =~ /^y/i && foo(); # do foo if desired
1497
1498 if (/Version: *([0-9.]*)/) { \$version = \$1; }
1499
1500 next if m#^/usr/spool/uucp#;
1501
1502 # poor man's grep
1503 \$arg = shift;
1504 while (<>) {
c543c01b 1505 print if /\$arg/o; # compile only once (no longer needed!)
a0d0e21e
LW
1506 }
1507
1508 if ((\$F1, \$F2, \$Etc) = (\$foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
1509
1510This last example splits \$foo into the first two words and the
5f05dabc 1511remainder of the line, and assigns those three fields to \$F1, \$F2, and
c543c01b
TC
1512\$Etc. The conditional is true if any variables were assigned; that is,
1513if the pattern matched.
a0d0e21e 1514
19799a22 1515The C</g> modifier specifies global pattern matching--that is,
3dd93342 1516matching as many times as possible within the string. How it behaves
1517depends on the context. In list context, it returns a list of the
19799a22 1518substrings matched by any capturing parentheses in the regular
3dd93342 1519expression. If there are no parentheses, it returns a list of all
19799a22
GS
1520the matched strings, as if there were parentheses around the whole
1521pattern.
a0d0e21e 1522
7e86de3e 1523In scalar context, each execution of C<m//g> finds the next match,
19799a22 1524returning true if it matches, and false if there is no further match.
3dd93342 1525The position after the last match can be read or set using the C<pos()>
1526function; see L<perlfunc/pos>. A failed match normally resets the
7e86de3e 1527search position to the beginning of the string, but you can avoid that
3dd93342 1528by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
7e86de3e 1529string also resets the search position.
c90c0ff4 1530
e9d89077
DN
1531=item \G assertion
1532
c90c0ff4 1533You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
3dd93342 1534zero-width assertion that matches the exact position where the
1535previous C<m//g>, if any, left off. Without the C</g> modifier, the
1536C<\G> assertion still anchors at C<pos()> as it was at the start of
1537the operation (see L<perlfunc/pos>), but the match is of course only
1538attempted once. Using C<\G> without C</g> on a target string that has
1539not previously had a C</g> match applied to it is the same as using
1540the C<\A> assertion to match the beginning of the string. Note also
1541that, currently, C<\G> is only properly supported when anchored at the
1542very beginning of the pattern.
c90c0ff4 1543
1544Examples:
a0d0e21e
LW
1545
1546 # list context
1547 (\$one,\$five,\$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1548
1549 # scalar context
c543c01b
TC
1550 local \$/ = "";
1551 while (\$paragraph = <>) {
1552 while (\$paragraph =~ /\p{Ll}['")]*[.!?]+['")]*\s/g) {
19799a22 1553 \$sentences++;
a0d0e21e
LW
1554 }
1555 }
c543c01b
TC
1556 say \$sentences;
1557
1558Here's another way to check for sentences in a paragraph:
1559
1560 my \$sentence_rx = qr{
1561 (?: (?<= ^ ) | (?<= \s ) ) # after start-of-string or whitespace
1562 \p{Lu} # capital letter
1563 .*? # a bunch of anything
1564 (?<= \S ) # that ends in non-whitespace
1565 (?<! \b [DMS]r ) # but isn't a common abbreviation
1566 (?<! \b Mrs )
1567 (?<! \b Sra )
1568 (?<! \b St )
1569 [.?!] # followed by a sentence ender
1570 (?= \$ | \s ) # in front of end-of-string or whitespace
1571 }sx;
1572 local \$/ = "";
1573 while (my \$paragraph = <>) {
1574 say "NEW PARAGRAPH";
1575 my \$count = 0;
1576 while (\$paragraph =~ /(\$sentence_rx)/g) {
1577 printf "\tgot sentence %d: <%s>\n", ++\$count, \$1;
1578 }
1579 }
1580
1581Here's how to use C<m//gc> with C<\G>:
a0d0e21e 1582
137443ea 1583 \$_ = "ppooqppqq";
44a8e56a 1584 while (\$i++ < 2) {
1585 print "1: '";
c90c0ff4 1586 print \$1 while /(o)/gc; print "', pos=", pos, "\n";
44a8e56a 1587 print "2: '";
c90c0ff4 1588 print \$1 if /\G(q)/gc; print "', pos=", pos, "\n";
44a8e56a 1589 print "3: '";
c90c0ff4 1590 print \$1 while /(p)/gc; print "', pos=", pos, "\n";
44a8e56a 1591 }
5d43e42d 1592 print "Final: '\$1', pos=",pos,"\n" if /\G(.)/;
44a8e56a 1593
1594The last example should print:
1595
1596 1: 'oo', pos=4
137443ea 1597 2: 'q', pos=5
44a8e56a 1598 3: 'pp', pos=7
1599 1: '', pos=7
137443ea 1600 2: 'q', pos=8
1601 3: '', pos=8
5d43e42d
DC
1602 Final: 'q', pos=8
1603
1604Notice that the final match matched C<q> instead of C<p>, which a match
1605without the C<\G> anchor would have done. Also note that the final match
ac036724 1606did not update C<pos>. C<pos> is only updated on a C</g> match. If the
c543c01b
TC
1607final match did indeed match C<p>, it's a good bet that you're running a
1608very old (pre-5.6.0) version of Perl.
44a8e56a 1609
c90c0ff4 1610A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
e7ea3e70 1611combine several regexps like this to process a string part-by-part,
c90c0ff4 1612doing different actions depending on which regexp matched. Each
1613regexp tries to match where the previous one leaves off.
e7ea3e70 1614
3fe9a6f1 1615 \$_ = <<'EOL';
950b09ed 1616 \$url = URI::URL->new( "http://example.com/" ); die if \$url eq "xXx";
3fe9a6f1 1617 EOL
c543c01b
TC
1618
1619 LOOP: {
950b09ed 1620 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
c543c01b
TC
1621 print(" lowercase"), redo LOOP if /\G\p{Ll}+\b[,.;]?\s*/gc;
1622 print(" UPPERCASE"), redo LOOP if /\G\p{Lu}+\b[,.;]?\s*/gc;
1623 print(" Capitalized"), redo LOOP if /\G\p{Lu}\p{Ll}+\b[,.;]?\s*/gc;
1624 print(" MiXeD"), redo LOOP if /\G\pL+\b[,.;]?\s*/gc;
1625 print(" alphanumeric"), redo LOOP if /\G[\p{Alpha}\pN]+\b[,.;]?\s*/gc;
1626 print(" line-noise"), redo LOOP if /\G\W+/gc;
950b09ed 1627 print ". That's all!\n";
c543c01b 1628 }
e7ea3e70
IZ
1629
1630Here is the output (split into several lines):
1631
c543c01b
TC
1632 line-noise lowercase line-noise UPPERCASE line-noise UPPERCASE
1633 line-noise lowercase line-noise lowercase line-noise lowercase
1634 lowercase line-noise lowercase lowercase line-noise lowercase
1635 lowercase line-noise MiXeD line-noise. That's all!
44a8e56a 1636
c543c01b 1637=item m?PATTERN?msixpodualgc
725a61d7 1638X<?> X<operator, match-once>
87e95b7f 1639
c543c01b 1640=item ?PATTERN?msixpodualgc
55d389e7 1641
725a61d7
Z
1642This is just like the C<m/PATTERN/> search, except that it matches
1643only once between calls to the reset() operator. This is a useful
87e95b7f 1644optimization when you want to see only the first occurrence of
ceb131e8 1645something in each file of a set of files, for instance. Only C<m??>
87e95b7f
YO
1646patterns local to the current package are reset.
1647
1648 while (<>) {
ceb131e8 1649 if (m?^\$?) {
87e95b7f
YO
1650 # blank line between header and body
1651 }
1652 } continue {
725a61d7 1653 reset if eof; # clear m?? status for next file
87e95b7f
YO
1654 }
1655
c543c01b
TC
1656Another example switched the first "latin1" encoding it finds
1657to "utf8" in a pod file:
1658
1659 s//utf8/ if m? ^ =encoding \h+ \K latin1 ?x;
1660
1661The match-once behavior is controlled by the match delimiter being
725a61d7
Z
1662C<?>; with any other delimiter this is the normal C<m//> operator.
