3 perlop - Perl operators and precedence
7 Perl operators have the following associativity and precedence,
8 listed from highest precedence to lowest. Note that all operators
9 borrowed from C keep the same precedence relationship with each other,
10 even where C's precedence is slightly screwy. (This makes learning
11 Perl easier for C folks.) With very few exceptions, these all
12 operate on scalar values only, not array values.
14 left terms and list operators (leftward)
18 right ! ~ \ and unary + and -
23 nonassoc named unary operators
24 nonassoc < > <= >= lt gt le ge
25 nonassoc == != <=> eq ne cmp
34 nonassoc list operators (rightward)
39 In the following sections, these operators are covered in precedence order.
41 Many operators can be overloaded for objects. See L<overload>.
45 =head2 Terms and List Operators (Leftward)
47 A TERM has the highest precedence in Perl. They include variables,
48 quote and quote-like operators, any expression in parentheses,
49 and any function whose arguments are parenthesized. Actually, there
50 aren't really functions in this sense, just list operators and unary
51 operators behaving as functions because you put parentheses around
52 the arguments. These are all documented in L<perlfunc>.
54 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
55 is followed by a left parenthesis as the next token, the operator and
56 arguments within parentheses are taken to be of highest precedence,
57 just like a normal function call.
59 In the absence of parentheses, the precedence of list operators such as
60 C<print>, C<sort>, or C<chmod> is either very high or very low depending on
61 whether you are looking at the left side or the right side of the operator.
64 @ary = (1, 3, sort 4, 2);
65 print @ary; # prints 1324
67 the commas on the right of the sort are evaluated before the sort, but
68 the commas on the left are evaluated after. In other words, list
69 operators tend to gobble up all the arguments that follow them, and
70 then act like a simple TERM with regard to the preceding expression.
71 Note that you have to be careful with parentheses:
73 # These evaluate exit before doing the print:
74 print($foo, exit); # Obviously not what you want.
75 print $foo, exit; # Nor is this.
77 # These do the print before evaluating exit:
78 (print $foo), exit; # This is what you want.
79 print($foo), exit; # Or this.
80 print ($foo), exit; # Or even this.
84 print ($foo & 255) + 1, "\n";
86 probably doesn't do what you expect at first glance. See
87 L<Named Unary Operators> for more discussion of this.
89 Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
90 well as subroutine and method calls, and the anonymous
91 constructors C<[]> and C<{}>.
93 See also L<Quote and Quote-like Operators> toward the end of this section,
94 as well as L<"I/O Operators">.
96 =head2 The Arrow Operator
98 Just as in C and C++, "C<-E<gt>>" is an infix dereference operator. If the
99 right side is either a C<[...]> or C<{...}> subscript, then the left side
100 must be either a hard or symbolic reference to an array or hash (or
101 a location capable of holding a hard reference, if it's an lvalue (assignable)).
104 Otherwise, the right side is a method name or a simple scalar variable
105 containing the method name, and the left side must either be an object
106 (a blessed reference) or a class name (that is, a package name).
109 =head2 Auto-increment and Auto-decrement
111 "++" and "--" work as in C. That is, if placed before a variable, they
112 increment or decrement the variable before returning the value, and if
113 placed after, increment or decrement the variable after returning the value.
115 The auto-increment operator has a little extra builtin magic to it. If
116 you increment a variable that is numeric, or that has ever been used in
117 a numeric context, you get a normal increment. If, however, the
118 variable has been used in only string contexts since it was set, and
119 has a value that is not the empty string and matches the pattern
120 C</^[a-zA-Z]*[0-9]*$/>, the increment is done as a string, preserving each
121 character within its range, with carry:
123 print ++($foo = '99'); # prints '100'
124 print ++($foo = 'a0'); # prints 'a1'
125 print ++($foo = 'Az'); # prints 'Ba'
126 print ++($foo = 'zz'); # prints 'aaa'
128 The auto-decrement operator is not magical.
130 =head2 Exponentiation
132 Binary "**" is the exponentiation operator. Note that it binds even more
133 tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
134 implemented using C's pow(3) function, which actually works on doubles
137 =head2 Symbolic Unary Operators
139 Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
140 precedence version of this.
142 Unary "-" performs arithmetic negation if the operand is numeric. If
143 the operand is an identifier, a string consisting of a minus sign
144 concatenated with the identifier is returned. Otherwise, if the string
145 starts with a plus or minus, a string starting with the opposite sign
146 is returned. One effect of these rules is that C<-bareword> is equivalent
149 Unary "~" performs bitwise negation, i.e., 1's complement. For example,
150 C<0666 &~ 027> is 0640. (See also L<Integer Arithmetic> and L<Bitwise
153 Unary "+" has no effect whatsoever, even on strings. It is useful
154 syntactically for separating a function name from a parenthesized expression
155 that would otherwise be interpreted as the complete list of function
156 arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
158 Unary "\" creates a reference to whatever follows it. See L<perlref>.
159 Do not confuse this behavior with the behavior of backslash within a
160 string, although both forms do convey the notion of protecting the next
161 thing from interpretation.
163 =head2 Binding Operators
165 Binary "=~" binds a scalar expression to a pattern match. Certain operations
166 search or modify the string $_ by default. This operator makes that kind
167 of operation work on some other string. The right argument is a search
168 pattern, substitution, or transliteration. The left argument is what is
169 supposed to be searched, substituted, or transliterated instead of the default
170 $_. The return value indicates the success of the operation. (If the
171 right argument is an expression rather than a search pattern,
172 substitution, or transliteration, it is interpreted as a search pattern at run
173 time. This can be is less efficient than an explicit search, because the
174 pattern must be compiled every time the expression is evaluated.
176 Binary "!~" is just like "=~" except the return value is negated in
179 =head2 Multiplicative Operators
181 Binary "*" multiplies two numbers.
183 Binary "/" divides two numbers.
185 Binary "%" computes the modulus of two numbers. Given integer
186 operands C<$a> and C<$b>: If C<$b> is positive, then C<$a % $b> is
187 C<$a> minus the largest multiple of C<$b> that is not greater than
188 C<$a>. If C<$b> is negative, then C<$a % $b> is C<$a> minus the
189 smallest multiple of C<$b> that is not less than C<$a> (i.e. the
190 result will be less than or equal to zero).
191 Note than when C<use integer> is in scope, "%" give you direct access
192 to the modulus operator as implemented by your C compiler. This
193 operator is not as well defined for negative operands, but it will
196 Binary "x" is the repetition operator. In scalar context, it
197 returns a string consisting of the left operand repeated the number of
198 times specified by the right operand. In list context, if the left
199 operand is a list in parentheses, it repeats the list.
201 print '-' x 80; # print row of dashes
203 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
205 @ones = (1) x 80; # a list of 80 1's
206 @ones = (5) x @ones; # set all elements to 5
209 =head2 Additive Operators
211 Binary "+" returns the sum of two numbers.
213 Binary "-" returns the difference of two numbers.
215 Binary "." concatenates two strings.
217 =head2 Shift Operators
219 Binary "<<" returns the value of its left argument shifted left by the
220 number of bits specified by the right argument. Arguments should be
221 integers. (See also L<Integer Arithmetic>.)
223 Binary ">>" returns the value of its left argument shifted right by
224 the number of bits specified by the right argument. Arguments should
225 be integers. (See also L<Integer Arithmetic>.)
227 =head2 Named Unary Operators
229 The various named unary operators are treated as functions with one
230 argument, with optional parentheses. These include the filetest
231 operators, like C<-f>, C<-M>, etc. See L<perlfunc>.
233 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
234 is followed by a left parenthesis as the next token, the operator and
235 arguments within parentheses are taken to be of highest precedence,
236 just like a normal function call. Examples:
238 chdir $foo || die; # (chdir $foo) || die
239 chdir($foo) || die; # (chdir $foo) || die
240 chdir ($foo) || die; # (chdir $foo) || die
241 chdir +($foo) || die; # (chdir $foo) || die
243 but, because * is higher precedence than ||:
245 chdir $foo * 20; # chdir ($foo * 20)
246 chdir($foo) * 20; # (chdir $foo) * 20
247 chdir ($foo) * 20; # (chdir $foo) * 20
248 chdir +($foo) * 20; # chdir ($foo * 20)
250 rand 10 * 20; # rand (10 * 20)
251 rand(10) * 20; # (rand 10) * 20
252 rand (10) * 20; # (rand 10) * 20
253 rand +(10) * 20; # rand (10 * 20)
255 See also L<"Terms and List Operators (Leftward)">.
257 =head2 Relational Operators
259 Binary "E<lt>" returns true if the left argument is numerically less than
262 Binary "E<gt>" returns true if the left argument is numerically greater
263 than the right argument.
265 Binary "E<lt>=" returns true if the left argument is numerically less than
266 or equal to the right argument.
268 Binary "E<gt>=" returns true if the left argument is numerically greater
269 than or equal to the right argument.
271 Binary "lt" returns true if the left argument is stringwise less than
274 Binary "gt" returns true if the left argument is stringwise greater
275 than the right argument.
277 Binary "le" returns true if the left argument is stringwise less than
278 or equal to the right argument.
280 Binary "ge" returns true if the left argument is stringwise greater
281 than or equal to the right argument.
283 =head2 Equality Operators
285 Binary "==" returns true if the left argument is numerically equal to
288 Binary "!=" returns true if the left argument is numerically not equal
289 to the right argument.
291 Binary "E<lt>=E<gt>" returns -1, 0, or 1 depending on whether the left
292 argument is numerically less than, equal to, or greater than the right
295 Binary "eq" returns true if the left argument is stringwise equal to
298 Binary "ne" returns true if the left argument is stringwise not equal
299 to the right argument.
301 Binary "cmp" returns -1, 0, or 1 depending on whether the left argument is stringwise
302 less than, equal to, or greater than the right argument.