1663
1664For historical reasons, the leading C<m> in C<m?PATTERN?> is optional,
1665but the resulting C<?PATTERN?> syntax is deprecated, will warn on
c543c01b
TC
1666usage and might be removed from a future stable release of Perl (without
1667further notice!).
87e95b7f 1668
b6fa137b 1669=item s/PATTERN/REPLACEMENT/msixpodualgcer
87e95b7f 1670X<substitute> X<substitution> X<replace> X<regexp, replace>
4f4d7508 1671X<regexp, substitute> X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c> X</e> X</r>
87e95b7f
YO
1672
1673Searches a string for a pattern, and if found, replaces that pattern
1674with the replacement text and returns the number of substitutions
1675made. Otherwise it returns false (specifically, the empty string).
1676
c543c01b 1677If the C</r> (non-destructive) option is used then it runs the
679563bb
KW
1678substitution on a copy of the string and instead of returning the
1679number of substitutions, it returns the copy whether or not a
c543c01b
TC
1680substitution occurred. The original string is never changed when
1681C</r> is used. The copy will always be a plain string, even if the
1682input is an object or a tied variable.
4f4d7508 1683
87e95b7f 1684If no string is specified via the C<=~> or C<!~> operator, the C<\$_>
c543c01b
TC
1685variable is searched and modified. Unless the C</r> option is used,
1686the string specified must be a scalar variable, an array element, a
1687hash element, or an assignment to one of those; that is, some sort of
1688scalar lvalue.
87e95b7f
YO
1689
1690If the delimiter chosen is a single quote, no interpolation is
1691done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1692PATTERN contains a \$ that looks like a variable rather than an
1693end-of-string test, the variable will be interpolated into the pattern
1694at run-time. If you want the pattern compiled only once the first time
1695the variable is interpolated, use the C</o> option. If the pattern
1696evaluates to the empty string, the last successfully executed regular
1697expression is used instead. See L<perlre> for further explanation on these.
87e95b7f
YO
1698
1699Options are as with m// with the addition of the following replacement
1700specific options:
1701
1702 e Evaluate the right side as an expression.
4f4d7508
DC
1703 ee Evaluate the right side as a string then eval the result.
1704 r Return substitution and leave the original string untouched.
87e95b7f 1705
ed02a3bf
DN
1706Any non-whitespace delimiter may replace the slashes. Add space after
1707the C<s> when using a character allowed in identifiers. If single quotes
1708are used, no interpretation is done on the replacement string (the C</e>
1709modifier overrides this, however). Unlike Perl 4, Perl 5 treats backticks
1710as normal delimiters; the replacement text is not evaluated as a command.
1711If the PATTERN is delimited by bracketing quotes, the REPLACEMENT has
1712its own pair of quotes, which may or may not be bracketing quotes, e.g.,
87e95b7f
YO
1713C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
1714replacement portion to be treated as a full-fledged Perl expression
1715and evaluated right then and there. It is, however, syntax checked at
1716compile-time. A second C<e> modifier will cause the replacement portion
1717to be C<eval>ed before being run as a Perl expression.
1718
1719Examples:
1720
1721 s/\bgreen\b/mauve/g; # don't change wintergreen
1722
1723 \$path =~ s|/usr/bin|/usr/local/bin|;
1724
1726
1727 (\$foo = \$bar) =~ s/this/that/; # copy first, then change
4f4d7508
DC
1728 (\$foo = "\$bar") =~ s/this/that/; # convert to string, copy, then change
1729 \$foo = \$bar =~ s/this/that/r; # Same as above using /r
1730 \$foo = \$bar =~ s/this/that/r
1731 =~ s/that/the other/r; # Chained substitutes using /r
1732 @foo = map { s/this/that/r } @bar # /r is very useful in maps
87e95b7f
YO
1733
1734 \$count = (\$paragraph =~ s/Mister\b/Mr./g); # get change-count
1735
1736 \$_ = 'abc123xyz';
1737 s/\d+/\$&*2/e; # yields 'abc246xyz'
1738 s/\d+/sprintf("%5d",\$&)/e; # yields 'abc 246xyz'
1739 s/\w/\$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1740
1741 s/%(.)/\$percent{\$1}/g; # change percent escapes; no /e
1742 s/%(.)/\$percent{\$1} || \$&/ge; # expr now, so /e
1743 s/^=(\w+)/pod(\$1)/ge; # use function call
1744
4f4d7508
DC
1745 \$_ = 'abc123xyz';
1746 \$a = s/abc/def/r; # \$a is 'def123xyz' and
1747 # \$_ remains 'abc123xyz'.
1748
87e95b7f
YO
1749 # expand variables in \$_, but dynamics only, using
1750 # symbolic dereferencing
1751 s/\\$(\w+)/\${\$1}/g;
1752
1753 # Add one to the value of any numbers in the string
1754 s/(\d+)/1 + \$1/eg;
1755
c543c01b
TC
1756 # Titlecase words in the last 30 characters only
1757 substr(\$str, -30) =~ s/\b(\p{Alpha}+)\b/\u\L\$1/g;
1758
87e95b7f
YO
1759 # This will expand any embedded scalar variable
1760 # (including lexicals) in \$_ : First \$1 is interpolated
1761 # to the variable name, and then evaluated
1762 s/(\\$\w+)/\$1/eeg;
1763
1764 # Delete (most) C comments.
1765 \$program =~ s {
1766 /\* # Match the opening delimiter.
1767 .*? # Match a minimal number of characters.
1768 \*/ # Match the closing delimiter.
1769 } []gsx;
1770
1771 s/^\s*(.*?)\s*\$/\$1/; # trim whitespace in \$_, expensively
1772
1773 for (\$variable) { # trim whitespace in \$variable, cheap
1774 s/^\s+//;
1775 s/\s+\$//;
1776 }
1777
1778 s/([^ ]*) *([^ ]*)/\$2 \$1/; # reverse 1st two fields
1779
1780Note the use of \$ instead of \ in the last example. Unlike
1781B<sed>, we use the \<I<digit>> form in only the left hand side.
1782Anywhere else it's \$<I<digit>>.
1783
1784Occasionally, you can't use just a C</g> to get all the changes
1785to occur that you might want. Here are two common cases:
1786
1787 # put commas in the right places in an integer
1788 1 while s/(\d)(\d\d\d)(?!\d)/\$1,\$2/g;
1789
1790 # expand tabs to 8-column spacing
1791 1 while s/\t+/' ' x (length(\$&)*8 - length(\$`)%8)/e;
1792
1c424184
FC
1793C<s///le> is treated as a substitution followed by the C<le> operator, not
1794the C</le> flags. This may change in a future version of Perl. It
1795produces a warning if warnings are enabled. To disambiguate, use a space
1796or change the order of the flags:
1797
1798 s/foo/bar/ le 5; # "le" infix operator
1799 s/foo/bar/el; # "e" and "l" flags
1800
87e95b7f
YO
1801=back
1802
1804X<operator, quote-like>
1805
01c6f5f4
RGS
1806=over 4
1807
a0d0e21e 1808=item q/STRING/
5d44bfff 1809X<q> X<quote, single> X<'> X<''>
a0d0e21e 1810
5d44bfff 1811=item 'STRING'
a0d0e21e 1812
19799a22 1813A single-quoted, literal string. A backslash represents a backslash
68dc0745 1814unless followed by the delimiter or another backslash, in which case
1815the delimiter or backslash is interpolated.
a0d0e21e
LW
1816
1817 \$foo = q!I said, "You said, 'She said it.'"!;
1818 \$bar = q('This is it.');
68dc0745 1819 \$baz = '\n'; # a two-character string
a0d0e21e
LW
1820
1821=item qq/STRING/
d74e8afc 1822X<qq> X<quote, double> X<"> X<"">
a0d0e21e
LW
1823
1824=item "STRING"
1825
1826A double-quoted, interpolated string.
1827
1828 \$_ .= qq
1829 (*** The previous line contains the naughty word "\$1".\n)
19799a22 1830 if /\b(tcl|java|python)\b/i; # :-)
68dc0745 1831 \$baz = "\n"; # a one-character string
a0d0e21e
LW
1832
1833=item qx/STRING/
d74e8afc 1834X<qx> X<`> X<``> X<backtick>
a0d0e21e
LW
1835
1836=item `STRING`
1837
43dd4d21
JH
1838A string which is (possibly) interpolated and then executed as a
1839system command with C</bin/sh> or its equivalent. Shell wildcards,
1840pipes, and redirections will be honored. The collected standard
1841output of the command is returned; standard error is unaffected. In
1842scalar context, it comes back as a single (potentially multi-line)
1843string, or undef if the command failed. In list context, returns a
1844list of lines (however you've defined lines with \$/ or
1845\$INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
5a964f20
TC
1846
1847Because backticks do not affect standard error, use shell file descriptor
1848syntax (assuming the shell supports this) if you care to address this.
1849To capture a command's STDERR and STDOUT together:
a0d0e21e 1850
5a964f20
TC
1851 \$output = `cmd 2>&1`;
1852
1853To capture a command's STDOUT but discard its STDERR:
1854
1855 \$output = `cmd 2>/dev/null`;
1856
1857To capture a command's STDERR but discard its STDOUT (ordering is
1858important here):
1859
1860 \$output = `cmd 2>&1 1>/dev/null`;
1861
1862To exchange a command's STDOUT and STDERR in order to capture the STDERR
1863but leave its STDOUT to come out the old STDERR:
1864
1865 \$output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1866
1867To read both a command's STDOUT and its STDERR separately, it's easiest
2359510d
SD
1868to redirect them separately to files, and then read from those files
1869when the program is done:
5a964f20 1870
2359510d 1871 system("program args 1>program.stdout 2>program.stderr");
5a964f20 1872
30398227
SP
1873The STDIN filehandle used by the command is inherited from Perl's STDIN.
1874For example:
1875
c543c01b
TC
1876 open(SPLAT, "stuff") || die "can't open stuff: \$!";
1877 open(STDIN, "<&SPLAT") || die "can't dupe SPLAT: \$!";
40bbb707 1878 print STDOUT `sort`;
30398227 1879
40bbb707 1880will print the sorted contents of the file named F<"stuff">.
30398227 1881
5a964f20
TC
1882Using single-quote as a delimiter protects the command from Perl's
1883double-quote interpolation, passing it on to the shell instead:
1884
1885 \$perl_info = qx(ps \$\$); # that's Perl's \$\$
1886 \$shell_info = qx'ps \$\$'; # that's the new shell's \$\$
1887
19799a22 1888How that string gets evaluated is entirely subject to the command
5a964f20
TC
1889interpreter on your system. On most platforms, you will have to protect
1890shell metacharacters if you want them treated literally. This is in
1891practice difficult to do, as it's unclear how to escape which characters.
1892See L<perlsec> for a clean and safe example of a manual fork() and exec()
1893to emulate backticks safely.
a0d0e21e 1894
bb32b41a
GS
1895On some platforms (notably DOS-like ones), the shell may not be
1896capable of dealing with multiline commands, so putting newlines in
1897the string may not get you what you want. You may be able to evaluate
1898multiple commands in a single line by separating them with the command
1899separator character, if your shell supports that (e.g. C<;> on many Unix
1900shells; C<&> on the Windows NT C<cmd> shell).
1901
0f897271
GS
1902Beginning with v5.6.0, Perl will attempt to flush all files opened for
1903output before starting the child process, but this may not be supported
1904on some platforms (see L<perlport>). To be safe, you may need to set
1905C<\$|> (\$AUTOFLUSH in English) or call the C<autoflush()> method of
1906C<IO::Handle> on any open handles.
1907
bb32b41a
GS
1908Beware that some command shells may place restrictions on the length
1909of the command line. You must ensure your strings don't exceed this
1910limit after any necessary interpolations. See the platform-specific
1912
5a964f20
TC
1913Using this operator can lead to programs that are difficult to port,
1914because the shell commands called vary between systems, and may in
1915fact not be present at all. As one example, the C<type> command under
1916the POSIX shell is very different from the C<type> command under DOS.
1917That doesn't mean you should go out of your way to avoid backticks
1918when they're the right way to get something done. Perl was made to be
1919a glue language, and one of the things it glues together is commands.
1920Just understand what you're getting yourself into.
bb32b41a 1921
da87341d 1922See L</"I/O Operators"> for more discussion.
a0d0e21e 1923
945c54fd 1924=item qw/STRING/
d74e8afc 1925X<qw> X<quote, list> X<quote, words>
945c54fd
JH
1926
1927Evaluates to a list of the words extracted out of STRING, using embedded
1928whitespace as the word delimiters. It can be understood as being roughly
1929equivalent to:
1930
c543c01b 1931 split(" ", q/STRING/);
945c54fd 1932
efb1e162
CW
1933the differences being that it generates a real list at compile time, and
1934in scalar context it returns the last element in the list. So
945c54fd
JH
1935this expression:
1936
1937 qw(foo bar baz)
1938
1939is semantically equivalent to the list:
1940
c543c01b 1941 "foo", "bar", "baz"
945c54fd
JH
1942
1943Some frequently seen examples:
1944
1945 use POSIX qw( setlocale localeconv )
1946 @EXPORT = qw( foo bar baz );
1947
1948A common mistake is to try to separate the words with comma or to
1949put comments into a multi-line C<qw>-string. For this reason, the
89d205f2 1950C<use warnings> pragma and the B<-w> switch (that is, the C<\$^W> variable)
945c54fd
JH
1951produces warnings if the STRING contains the "," or the "#" character.
1952
8ff32507 1953=item tr/SEARCHLIST/REPLACEMENTLIST/cdsr
d74e8afc 1954X<tr> X<y> X<transliterate> X</c> X</d> X</s>
a0d0e21e 1955
8ff32507 1956=item y/SEARCHLIST/REPLACEMENTLIST/cdsr
a0d0e21e 1957
2c268ad5 1958Transliterates all occurrences of the characters found in the search list
a0d0e21e
LW
1959with the corresponding character in the replacement list. It returns
1960the number of characters replaced or deleted. If no string is
c543c01b
TC
1961specified via the C<=~> or C<!~> operator, the \$_ string is transliterated.
1962
1963If the C</r> (non-destructive) option is present, a new copy of the string
1964is made and its characters transliterated, and this copy is returned no
1965matter whether it was modified or not: the original string is always
1966left unchanged. The new copy is always a plain string, even if the input
1967string is an object or a tied variable.
c543c01b
TC
1969Unless the C</r> option is used, the string specified with C<=~> must be a
1970scalar variable, an array element, a hash element, or an assignment to one
1971of those; in other words, an lvalue.
8ff32507 1972
89d205f2 1973A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
2c268ad5 1974does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
54310121 1975For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1976SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
c543c01b
TC
1977its own pair of quotes, which may or may not be bracketing quotes;
1978for example, C<tr[aeiouy][yuoiea]> or C<tr(+\-*/)/ABCD/>.
1979
1980Note that C<tr> does B<not> do regular expression character classes such as
1981C<\d> or C<\pL>. The C<tr> operator is not equivalent to the tr(1)
1982utility. If you want to map strings between lower/upper cases, see
1983L<perlfunc/lc> and L<perlfunc/uc>, and in general consider using the C<s>
1984operator if you need regular expressions. The C<\U>, C<\u>, C<\L>, and
1985C<\l> string-interpolation escapes on the right side of a substitution
1986operator will perform correct case-mappings, but C<tr[a-z][A-Z]> will not
1987(except sometimes on legacy 7-bit data).
cc255d5f 1988
JH
1989Note also that the whole range idea is rather unportable between
1990character sets--and even within character sets they may cause results
1991you probably didn't expect. A sound principle is to use only ranges
1992that begin from and end at either alphabets of equal case (a-e, A-E),
1993or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1994character sets in full.
1995
a0d0e21e
LW
1996Options:
1997
1998 c Complement the SEARCHLIST.
1999 d Delete found but unreplaced characters.
2000 s Squash duplicate replaced characters.