304 "lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
305 by the current locale if C<use locale> is in effect. See L<perllocale>.
309 Binary "&" returns its operators ANDed together bit by bit.
310 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
312 =head2 Bitwise Or and Exclusive Or
314 Binary "|" returns its operators ORed together bit by bit.
315 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
317 Binary "^" returns its operators XORed together bit by bit.
318 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
320 =head2 C-style Logical And
322 Binary "&&" performs a short-circuit logical AND operation. That is,
323 if the left operand is false, the right operand is not even evaluated.
324 Scalar or list context propagates down to the right operand if it
327 =head2 C-style Logical Or
329 Binary "||" performs a short-circuit logical OR operation. That is,
330 if the left operand is true, the right operand is not even evaluated.
331 Scalar or list context propagates down to the right operand if it
334 The C<||> and C<&&> operators differ from C's in that, rather than returning
335 0 or 1, they return the last value evaluated. Thus, a reasonably portable
336 way to find out the home directory (assuming it's not "0") might be:
338 $home = $ENV{'HOME'} || $ENV{'LOGDIR'} ||
339 (getpwuid($<))[7] || die "You're homeless!\n";
341 In particular, this means that you shouldn't use this
342 for selecting between two aggregates for assignment:
344 @a = @b || @c; # this is wrong
345 @a = scalar(@b) || @c; # really meant this
346 @a = @b ? @b : @c; # this works fine, though
348 As more readable alternatives to C<&&> and C<||> when used for
349 control flow, Perl provides C<and> and C<or> operators (see below).
350 The short-circuit behavior is identical. The precedence of "and" and
351 "or" is much lower, however, so that you can safely use them after a
352 list operator without the need for parentheses:
354 unlink "alpha", "beta", "gamma"
355 or gripe(), next LINE;
357 With the C-style operators that would have been written like this:
359 unlink("alpha", "beta", "gamma")
360 || (gripe(), next LINE);
362 Use "or" for assignment is unlikely to do what you want; see below.
364 =head2 Range Operators
366 Binary ".." is the range operator, which is really two different
367 operators depending on the context. In list context, it returns an
368 array of values counting (up by ones) from the left value to the right
369 value. If the left value is greater than the right value then it
370 returns the empty array. The range operator is useful for writing
371 C<foreach (1..10)> loops and for doing slice operations on arrays. In
372 the current implementation, no temporary array is created when the
373 range operator is used as the expression in C<foreach> loops, but older
374 versions of Perl might burn a lot of memory when you write something
377 for (1 .. 1_000_000) {
381 In scalar context, ".." returns a boolean value. The operator is
382 bistable, like a flip-flop, and emulates the line-range (comma) operator
383 of B<sed>, B<awk>, and various editors. Each ".." operator maintains its
384 own boolean state. It is false as long as its left operand is false.
385 Once the left operand is true, the range operator stays true until the
386 right operand is true, I<AFTER> which the range operator becomes false
387 again. (It doesn't become false till the next time the range operator is
388 evaluated. It can test the right operand and become false on the same
389 evaluation it became true (as in B<awk>), but it still returns true once.
390 If you don't want it to test the right operand till the next evaluation
391 (as in B<sed>), use three dots ("...") instead of two.) The right
392 operand is not evaluated while the operator is in the "false" state, and
393 the left operand is not evaluated while the operator is in the "true"
394 state. The precedence is a little lower than || and &&. The value
395 returned is either the empty string for false, or a sequence number
396 (beginning with 1) for true. The sequence number is reset for each range
397 encountered. The final sequence number in a range has the string "E0"
398 appended to it, which doesn't affect its numeric value, but gives you
399 something to search for if you want to exclude the endpoint. You can
400 exclude the beginning point by waiting for the sequence number to be
401 greater than 1. If either operand of scalar ".." is a constant expression,
402 that operand is implicitly compared to the C<$.> variable, the current
403 line number. Examples:
405 As a scalar operator:
407 if (101 .. 200) { print; } # print 2nd hundred lines
408 next line if (1 .. /^$/); # skip header lines
409 s/^/> / if (/^$/ .. eof()); # quote body
411 # parse mail messages
413 $in_header = 1 .. /^$/;
414 $in_body = /^$/ .. eof();
415 # do something based on those
417 close ARGV if eof; # reset $. each file
422 for (101 .. 200) { print; } # print $_ 100 times
423 @foo = @foo[0 .. $#foo]; # an expensive no-op
424 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
426 The range operator (in list context) makes use of the magical
427 auto-increment algorithm if the operands are strings. You
430 @alphabet = ('A' .. 'Z');
432 to get all the letters of the alphabet, or
434 $hexdigit = (0 .. 9, 'a' .. 'f')[$num & 15];
436 to get a hexadecimal digit, or
438 @z2 = ('01' .. '31'); print $z2[$mday];
440 to get dates with leading zeros. If the final value specified is not
441 in the sequence that the magical increment would produce, the sequence
442 goes until the next value would be longer than the final value
445 =head2 Conditional Operator
447 Ternary "?:" is the conditional operator, just as in C. It works much
448 like an if-then-else. If the argument before the ? is true, the
449 argument before the : is returned, otherwise the argument after the :
450 is returned. For example:
452 printf "I have %d dog%s.\n", $n,
453 ($n == 1) ? '' : "s";
455 Scalar or list context propagates downward into the 2nd
456 or 3rd argument, whichever is selected.