8ff32507
FC
2001 r Return the modified string and leave the original string
2002 untouched.
a0d0e21e 2003
19799a22
GS
2004If the C</c> modifier is specified, the SEARCHLIST character set
2005is complemented. If the C</d> modifier is specified, any characters
2007(Note that this is slightly more flexible than the behavior of some
2008B<tr> programs, which delete anything they find in the SEARCHLIST,
2009period.) If the C</s> modifier is specified, sequences of characters
2010that were transliterated to the same character are squashed down
2011to a single instance of the character.
a0d0e21e
LW
2012
2013If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
2014exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
2015than the SEARCHLIST, the final character is replicated till it is long
5a964f20 2016enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
a0d0e21e
LW
2017This latter is useful for counting characters in a class or for
2018squashing character sequences in a class.
2019
2020Examples:
2021
c543c01b 2022 \$ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case ASCII
a0d0e21e
LW
2023
2024 \$cnt = tr/*/*/; # count the stars in \$_
2025
2026 \$cnt = \$sky =~ tr/*/*/; # count the stars in \$sky
2027
2028 \$cnt = tr/0-9//; # count the digits in \$_
2029
2030 tr/a-zA-Z//s; # bookkeeper -> bokeper
2031
2032 (\$HOST = \$host) =~ tr/a-z/A-Z/;
c543c01b 2033 \$HOST = \$host =~ tr/a-z/A-Z/r; # same thing
8ff32507 2034
c543c01b 2035 \$HOST = \$host =~ tr/a-z/A-Z/r # chained with s///r
8ff32507 2036 =~ s/:/ -p/r;
a0d0e21e
LW
2037
2038 tr/a-zA-Z/ /cs; # change non-alphas to single space
2039
8ff32507
FC
2040 @stripped = map tr/a-zA-Z/ /csr, @original;
2041 # /r with map
2042
a0d0e21e 2043 tr [\200-\377]
c543c01b 2044 [\000-\177]; # wickedly delete 8th bit
a0d0e21e 2045
19799a22
GS
2046If multiple transliterations are given for a character, only the
2047first one is used:
748a9306
LW
2048
2049 tr/AAA/XYZ/
2050
2c268ad5 2051will transliterate any A to X.
748a9306 2052
19799a22 2053Because the transliteration table is built at compile time, neither
a0d0e21e 2054the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
19799a22
GS
2055interpolation. That means that if you want to use variables, you
2056must use an eval():
a0d0e21e
LW
2057
2058 eval "tr/\$oldlist/\$newlist/";
2059 die \$@ if \$@;
2060
2061 eval "tr/\$oldlist/\$newlist/, 1" or die \$@;
2062
7e3b091d 2063=item <<EOF
d74e8afc 2064X<here-doc> X<heredoc> X<here-document> X<<< << >>>
7e3b091d
DA
2065
2066A line-oriented form of quoting is based on the shell "here-document"
2067syntax. Following a C<< << >> you specify a string to terminate
2068the quoted material, and all lines following the current line down to
89d205f2
YO
2069the terminating string are the value of the item.
2070
2071The terminating string may be either an identifier (a word), or some
2072quoted text. An unquoted identifier works like double quotes.
2073There may not be a space between the C<< << >> and the identifier,
2074unless the identifier is explicitly quoted. (If you put a space it
2075will be treated as a null identifier, which is valid, and matches the
2076first empty line.) The terminating string must appear by itself
2077(unquoted and with no surrounding whitespace) on the terminating line.
2078
2079If the terminating string is quoted, the type of quotes used determine
2080the treatment of the text.
2081
2082=over 4
2083
2084=item Double Quotes
2085
2086Double quotes indicate that the text will be interpolated using exactly
2087the same rules as normal double quoted strings.
7e3b091d
DA
2088
2089 print <<EOF;
2090 The price is \$Price.
2091 EOF
2092
2093 print << "EOF"; # same as above
2094 The price is \$Price.
2095 EOF
2096
89d205f2
YO
2097
2098=item Single Quotes
2099
2100Single quotes indicate the text is to be treated literally with no
2101interpolation of its content. This is similar to single quoted
2102strings except that backslashes have no special meaning, with C<\\>
2103being treated as two backslashes and not one as they would in every
2104other quoting construct.
2105
c543c01b
TC
2106Just as in the shell, a backslashed bareword following the C<<< << >>>
2107means the same thing as a single-quoted string does:
2108
2109 \$cost = <<'VISTA'; # hasta la ...
2110 That'll be \$10 please, ma'am.
2111 VISTA
2112
2113 \$cost = <<\VISTA; # Same thing!
2114 That'll be \$10 please, ma'am.
2115 VISTA
2116
89d205f2
YO
2117This is the only form of quoting in perl where there is no need
2118to worry about escaping content, something that code generators
2119can and do make good use of.
2120
2121=item Backticks
2122
2123The content of the here doc is treated just as it would be if the
2124string were embedded in backticks. Thus the content is interpolated
2125as though it were double quoted and then executed via the shell, with
2126the results of the execution returned.
2127
2128 print << `EOC`; # execute command and get results
7e3b091d 2129 echo hi there
7e3b091d
DA
2130 EOC
2131
89d205f2
YO
2132=back
2133
2134It is possible to stack multiple here-docs in a row:
2135
7e3b091d
DA
2136 print <<"foo", <<"bar"; # you can stack them
2137 I said foo.
2138 foo
2139 I said bar.
2140 bar
2141
2142 myfunc(<< "THIS", 23, <<'THAT');
2143 Here's a line
2144 or two.
2145 THIS
2146 and here's another.
2147 THAT
2148
2149Just don't forget that you have to put a semicolon on the end
2150to finish the statement, as Perl doesn't know you're not going to
2151try to do this:
2152
2153 print <<ABC
2154 179231
2155 ABC
2156 + 20;
2157
872d7e53
TS
2158If you want to remove the line terminator from your here-docs,
2159use C<chomp()>.
2160
2161 chomp(\$string = <<'END');
2162 This is a string.
2163 END
2164
2165If you want your here-docs to be indented with the rest of the code,
2166you'll need to remove leading whitespace from each line manually:
7e3b091d
DA
2167
2168 (\$quote = <<'FINIS') =~ s/^\s+//gm;
89d205f2 2169 The Road goes ever on and on,
7e3b091d
DA
2170 down from the door where it began.
2171 FINIS
2172
2173If you use a here-doc within a delimited construct, such as in C<s///eg>,
2174the quoted material must come on the lines following the final delimiter.
2176
2177 s/this/<<E . 'that'
2178 the other
2179 E
2180 . 'more '/eg;
2181
2182you have to write
2183
89d205f2
YO
2184 s/this/<<E . 'that'
2185 . 'more '/eg;
2186 the other
2187 E
7e3b091d
DA
2188
2189If the terminating identifier is on the last line of the program, you
2190must be sure there is a newline after it; otherwise, Perl will give the
2191warning B<Can't find string terminator "END" anywhere before EOF...>.
2192
c543c01b
TC
2193Additionally, quoting rules for the end-of-string identifier are
2194unrelated to Perl's quoting rules. C<q()>, C<qq()>, and the like are not
89d205f2
YO
2195supported in place of C<''> and C<"">, and the only interpolation is for
2196backslashing the quoting character:
7e3b091d
DA
2197
2198 print << "abc\"def";
2199 testing...
2200 abc"def
2201
2202Finally, quoted strings cannot span multiple lines. The general rule is
2203that the identifier must be a string literal. Stick with that, and you
2204should be safe.
2205
a0d0e21e
LW
2206=back
2207
75e14d17 2208=head2 Gory details of parsing quoted constructs
d74e8afc 2209X<quote, gory details>
75e14d17 2210
19799a22
GS
2211When presented with something that might have several different
2212interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
2213principle to pick the most probable interpretation. This strategy
2214is so successful that Perl programmers often do not suspect the
2215ambivalence of what they write. But from time to time, Perl's
2216notions differ substantially from what the author honestly meant.
2217
2218This section hopes to clarify how Perl handles quoted constructs.
2219Although the most common reason to learn this is to unravel labyrinthine
2220regular expressions, because the initial steps of parsing are the
2221same for all quoting operators, they are all discussed together.
2222
2223The most important Perl parsing rule is the first one discussed
2224below: when processing a quoted construct, Perl first finds the end
2225of that construct, then interprets its contents. If you understand
2226this rule, you may skip the rest of this section on the first
2228expectations much less frequently than this first one.
2229
2230Some passes discussed below are performed concurrently, but because
2231their results are the same, we consider them individually. For different
2232quoting constructs, Perl performs different numbers of passes, from
6deea57f 2233one to four, but these passes are always performed in the same order.
75e14d17 2234
13a2d996 2235=over 4
75e14d17
IZ
2236
2237=item Finding the end
2238
6deea57f
TS
2239The first pass is finding the end of the quoted construct, where
2240the information about the delimiters is used in parsing.
2241During this search, text between the starting and ending delimiters
2242is copied to a safe location. The text copied gets delimiter-independent.
2243
2244If the construct is a here-doc, the ending delimiter is a line
2245that has a terminating string as the content. Therefore C<<<EOF> is
2246terminated by C<EOF> immediately followed by C<"\n"> and starting
2247from the first column of the terminating line.
2248When searching for the terminating line of a here-doc, nothing
2249is skipped. In other words, lines after the here-doc syntax
2250are compared with the terminating string line by line.
2251
2252For the constructs except here-docs, single characters are used as starting
2253and ending delimiters. If the starting delimiter is an opening punctuation
2254(that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
2255corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
2256If the starting delimiter is an unpaired character like C</> or a closing
2257punctuation, the ending delimiter is same as the starting delimiter.
2258Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
2259C<qq[]> and C<qq]]> constructs.
2260
2261When searching for single-character delimiters, escaped delimiters
2262and C<\\> are skipped. For example, while searching for terminating C</>,
2263combinations of C<\\> and C<\/> are skipped. If the delimiters are
2264bracketing, nested pairs are also skipped. For example, while searching
2265for closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
2266and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
2267However, when backslashes are used as the delimiters (like C<qq\\> and
2268C<tr\\\>), nothing is skipped.
2269During the search for the end, backslashes that escape delimiters
2270are removed (exactly speaking, they are not copied to the safe location).
75e14d17 2271
19799a22
GS
2272For constructs with three-part delimiters (C<s///>, C<y///>, and
2273C<tr///>), the search is repeated once more.
6deea57f
TS
2274If the first delimiter is not an opening punctuation, three delimiters must
2275be same such as C<s!!!> and C<tr)))>, in which case the second delimiter
2276terminates the left part and starts the right part at once.
b6538e4f 2277If the left part is delimited by bracketing punctuation (that is C<()>,
6deea57f 2278C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
b6538e4f 2279delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespace
6deea57f 2280and comments are allowed between both parts, though the comment must follow
b6538e4f
TC
2281at least one whitespace character; otherwise a character expected as the
2282start of the comment may be regarded as the starting delimiter of the right part.
75e14d17 2283
19799a22
GS
2284During this search no attention is paid to the semantics of the construct.
2285Thus:
75e14d17
IZ
2286
2287 "\$hash{"\$foo/\$bar"}"
2288
2a94b7ce 2289or:
75e14d17 2290
89d205f2 2291 m/
2a94b7ce 2292 bar # NOT a comment, this slash / terminated m//!
75e14d17
IZ
2293 /x
2294
19799a22
GS
2295do not form legal quoted expressions. The quoted part ends on the
2296first C<"> and C</>, and the rest happens to be a syntax error.
2297Because the slash that terminated C<m//> was followed by a C<SPACE>,
2298the example above is not C<m//x>, but rather C<m//> with no C</x>
2299modifier. So the embedded C<#> is interpreted as a literal C<#>.
75e14d17 2300
89d205f2
YO
2301Also no attention is paid to C<\c\> (multichar control char syntax) during
2302this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
2303of C<\/>, and the following C</> is not recognized as a delimiter.
0d594e51
TS
2304Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
2305
75e14d17 2306=item Interpolation
d74e8afc 2307X<interpolation>
75e14d17 2308
19799a22 2309The next step is interpolation in the text obtained, which is now
89d205f2 2310delimiter-independent. There are multiple cases.
75e14d17 2311
13a2d996 2312=over 4
75e14d17 2313
89d205f2 2314=item C<<<'EOF'>
75e14d17
IZ
2315
2316No interpolation is performed.
6deea57f
TS
2317Note that the combination C<\\> is left intact, since escaped delimiters
2318are not available for here-docs.
75e14d17 2319
6deea57f 2320=item C<m''>, the pattern of C<s'''>
89d205f2 2321
6deea57f
TS
2322No interpolation is performed at this stage.
2323Any backslashed sequences including C<\\> are treated at the stage
2324to L</"parsing regular expressions">.
89d205f2 2325
6deea57f 2326=item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
75e14d17 2327
89d205f2 2328The only interpolation is removal of C<\> from pairs of C<\\>.
6deea57f
TS
2329Therefore C<-> in C<tr'''> and C<y'''> is treated literally
2330as a hyphen and no character range is available.
2331C<\1> in the replacement of C<s'''> does not work as C<\$1>.
89d205f2
YO
2332
2333=item C<tr///>, C<y///>
2334
6deea57f
TS
2335No variable interpolation occurs. String modifying combinations for
2336case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
2337The other escape sequences such as C<\200> and C<\t> and backslashed
2338characters such as C<\\> and C<\-> are converted to appropriate literals.
89d205f2
YO
2339The character C<-> is treated specially and therefore C<\-> is treated
2340as a literal C<->.
75e14d17 2341
89d205f2 2342=item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
75e14d17 2343
19799a22
GS
2344C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
2345converted to corresponding Perl constructs. Thus, C<"\$foo\Qbaz\$bar">
2346is converted to C<\$foo . (quotemeta("baz" . \$bar))> internally.
6deea57f
TS
2347The other escape sequences such as C<\200> and C<\t> and backslashed
2348characters such as C<\\> and C<\-> are replaced with appropriate
2349expansions.
2a94b7ce 2350
19799a22
GS
2351Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
2352is interpolated in the usual way. Something like C<"\Q\\E"> has
2353no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
2354result is the same as for C<"\\\\E">. As a general rule, backslashes
2355between C<\Q> and C<\E> may lead to counterintuitive results. So,
2356C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
2357as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2a94b7ce
IZ
2358
2359 \$str = '\t';
2360 return "\Q\$str";
2361
2362may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
2363
19799a22 2364Interpolated scalars and arrays are converted internally to the C<join> and
92d29cee 2365C<.> catenation operations. Thus, C<"\$foo XXX '@arr'"> becomes:
75e14d17 2366
19799a22 2367 \$foo . " XXX '" . (join \$", @arr) . "'";
75e14d17 2368
19799a22 2369All operations above are performed simultaneously, left to right.
75e14d17 2370
19799a22
GS
2371Because the result of C<"\Q STRING \E"> has all metacharacters
2372quoted, there is no way to insert a literal C<\$> or C<@> inside a
2373C<\Q\E> pair. If protected by C<\>, C<\$> will be quoted to became
2374C<"\\\\$">; if not, it is interpreted as the start of an interpolated
2375scalar.
75e14d17 2376
19799a22 2377Note also that the interpolation code needs to make a decision on
89d205f2 2378where the interpolated scalar ends. For instance, whether
35f2feb0 2379C<< "a \$b -> {c}" >> really means:
75e14d17
IZ
2380
2381 "a " . \$b . " -> {c}";
2382
2a94b7ce 2383or:
75e14d17
IZ
2384
2385 "a " . \$b -> {c};
2386
19799a22
GS
2387Most of the time, the longest possible text that does not include
2388spaces between components and which contains matching braces or
2389brackets. because the outcome may be determined by voting based
2390on heuristic estimators, the result is not strictly predictable.
2391Fortunately, it's usually correct for ambiguous cases.
75e14d17 2392
6deea57f 2393=item the replacement of C<s///>
75e14d17 2394
19799a22 2395Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
6deea57f
TS
2396happens as with C<qq//> constructs.
2397
2398It is at this step that C<\1> is begrudgingly converted to C<\$1> in
2399the replacement text of C<s///>, in order to correct the incorrigible
2400I<sed> hackers who haven't picked up the saner idiom yet. A warning
2401is emitted if the C<use warnings> pragma or the B<-w> command-line flag
2402(that is, the C<\$^W> variable) was set.
2403
2404=item C<RE> in C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
2405
cc74c5bd
TS
2406Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\E>,
2407and interpolation happens (almost) as with C<qq//> constructs.