458 $a = $ok ? $b : $c; # get a scalar
459 @a = $ok ? @b : @c; # get an array
460 $a = $ok ? @b : @c; # oops, that's just a count!
462 The operator may be assigned to if both the 2nd and 3rd arguments are
463 legal lvalues (meaning that you can assign to them):
465 ($a_or_b ? $a : $b) = $c;
467 This is not necessarily guaranteed to contribute to the readability of your program.
469 Because this operator produces an assignable result, using assignments
470 without parentheses will get you in trouble. For example, this:
472 $a % 2 ? $a += 10 : $a += 2
476 (($a % 2) ? ($a += 10) : $a) += 2
480 ($a % 2) ? ($a += 10) : ($a += 2)
482 =head2 Assignment Operators
484 "=" is the ordinary assignment operator.
486 Assignment operators work as in C. That is,
494 although without duplicating any side effects that dereferencing the lvalue
495 might trigger, such as from tie(). Other assignment operators work similarly.
496 The following are recognized:
503 Note that while these are grouped by family, they all have the precedence
506 Unlike in C, the assignment operator produces a valid lvalue. Modifying
507 an assignment is equivalent to doing the assignment and then modifying
508 the variable that was assigned to. This is useful for modifying
509 a copy of something, like this:
511 ($tmp = $global) =~ tr [A-Z] [a-z];
522 =head2 Comma Operator
524 Binary "," is the comma operator. In scalar context it evaluates
525 its left argument, throws that value away, then evaluates its right
526 argument and returns that value. This is just like C's comma operator.
528 In list context, it's just the list argument separator, and inserts
529 both its arguments into the list.
531 The =E<gt> digraph is mostly just a synonym for the comma operator. It's useful for
532 documenting arguments that come in pairs. As of release 5.001, it also forces
533 any word to the left of it to be interpreted as a string.
535 =head2 List Operators (Rightward)
537 On the right side of a list operator, it has very low precedence,
538 such that it controls all comma-separated expressions found there.
539 The only operators with lower precedence are the logical operators
540 "and", "or", and "not", which may be used to evaluate calls to list
541 operators without the need for extra parentheses:
543 open HANDLE, "filename"
544 or die "Can't open: $!\n";
546 See also discussion of list operators in L<Terms and List Operators (Leftward)>.
550 Unary "not" returns the logical negation of the expression to its right.
551 It's the equivalent of "!" except for the very low precedence.
555 Binary "and" returns the logical conjunction of the two surrounding
556 expressions. It's equivalent to && except for the very low
557 precedence. This means that it short-circuits: i.e., the right
558 expression is evaluated only if the left expression is true.
560 =head2 Logical or and Exclusive Or
562 Binary "or" returns the logical disjunction of the two surrounding
563 expressions. It's equivalent to || except for the very low precedence.
564 This makes it useful for control flow
566 print FH $data or die "Can't write to FH: $!";
568 This means that it short-circuits: i.e., the right expression is evaluated
569 only if the left expression is false. Due to its precedence, you should
570 probably avoid using this for assignment, only for control flow.
572 $a = $b or $c; # bug: this is wrong
573 ($a = $b) or $c; # really means this
574 $a = $b || $c; # better written this way
576 However, when it's a list context assignment and you're trying to use
577 "||" for control flow, you probably need "or" so that the assignment
578 takes higher precedence.
580 @info = stat($file) || die; # oops, scalar sense of stat!
581 @info = stat($file) or die; # better, now @info gets its due
583 Then again, you could always use parentheses.
585 Binary "xor" returns the exclusive-OR of the two surrounding expressions.
586 It cannot short circuit, of course.
588 =head2 C Operators Missing From Perl
590 Here is what C has that Perl doesn't:
596 Address-of operator. (But see the "\" operator for taking a reference.)
600 Dereference-address operator. (Perl's prefix dereferencing
601 operators are typed: $, @, %, and &.)
605 Type casting operator.
609 =head2 Quote and Quote-like Operators
611 While we usually think of quotes as literal values, in Perl they
612 function as operators, providing various kinds of interpolating and
613 pattern matching capabilities. Perl provides customary quote characters
614 for these behaviors, but also provides a way for you to choose your
615 quote character for any of them. In the following table, a C<{}> represents
616 any pair of delimiters you choose. Non-bracketing delimiters use
617 the same character fore and aft, but the 4 sorts of brackets
618 (round, angle, square, curly) will all nest.
620 Customary Generic Meaning Interpolates
623 `` qx{} Command yes (unless '' is delimiter)
625 // m{} Pattern match yes (unless '' is delimiter)
626 qr{} Pattern yes (unless '' is delimiter)
627 s{}{} Substitution yes (unless '' is delimiter)
628 tr{}{} Transliteration no (but see below)
630 Note that there can be whitespace between the operator and the quoting
631 characters, except when C<#> is being used as the quoting character.
632 C<q#foo#> is parsed as being the string C<foo>, while C<q #foo#> is the
633 operator C<q> followed by a comment. Its argument will be taken from the
634 next line. This allows you to write:
636 s {foo} # Replace foo
639 For constructs that do interpolation, variables beginning with "C<$>"
640 or "C<@>" are interpolated, as are the following sequences. Within
641 a transliteration, the first eleven of these sequences may be used.
648 \a alarm (bell) (BEL)
650 \033 octal char (ESC)
652 \x{263a} wide hex char (SMILEY)
655 \l lowercase next char
656 \u uppercase next char
659 \E end case modification
660 \Q quote non-word characters till \E
662 If C<use locale> is in effect, the case map used by C<\l>, C<\L>, C<\u>
663 and C<\U> is taken from the current locale. See L<perllocale>.
665 All systems use the virtual C<"\n"> to represent a line terminator,
666 called a "newline". There is no such thing as an unvarying, physical
667 newline character. It is an illusion that the operating system,
668 device drivers, C libraries, and Perl all conspire to preserve. Not all
669 systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
670 on a Mac, these are reversed, and on systems without line terminator,
671 printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
672 you mean a "newline" for your system, but use the literal ASCII when you
673 need an exact character. For example, most networking protocols expect
674 and prefer a CR+LF (C<"\012\015"> or C<"\cJ\cM">) for line terminators,
675 and although they often accept just C<"\012">, they seldom tolerate just
676 C<"\015">. If you get in the habit of using C<"\n"> for networking,
677 you may be burned some day.
679 You cannot include a literal C<$> or C<@> within a C<\Q> sequence.
680 An unescaped C<$> or C<@> interpolates the corresponding variable,
681 while escaping will cause the literal string C<\$> to be inserted.
682 You'll need to write something like C<m/\Quser\E\@\Qhost/>.
684 Patterns are subject to an additional level of interpretation as a
685 regular expression. This is done as a second pass, after variables are
686 interpolated, so that regular expressions may be incorporated into the
687 pattern from the variables. If this is not what you want, use C<\Q> to
688 interpolate a variable literally.
690 Apart from the above, there are no multiple levels of interpolation. In
691 particular, contrary to the expectations of shell programmers, back-quotes
692 do I<NOT> interpolate within double quotes, nor do single quotes impede
693 evaluation of variables when used within double quotes.
695 =head2 Regexp Quote-Like Operators
697 Here are the quote-like operators that apply to pattern
698 matching and related activities.
700 Most of this section is related to use of regular expressions from Perl.
701 Such a use may be considered from two points of view: Perl handles a
702 a string and a "pattern" to RE (regular expression) engine to match,
703 RE engine finds (or does not find) the match, and Perl uses the findings
704 of RE engine for its operation, possibly asking the engine for other matches.
706 RE engine has no idea what Perl is going to do with what it finds,
707 similarly, the rest of Perl has no idea what a particular regular expression
708 means to RE engine. This creates a clean separation, and in this section
709 we discuss matching from Perl point of view only. The other point of
710 view may be found in L<perlre>.
716 This is just like the C</pattern/> search, except that it matches only
717 once between calls to the reset() operator. This is a useful
718 optimization when you want to see only the first occurrence of
719 something in each file of a set of files, for instance. Only C<??>
720 patterns local to the current package are reset.
724 # blank line between header and body
727 reset if eof; # clear ?? status for next file
730 This usage is vaguely deprecated, and may be removed in some future
733 =item m/PATTERN/cgimosx
735 =item /PATTERN/cgimosx
737 Searches a string for a pattern match, and in scalar context returns
738 true (1) or false (''). If no string is specified via the C<=~> or
739 C<!~> operator, the $_ string is searched. (The string specified with
740 C<=~> need not be an lvalue--it may be the result of an expression
741 evaluation, but remember the C<=~> binds rather tightly.) See also
743 See L<perllocale> for discussion of additional considerations that apply
744 when C<use locale> is in effect.
748 c Do not reset search position on a failed match when /g is in effect.
749 g Match globally, i.e., find all occurrences.
750 i Do case-insensitive pattern matching.
751 m Treat string as multiple lines.
752 o Compile pattern only once.
753 s Treat string as single line.
754 x Use extended regular expressions.
756 If "/" is the delimiter then the initial C<m> is optional. With the C<m>
757 you can use any pair of non-alphanumeric, non-whitespace characters
758 as delimiters. This is particularly useful for matching Unix path names
759 that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
760 the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
761 If "'" is the delimiter, no variable interpolation is performed on the
764 PATTERN may contain variables, which will be interpolated (and the
765 pattern recompiled) every time the pattern search is evaluated, except
766 for when the delimiter is a single quote. (Note that C<$)> and C<$|>
767 might not be interpolated because they look like end-of-string tests.)