2408
5d03b57c
KW
2409Processing of C<\N{...}> is also done here, and compiled into an intermediate
2410form for the regex compiler. (This is because, as mentioned below, the regex
2411compilation may be done at execution time, and C<\N{...}> is a compile-time
2412construct.)
2413
cc74c5bd
TS
2414However any other combinations of C<\> followed by a character
2415are not substituted but only skipped, in order to parse them
2416as regular expressions at the following step.
6deea57f 2417As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
1749ea0d 2418treated as an array symbol (for example C<@foo>),
6deea57f 2419even though the same text in C<qq//> gives interpolation of C<\c@>.
6deea57f
TS
2420
2421Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
19799a22
GS
2422a C<#>-comment in a C<//x>-regular expression, no processing is
2423performed whatsoever. This is the first step at which the presence
2424of the C<//x> modifier is relevant.
2425
1749ea0d
TS
2426Interpolation in patterns has several quirks: C<\$|>, C<\$(>, C<\$)>, C<@+>
2427and C<@-> are not interpolated, and constructs C<\$var[SOMETHING]> are
2428voted (by several different estimators) to be either an array element
2429or C<\$var> followed by an RE alternative. This is where the notation
19799a22
GS
2430C<\${arr[\$bar]}> comes handy: C</\${arr[0-9]}/> is interpreted as
2431array element C<-9>, not as a regular expression from the variable
2432C<\$arr> followed by a digit, which would be the interpretation of
2433C</\$arr[0-9]/>. Since voting among different estimators may occur,
2434the result is not predictable.
2435
19799a22
GS
2436The lack of processing of C<\\> creates specific restrictions on
2437the post-processed text. If the delimiter is C</>, one cannot get
2438the combination C<\/> into the result of this step. C</> will
2439finish the regular expression, C<\/> will be stripped to C</> on
2440the previous step, and C<\\/> will be left as is. Because C</> is
2441equivalent to C<\/> inside a regular expression, this does not
2442matter unless the delimiter happens to be character special to the
2443RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
2444alphanumeric char, as in:
2a94b7ce
IZ
2445
2446 m m ^ a \s* b mmx;
2447
19799a22 2448In the RE above, which is intentionally obfuscated for illustration, the
6deea57f 2449delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
89d205f2 2450RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
19799a22
GS
2451reason you're encouraged to restrict your delimiters to non-alphanumeric,
2452non-whitespace choices.
75e14d17
IZ
2453
2454=back
2455
19799a22 2456This step is the last one for all constructs except regular expressions,
75e14d17
IZ
2457which are processed further.
2458
6deea57f
TS
2459=item parsing regular expressions
2460X<regexp, parse>
75e14d17 2461
19799a22 2462Previous steps were performed during the compilation of Perl code,
ac036724 2463but this one happens at run time, although it may be optimized to
19799a22 2464be calculated at compile time if appropriate. After preprocessing
6deea57f 2465described above, and possibly after evaluation if concatenation,
19799a22
GS
2466joining, casing translation, or metaquoting are involved, the
2467resulting I<string> is passed to the RE engine for compilation.
2468
2469Whatever happens in the RE engine might be better discussed in L<perlre>,
2470but for the sake of continuity, we shall do so here.
2471
2472This is another step where the presence of the C<//x> modifier is
2473relevant. The RE engine scans the string from left to right and
2474converts it to a finite automaton.
2475
2476Backslashed characters are either replaced with corresponding
2477literal strings (as with C<\{>), or else they generate special nodes
2478in the finite automaton (as with C<\b>). Characters special to the
2479RE engine (such as C<|>) generate corresponding nodes or groups of
2480nodes. C<(?#...)> comments are ignored. All the rest is either
2481converted to literal strings to match, or else is ignored (as is
2482whitespace and C<#>-style comments if C<//x> is present).
2483
2484Parsing of the bracketed character class construct, C<[...]>, is
2485rather different than the rule used for the rest of the pattern.
2486The terminator of this construct is found using the same rules as
2487for finding the terminator of a C<{}>-delimited construct, the only
2488exception being that C<]> immediately following C<[> is treated as
2489though preceded by a backslash. Similarly, the terminator of
2490C<(?{...})> is found using the same rules as for finding the
2491terminator of a C<{}>-delimited construct.
2492
2493It is possible to inspect both the string given to RE engine and the
2494resulting finite automaton. See the arguments C<debug>/C<debugcolor>
2495in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
4a4eefd0 2496switch documented in L<perlrun/"Command Switches">.
75e14d17
IZ
2497
2498=item Optimization of regular expressions
d74e8afc 2499X<regexp, optimization>
75e14d17 2500
7522fed5 2501This step is listed for completeness only. Since it does not change
75e14d17 2502semantics, details of this step are not documented and are subject
19799a22
GS
2503to change without notice. This step is performed over the finite
2504automaton that was generated during the previous pass.
2a94b7ce 2505
19799a22
GS
2506It is at this stage that C<split()> silently optimizes C</^/> to
2507mean C</^/m>.
75e14d17
IZ
2508
2509=back
2510
d74e8afc
ITB
2512X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
2513X<< <> >> X<@ARGV>
a0d0e21e 2514
54310121 2515There are several I/O operators you should know about.
7b8d334a 2517A string enclosed by backticks (grave accents) first undergoes
19799a22
GS
2518double-quote interpolation. It is then interpreted as an external
2519command, and the output of that command is the value of the
e9c56f9b
JH
2520backtick string, like in a shell. In scalar context, a single string
2521consisting of all output is returned. In list context, a list of
2522values is returned, one per line of output. (You can set C<\$/> to use
2523a different line terminator.) The command is executed each time the
2524pseudo-literal is evaluated. The status value of the command is
2525returned in C<\$?> (see L<perlvar> for the interpretation of C<\$?>).
2526Unlike in B<csh>, no translation is done on the return data--newlines
2527remain newlines. Unlike in any of the shells, single quotes do not
2528hide variable names in the command from interpretation. To pass a
2529literal dollar-sign through to the shell you need to hide it with a
2530backslash. The generalized form of backticks is C<qx//>. (Because
2531backticks always undergo shell expansion as well, see L<perlsec> for
2532security concerns.)
d74e8afc 2533X<qx> X<`> X<``> X<backtick> X<glob>
19799a22
GS
2534
2535In scalar context, evaluating a filehandle in angle brackets yields
2536the next line from that file (the newline, if any, included), or
2537C<undef> at end-of-file or on error. When C<\$/> is set to C<undef>
2538(sometimes known as file-slurp mode) and the file is empty, it
2539returns C<''> the first time, followed by C<undef> subsequently.
2540
2541Ordinarily you must assign the returned value to a variable, but
2542there is one situation where an automatic assignment happens. If
2543and only if the input symbol is the only thing inside the conditional
2544of a C<while> statement (even if disguised as a C<for(;;)> loop),
2545the value is automatically assigned to the global variable \$_,
2546destroying whatever was there previously. (This may seem like an
2547odd thing to you, but you'll use the construct in almost every Perl
17b829fa 2548script you write.) The \$_ variable is not implicitly localized.
19799a22
GS
2549You'll have to put a C<local \$_;> before the loop if you want that
2550to happen.
2551
2552The following lines are equivalent:
a0d0e21e 2553
748a9306 2554 while (defined(\$_ = <STDIN>)) { print; }
7b8d334a 2555 while (\$_ = <STDIN>) { print; }
a0d0e21e
LW
2556 while (<STDIN>) { print; }
2557 for (;<STDIN>;) { print; }
748a9306 2558 print while defined(\$_ = <STDIN>);
7b8d334a 2559 print while (\$_ = <STDIN>);
a0d0e21e
LW
2560 print while <STDIN>;
2561
19799a22 2562This also behaves similarly, but avoids \$_ :
7b8d334a 2563
89d205f2 2564 while (my \$line = <STDIN>) { print \$line }
7b8d334a 2565
19799a22
GS
2566In these loop constructs, the assigned value (whether assignment
2567is automatic or explicit) is then tested to see whether it is
2568defined. The defined test avoids problems where line has a string
2569value that would be treated as false by Perl, for example a "" or
2570a "0" with no trailing newline. If you really mean for such values
2571to terminate the loop, they should be tested for explicitly:
7b8d334a
GS
2572
2573 while ((\$_ = <STDIN>) ne '0') { ... }
2574 while (<STDIN>) { last unless \$_; ... }
2575
5ef4d93e 2576In other boolean contexts, C<< <filehandle> >> without an
2577explicit C<defined> test or comparison elicits a warning if the
9f1b1f2d 2578C<use warnings> pragma or the B<-w>
19799a22 2579command-line switch (the C<\$^W> variable) is in effect.
7b8d334a 2580
5f05dabc 2581The filehandles STDIN, STDOUT, and STDERR are predefined. (The
19799a22
GS
2582filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
2583in packages, where they would be interpreted as local identifiers
2584rather than global.) Additional filehandles may be created with
2585the open() function, amongst others. See L<perlopentut> and
2586L<perlfunc/open> for details on this.
d74e8afc 2587X<stdin> X<stdout> X<sterr>
a0d0e21e 2588
35f2feb0 2589If a <FILEHANDLE> is used in a context that is looking for
19799a22
GS
2590a list, a list comprising all input lines is returned, one line per
2591list element. It's easy to grow to a rather large data space this
2592way, so use with care.
a0d0e21e 2593
35f2feb0 2594<FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
35f2feb0
GS
2597The null filehandle <> is special: it can be used to emulate the
2598behavior of B<sed> and B<awk>. Input from <> comes either from
a0d0e21e 2599standard input, or from each file listed on the command line. Here's
35f2feb0 2600how it works: the first time <> is evaluated, the @ARGV array is
5a964f20 2601checked, and if it is empty, C<\$ARGV[0]> is set to "-", which when opened
a0d0e21e
LW
2602gives you standard input. The @ARGV array is then processed as a list
2603of filenames. The loop
2604
2605 while (<>) {
2606 ... # code for each line
2607 }
2608
2609is equivalent to the following Perl-like pseudo code:
2610
3e3baf6d 2611 unshift(@ARGV, '-') unless @ARGV;
a0d0e21e
LW
2612 while (\$ARGV = shift) {
2613 open(ARGV, \$ARGV);
2614 while (<ARGV>) {
2615 ... # code for each line
2616 }
2617 }
2618
19799a22
GS
2619except that it isn't so cumbersome to say, and will actually work.
2620It really does shift the @ARGV array and put the current filename
2621into the \$ARGV variable. It also uses filehandle I<ARGV>
ac036724 2622internally. <> is just a synonym for <ARGV>, which
19799a22 2623is magical. (The pseudo code above doesn't work because it treats
35f2feb0 2624<ARGV> as non-magical.)
a0d0e21e 2625
48ab5743
ML
2626Since the null filehandle uses the two argument form of L<perlfunc/open>
2627it interprets special characters, so if you have a script like this:
2628
2629 while (<>) {
2630 print;
2631 }
2632
2633and call it with C<perl dangerous.pl 'rm -rfv *|'>, it actually opens a
2634pipe, executes the C<rm> command and reads C<rm>'s output from that pipe.
2635If you want all items in C<@ARGV> to be interpreted as file names, you
2636can use the module C<ARGV::readonly> from CPAN.
2637
35f2feb0 2638You can modify @ARGV before the first <> as long as the array ends up
a0d0e21e 2639containing the list of filenames you really want. Line numbers (C<\$.>)
19799a22
GS
2640continue as though the input were one big happy file. See the example
2641in L<perlfunc/eof> for how to reset line numbers on each file.
5a964f20 2642
89d205f2 2643If you want to set @ARGV to your own list of files, go right ahead.
5a964f20
TC
2644This sets @ARGV to all plain text files if no @ARGV was given:
2645
2646 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
a0d0e21e 2647
5a964f20
TC
2648You can even set them to pipe commands. For example, this automatically
2649filters compressed arguments through B<gzip>:
2650
2651 @ARGV = map { /\.(gz|Z)\$/ ? "gzip -dc < \$_ |" : \$_ } @ARGV;
2652
2653If you want to pass switches into your script, you can use one of the
a0d0e21e
LW
2654Getopts modules or put a loop on the front like this:
2655
2656 while (\$_ = \$ARGV[0], /^-/) {
2657 shift;
2658 last if /^--\$/;
2659 if (/^-D(.*)/) { \$debug = \$1 }
2660 if (/^-v/) { \$verbose++ }
5a964f20 2661 # ... # other switches
a0d0e21e 2662 }
5a964f20 2663
a0d0e21e 2664 while (<>) {
5a964f20 2665 # ... # code for each line
a0d0e21e
LW
2666 }
2667
89d205f2
YO
2668The <> symbol will return C<undef> for end-of-file only once.
2669If you call it again after this, it will assume you are processing another
19799a22 2670@ARGV list, and if you haven't set @ARGV, will read input from STDIN.
a0d0e21e 2671
b159ebd3 2672If what the angle brackets contain is a simple scalar variable (e.g.,
35f2feb0 2673<\$foo>), then that variable contains the name of the
19799a22
GS
2674filehandle to input from, or its typeglob, or a reference to the
2675same. For example:
cb1a09d0
2676
2677 \$fh = \*STDIN;
2678 \$line = <\$fh>;
a0d0e21e 2679
5a964f20
TC
2680If what's within the angle brackets is neither a filehandle nor a simple
2681scalar variable containing a filehandle name, typeglob, or typeglob
2682reference, it is interpreted as a filename pattern to be globbed, and
2683either a list of filenames or the next filename in the list is returned,
19799a22 2684depending on context. This distinction is determined on syntactic
35f2feb0
GS
2685grounds alone. That means C<< <\$x> >> is always a readline() from
2686an indirect handle, but C<< <\$hash{key}> >> is always a glob().
5a964f20 2687That's because \$x is a simple scalar variable, but C<\$hash{key}> is
ef191992
YST
2688not--it's a hash element. Even C<< <\$x > >> (note the extra space)
2689is treated as C<glob("\$x ")>, not C<readline(\$x)>.
5a964f20
TC
2690
2691One level of double-quote interpretation is done first, but you can't
35f2feb0 2692say C<< <\$foo> >> because that's an indirect filehandle as explained
5a964f20
TC
2693in the previous paragraph. (In older versions of Perl, programmers
2694would insert curly brackets to force interpretation as a filename glob:
35f2feb0 2695C<< <\${foo}> >>. These days, it's considered cleaner to call the
5a964f20 2696internal function directly as C<glob(\$foo)>, which is probably the right
19799a22 2697way to have done it in the first place.) For example:
a0d0e21e
LW
2698
2699 while (<*.c>) {
2700 chmod 0644, \$_;
2701 }
2702
3a4b19e4 2703is roughly equivalent to:
a0d0e21e
LW
2704
2705 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
2706 while (<FOO>) {
5b3eff12 2707 chomp;
a0d0e21e
LW
2708 chmod 0644, \$_;
2709 }
2710
3a4b19e4
GS
2711except that the globbing is actually done internally using the standard
2712C<File::Glob> extension. Of course, the shortest way to do the above is:
a0d0e21e
LW
2713
2714 chmod 0644, <*.c>;
2715
19799a22
GS
2716A (file)glob evaluates its (embedded) argument only when it is
2717starting a new list. All values must be read before it will start
2718over. In list context, this isn't important because you automatically
2719get them all anyway. However, in scalar context the operator returns
069e01df 2720the next value each time it's called, or C<undef> when the list has
19799a22
GS
2721run out. As with filehandle reads, an automatic C<defined> is
2722generated when the glob occurs in the test part of a C<while>,
2723because legal glob returns (e.g. a file called F<0>) would otherwise
2724terminate the loop. Again, C<undef> is returned only once. So if
2725you're expecting a single value from a glob, it is much better to
2726say
4633a7c4
LW
2727
2728 (\$file) = <blurch*>;
2729
2730than
2731
2732 \$file = <blurch*>;
2733
2734because the latter will alternate between returning a filename and
19799a22 2735returning false.
4633a7c4 2736
b159ebd3 2737If you're trying to do variable interpolation, it's definitely better
4633a7c4 2738to use the glob() function, because the older notation can cause people
e37d713d 2739to become confused with the indirect filehandle notation.