768 If you want such a pattern to be compiled only once, add a C</o> after
769 the trailing delimiter. This avoids expensive run-time recompilations,
770 and is useful when the value you are interpolating won't change over
771 the life of the script. However, mentioning C</o> constitutes a promise
772 that you won't change the variables in the pattern. If you change them,
773 Perl won't even notice.
775 If the PATTERN evaluates to the empty string, the last
776 I<successfully> matched regular expression is used instead.
778 If the C</g> option is not used, C<m//> in a list context returns a
779 list consisting of the subexpressions matched by the parentheses in the
780 pattern, i.e., (C<$1>, C<$2>, C<$3>...). (Note that here C<$1> etc. are
781 also set, and that this differs from Perl 4's behavior.) When there are
782 no parentheses in the pattern, the return value is the list C<(1)> for
783 success. With or without parentheses, an empty list is returned upon
788 open(TTY, '/dev/tty');
789 <TTY> =~ /^y/i && foo(); # do foo if desired
791 if (/Version: *([0-9.]*)/) { $version = $1; }
793 next if m#^/usr/spool/uucp#;
798 print if /$arg/o; # compile only once
801 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
803 This last example splits $foo into the first two words and the
804 remainder of the line, and assigns those three fields to $F1, $F2, and
805 $Etc. The conditional is true if any variables were assigned, i.e., if
808 The C</g> modifier specifies global pattern matching--that is, matching
809 as many times as possible within the string. How it behaves depends on
810 the context. In list context, it returns a list of all the
811 substrings matched by all the parentheses in the regular expression.
812 If there are no parentheses, it returns a list of all the matched
813 strings, as if there were parentheses around the whole pattern.
815 In scalar context, each execution of C<m//g> finds the next match,
816 returning TRUE if it matches, and FALSE if there is no further match.
817 The position after the last match can be read or set using the pos()
818 function; see L<perlfunc/pos>. A failed match normally resets the
819 search position to the beginning of the string, but you can avoid that
820 by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
821 string also resets the search position.
823 You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
824 zero-width assertion that matches the exact position where the previous
825 C<m//g>, if any, left off. The C<\G> assertion is not supported without
826 the C</g> modifier; currently, without C</g>, C<\G> behaves just like
827 C<\A>, but that's accidental and may change in the future.
832 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
837 while (defined($paragraph = <>)) {
838 while ($paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
843 print "$sentences\n";
845 # using m//gc with \G
849 print $1 while /(o)/gc; print "', pos=", pos, "\n";
851 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
853 print $1 while /(p)/gc; print "', pos=", pos, "\n";
856 The last example should print:
865 A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
866 combine several regexps like this to process a string part-by-part,
867 doing different actions depending on which regexp matched. Each
868 regexp tries to match where the previous one leaves off.
871 $url = new URI::URL "http://www/"; die if $url eq "xXx";
875 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
876 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
877 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
878 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
879 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
880 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
881 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
882 print ". That's all!\n";
885 Here is the output (split into several lines):
887 line-noise lowercase line-noise lowercase UPPERCASE line-noise
888 UPPERCASE line-noise lowercase line-noise lowercase line-noise
889 lowercase lowercase line-noise lowercase lowercase line-noise
890 MiXeD line-noise. That's all!
896 A single-quoted, literal string. A backslash represents a backslash
897 unless followed by the delimiter or another backslash, in which case
898 the delimiter or backslash is interpolated.
900 $foo = q!I said, "You said, 'She said it.'"!;
901 $bar = q('This is it.');
902 $baz = '\n'; # a two-character string
908 A double-quoted, interpolated string.
911 (*** The previous line contains the naughty word "$1".\n)
912 if /(tcl|rexx|python)/; # :-)
913 $baz = "\n"; # a one-character string
915 =item qr/STRING/imosx
917 Quote-as-a-regular-expression operator. I<STRING> is interpolated the
918 same way as I<PATTERN> in C<m/PATTERN/>. If "'" is used as the
919 delimiter, no variable interpolation is done. Returns a Perl value
920 which may be used instead of the corresponding C</STRING/imosx> expression.
924 $rex = qr/my.STRING/is;
931 The result may be used as a subpattern in a match:
934 $string =~ /foo${re}bar/; # can be interpolated in other patterns
935 $string =~ $re; # or used standalone
936 $string =~ /$re/; # or this way
938 Since Perl may compile the pattern at the moment of execution of qr()
939 operator, using qr() may have speed advantages in I<some> situations,
940 notably if the result of qr() is used standalone:
943 my $patterns = shift;
944 my @compiled = map qr/$_/i, @$patterns;
947 foreach my $pat @compiled {
948 $success = 1, last if /$pat/;
954 Precompilation of the pattern into an internal representation at the
955 moment of qr() avoids a need to recompile the pattern every time a
956 match C</$pat/> is attempted. (Note that Perl has many other
957 internal optimizations, but none would be triggered in the above
958 example if we did not use qr() operator.)
962 i Do case-insensitive pattern matching.
963 m Treat string as multiple lines.
964 o Compile pattern only once.
965 s Treat string as single line.
966 x Use extended regular expressions.
968 See L<perlre> for additional information on valid syntax for STRING, and
969 for a detailed look at the semantics of regular expressions.
975 A string which is (possibly) interpolated and then executed as a system
976 command with C</bin/sh> or its equivalent. Shell wildcards, pipes,
977 and redirections will be honored. The collected standard output of the
978 command is returned; standard error is unaffected. In scalar context,
979 it comes back as a single (potentially multi-line) string. In list
980 context, returns a list of lines (however you've defined lines with $/
981 or $INPUT_RECORD_SEPARATOR).
983 Because backticks do not affect standard error, use shell file descriptor
984 syntax (assuming the shell supports this) if you care to address this.
985 To capture a command's STDERR and STDOUT together:
987 $output = `cmd 2>&1`;
989 To capture a command's STDOUT but discard its STDERR:
991 $output = `cmd 2>/dev/null`;
993 To capture a command's STDERR but discard its STDOUT (ordering is
996 $output = `cmd 2>&1 1>/dev/null`;
998 To exchange a command's STDOUT and STDERR in order to capture the STDERR
999 but leave its STDOUT to come out the old STDERR:
1001 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1003 To read both a command's STDOUT and its STDERR separately, it's easiest
1004 and safest to redirect them separately to files, and then read from those
1005 files when the program is done:
1007 system("program args 1>/tmp/program.stdout 2>/tmp/program.stderr");
1009 Using single-quote as a delimiter protects the command from Perl's
1010 double-quote interpolation, passing it on to the shell instead:
1012 $perl_info = qx(ps $$); # that's Perl's $$
1013 $shell_info = qx'ps $$'; # that's the new shell's $$
1015 Note that how the string gets evaluated is entirely subject to the command
1016 interpreter on your system. On most platforms, you will have to protect
1017 shell metacharacters if you want them treated literally. This is in
1018 practice difficult to do, as it's unclear how to escape which characters.
1019 See L<perlsec> for a clean and safe example of a manual fork() and exec()
1020 to emulate backticks safely.
1022 On some platforms (notably DOS-like ones), the shell may not be
1023 capable of dealing with multiline commands, so putting newlines in
1024 the string may not get you what you want. You may be able to evaluate
1025 multiple commands in a single line by separating them with the command
1026 separator character, if your shell supports that (e.g. C<;> on many Unix
1027 shells; C<&> on the Windows NT C<cmd> shell).
1029 Beware that some command shells may place restrictions on the length
1030 of the command line. You must ensure your strings don't exceed this
1031 limit after any necessary interpolations. See the platform-specific
1032 release notes for more details about your particular environment.
1034 Using this operator can lead to programs that are difficult to port,
1035 because the shell commands called vary between systems, and may in
1036 fact not be present at all. As one example, the C<type> command under
1037 the POSIX shell is very different from the C<type> command under DOS.
1038 That doesn't mean you should go out of your way to avoid backticks
1039 when they're the right way to get something done. Perl was made to be
1040 a glue language, and one of the things it glues together is commands.
1041 Just understand what you're getting yourself into.
1043 See L<"I/O Operators"> for more discussion.
1047 Evaluates to a list of the words extracted out of STRING, using embedded
1048 whitespace as the word delimiters. It can be understood as being roughly
1051 split(' ', q/STRING/);
1053 the difference being that it generates a real list at compile time. So
1058 is exactly equivalent to the list:
1060 ('foo', 'bar', 'baz')
1062 Some frequently seen examples:
1064 use POSIX qw( setlocale localeconv )
1065 @EXPORT = qw( foo bar baz );
1067 A common mistake is to try to separate the words with comma or to put
1068 comments into a multi-line C<qw>-string. For this reason the C<-w>
1069 switch produce warnings if the STRING contains the "," or the "#"
1072 =item s/PATTERN/REPLACEMENT/egimosx
1074 Searches a string for a pattern, and if found, replaces that pattern
1075 with the replacement text and returns the number of substitutions
1076 made. Otherwise it returns false (specifically, the empty string).
1078 If no string is specified via the C<=~> or C<!~> operator, the C<$_>
1079 variable is searched and modified. (The string specified with C<=~> must
1080 be scalar variable, an array element, a hash element, or an assignment
1081 to one of those, i.e., an lvalue.)
1083 If the delimiter chosen is a single quote, no variable interpolation is
1084 done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1085 PATTERN contains a $ that looks like a variable rather than an
1086 end-of-string test, the variable will be interpolated into the pattern
1087 at run-time. If you want the pattern compiled only once the first time
1088 the variable is interpolated, use the C</o> option. If the pattern
1089 evaluates to the empty string, the last successfully executed regular
1090 expression is used instead. See L<perlre> for further explanation on these.