4633a7c4
LW
2740
2741 @files = glob("\$dir/*.[ch]");
2742 @files = glob(\$files[\$i]);
2743
d74e8afc 2745X<constant folding> X<folding>
a0d0e21e
LW
2746
2747Like C, Perl does a certain amount of expression evaluation at
19799a22 2748compile time whenever it determines that all arguments to an
a0d0e21e
LW
2749operator are static and have no side effects. In particular, string
2750concatenation happens at compile time between literals that don't do
19799a22 2751variable substitution. Backslash interpolation also happens at
a0d0e21e
LW
2752compile time. You can say
2753
2754 'Now is the time for all' . "\n" .
2755 'good men to come to.'
2756
54310121 2757and this all reduces to one string internally. Likewise, if
a0d0e21e
LW
2758you say
2759
2760 foreach \$file (@filenames) {
5a964f20 2761 if (-s \$file > 5 + 100 * 2**16) { }
54310121 2762 }
a0d0e21e 2763
19799a22
GS
2764the compiler will precompute the number which that expression
2765represents so that the interpreter won't have to.
a0d0e21e 2766
d74e8afc 2768X<no-op> X<nop>
fd1abbef
DN
2769
2770Perl doesn't officially have a no-op operator, but the bare constants
2771C<0> and C<1> are special-cased to not produce a warning in a void
2772context, so you can for example safely do
2773
2774 1 while foo();
2775
d74e8afc 2777X<operator, bitwise, string>
TP
2778
2779Bitstrings of any size may be manipulated by the bitwise operators
2780(C<~ | & ^>).
2781
19799a22
GS
2782If the operands to a binary bitwise op are strings of different
2783sizes, B<|> and B<^> ops act as though the shorter operand had
2784additional zero bits on the right, while the B<&> op acts as though
2785the longer operand were truncated to the length of the shorter.
2786The granularity for such extension or truncation is one or more
2787bytes.
89d205f2 2789 # ASCII-based examples
TP
2790 print "j p \n" ^ " a h"; # prints "JAPH\n"
2791 print "JA" | " ph\n"; # prints "japh\n"
2792 print "japh\nJunk" & '_____'; # prints "JAPH\n";
2793 print 'p N\$' ^ " E<H\n"; # prints "Perl\n";
2794
19799a22 2795If you are intending to manipulate bitstrings, be certain that
2c268ad5 2796you're supplying bitstrings: If an operand is a number, that will imply
19799a22 2797a B<numeric> bitwise operation. You may explicitly show which type of
TP
2798operation you intend by using C<""> or C<0+>, as in the examples below.
2799
4358a253
SS
2800 \$foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
2801 \$foo = '150' | 105; # yields 255
TP
2802 \$foo = 150 | '105'; # yields 255
2803 \$foo = '150' | '105'; # yields string '155' (under ASCII)
2804
2805 \$baz = 0+\$foo & 0+\$bar; # both ops explicitly numeric
2806 \$biz = "\$foo" ^ "\$bar"; # both ops explicitly stringy
a0d0e21e 2807
1ae175c8
GS
2808See L<perlfunc/vec> for information on how to manipulate individual bits
2809in a bit vector.
2810
d74e8afc 2812X<integer>
a0d0e21e 2813
19799a22 2814By default, Perl assumes that it must do most of its arithmetic in
a0d0e21e
LW
2815floating point. But by saying
2816
2817 use integer;
2818
3eab78e3
CW
2819you may tell the compiler to use integer operations
2820(see L<integer> for a detailed explanation) from here to the end of
2821the enclosing BLOCK. An inner BLOCK may countermand this by saying
a0d0e21e
LW
2822
2823 no integer;
2824
19799a22 2825which lasts until the end of that BLOCK. Note that this doesn't
3eab78e3
CW
2826mean everything is an integer, merely that Perl will use integer
2827operations for arithmetic, comparison, and bitwise operators. For
2828example, even under C<use integer>, if you take the C<sqrt(2)>, you'll
2829still get C<1.4142135623731> or so.
19799a22
GS
2830
2831Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
13a2d996 2833L<Bitwise String Operators>.) However, C<use integer> still has meaning for
19799a22
GS
2834them. By default, their results are interpreted as unsigned integers, but
2835if C<use integer> is in effect, their results are interpreted
2836as signed integers. For example, C<~0> usually evaluates to a large
0be96356 2837integral value. However, C<use integer; ~0> is C<-1> on two's-complement
19799a22 2838machines.
68dc0745 2839
06ce2fa3 2841
d74e8afc 2842X<floating-point> X<floating point> X<float> X<real>
68dc0745 2843
2844While C<use integer> provides integer-only arithmetic, there is no
19799a22
GS
2845analogous mechanism to provide automatic rounding or truncation to a
2846certain number of decimal places. For rounding to a certain number
2847of digits, sprintf() or printf() is usually the easiest route.
2848See L<perlfaq4>.
68dc0745 2849
5a964f20
TC
2850Floating-point numbers are only approximations to what a mathematician
2851would call real numbers. There are infinitely more reals than floats,
2852so some corners must be cut. For example:
2853
2854 printf "%.20g\n", 123456789123456789;
2855 # produces 123456789123456784
2856
8548cb57
RGS
2857Testing for exact floating-point equality or inequality is not a
2858good idea. Here's a (relatively expensive) work-around to compare
5a964f20
TC
2859whether two floating-point numbers are equal to a particular number of
2860decimal places. See Knuth, volume II, for a more robust treatment of
2861this topic.
2862
2863 sub fp_equal {
2864 my (\$X, \$Y, \$POINTS) = @_;
2865 my (\$tX, \$tY);
2866 \$tX = sprintf("%.\${POINTS}g", \$X);
2867 \$tY = sprintf("%.\${POINTS}g", \$Y);
2868 return \$tX eq \$tY;
2869 }
2870
68dc0745 2871The POSIX module (part of the standard perl distribution) implements
19799a22
GS
2872ceil(), floor(), and other mathematical and trigonometric functions.
2873The Math::Complex module (part of the standard perl distribution)
2874defines mathematical functions that work on both the reals and the
2875imaginary numbers. Math::Complex not as efficient as POSIX, but
68dc0745 2876POSIX can't work with complex numbers.
2877
2878Rounding in financial applications can have serious implications, and
2879the rounding method used should be specified precisely. In these
2880cases, it probably pays not to trust whichever system rounding is
2881being used by Perl, but to instead implement the rounding function you
2882need yourself.
5a964f20
TC
2883
d74e8afc 2885X<number, arbitrary precision>
5a964f20 2886
c543c01b
TC
2887The standard C<Math::BigInt>, C<Math::BigRat>, and C<Math::BigFloat> modules,
2888along with the C<bigint>, C<bigrat>, and C<bitfloat> pragmas, provide
19799a22 2889variable-precision arithmetic and overloaded operators, although
cd5c4fce 2890they're currently pretty slow. At the cost of some space and
19799a22
GS
2891considerable speed, they avoid the normal pitfalls associated with
2892limited-precision representations.
5a964f20 2893
c543c01b
TC
2894 use 5.010;
2895 use bigint; # easy interface to Math::BigInt
2896 \$x = 123456789123456789;
2897 say \$x * \$x;
2898 +15241578780673678515622620750190521
2899
2900Or with rationals:
2901
2902 use 5.010;
2903 use bigrat;
2904 \$a = 3/22;
2905 \$b = 4/6;
2906 say "a/b is ", \$a/\$b;
2907 say "a*b is ", \$a*\$b;
2908 a/b is 9/44
2909 a*b is 1/11
2910
2911Several modules let you calculate with (bound only by memory and CPU time)
2912unlimited or fixed precision. There are also some non-standard modules that
2913provide faster implementations via external C libraries.
cd5c4fce
T
2914
2915Here is a short, but incomplete summary:
2916
950b09ed
KW
2917 Math::Fraction big, unlimited fractions like 9973 / 12967
2918 Math::String treat string sequences like numbers
2919 Math::FixedPrecision calculate with a fixed precision
2920 Math::Currency for currency calculations
2921 Bit::Vector manipulate bit vectors fast (uses C)
2922 Math::BigIntFast Bit::Vector wrapper for big numbers