1091 See L<perllocale> for discussion of additional considerations that apply
1092 when C<use locale> is in effect.
1096 e Evaluate the right side as an expression.
1097 g Replace globally, i.e., all occurrences.
1098 i Do case-insensitive pattern matching.
1099 m Treat string as multiple lines.
1100 o Compile pattern only once.
1101 s Treat string as single line.
1102 x Use extended regular expressions.
1104 Any non-alphanumeric, non-whitespace delimiter may replace the
1105 slashes. If single quotes are used, no interpretation is done on the
1106 replacement string (the C</e> modifier overrides this, however). Unlike
1107 Perl 4, Perl 5 treats backticks as normal delimiters; the replacement
1108 text is not evaluated as a command. If the
1109 PATTERN is delimited by bracketing quotes, the REPLACEMENT has its own
1110 pair of quotes, which may or may not be bracketing quotes, e.g.,
1111 C<s(foo)(bar)> or C<sE<lt>fooE<gt>/bar/>. A C</e> will cause the
1112 replacement portion to be interpreted as a full-fledged Perl expression
1113 and eval()ed right then and there. It is, however, syntax checked at
1118 s/\bgreen\b/mauve/g; # don't change wintergreen
1120 $path =~ s|/usr/bin|/usr/local/bin|;
1122 s/Login: $foo/Login: $bar/; # run-time pattern
1124 ($foo = $bar) =~ s/this/that/; # copy first, then change
1126 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-count
1129 s/\d+/$&*2/e; # yields 'abc246xyz'
1130 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
1131 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1133 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
1134 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
1135 s/^=(\w+)/&pod($1)/ge; # use function call
1137 # expand variables in $_, but dynamics only, using
1138 # symbolic dereferencing
1141 # /e's can even nest; this will expand
1142 # any embedded scalar variable (including lexicals) in $_
1145 # Delete (most) C comments.
1147 /\* # Match the opening delimiter.
1148 .*? # Match a minimal number of characters.
1149 \*/ # Match the closing delimiter.
1152 s/^\s*(.*?)\s*$/$1/; # trim white space in $_, expensively
1154 for ($variable) { # trim white space in $variable, cheap
1159 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
1161 Note the use of $ instead of \ in the last example. Unlike
1162 B<sed>, we use the \E<lt>I<digit>E<gt> form in only the left hand side.
1163 Anywhere else it's $E<lt>I<digit>E<gt>.
1165 Occasionally, you can't use just a C</g> to get all the changes
1166 to occur. Here are two common cases:
1168 # put commas in the right places in an integer
1169 1 while s/(.*\d)(\d\d\d)/$1,$2/g; # perl4
1170 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g; # perl5
1172 # expand tabs to 8-column spacing
1173 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
1176 =item tr/SEARCHLIST/REPLACEMENTLIST/cdsUC
1178 =item y/SEARCHLIST/REPLACEMENTLIST/cdsUC
1180 Transliterates all occurrences of the characters found in the search list
1181 with the corresponding character in the replacement list. It returns
1182 the number of characters replaced or deleted. If no string is
1183 specified via the =~ or !~ operator, the $_ string is transliterated. (The
1184 string specified with =~ must be a scalar variable, an array element, a
1185 hash element, or an assignment to one of those, i.e., an lvalue.)
1187 A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
1188 does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
1189 For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1190 SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1191 its own pair of quotes, which may or may not be bracketing quotes,
1192 e.g., C<tr[A-Z][a-z]> or C<tr(+\-*/)/ABCD/>.
1194 Note also that the whole range idea is rather unportable between
1195 character sets--and even within character sets they may cause results
1196 you probably didn't expect. A sound principle is to use only ranges
1197 that begin from and end at either alphabets of equal case (a-e, A-E),
1198 or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1199 character sets in full.
1203 c Complement the SEARCHLIST.
1204 d Delete found but unreplaced characters.
1205 s Squash duplicate replaced characters.
1206 U Translate to/from UTF-8.
1207 C Translate to/from 8-bit char (octet).
1209 If the C</c> modifier is specified, the SEARCHLIST character set is
1210 complemented. If the C</d> modifier is specified, any characters specified
1211 by SEARCHLIST not found in REPLACEMENTLIST are deleted. (Note
1212 that this is slightly more flexible than the behavior of some B<tr>
1213 programs, which delete anything they find in the SEARCHLIST, period.)
1214 If the C</s> modifier is specified, sequences of characters that were
1215 transliterated to the same character are squashed down to a single instance of the
1218 If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1219 exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1220 than the SEARCHLIST, the final character is replicated till it is long
1221 enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
1222 This latter is useful for counting characters in a class or for
1223 squashing character sequences in a class.
1225 The first C</U> or C</C> modifier applies to the left side of the translation.
1226 The second one applies to the right side. If present, these modifiers override
1227 the current utf8 state.
1231 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1233 $cnt = tr/*/*/; # count the stars in $_
1235 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
1237 $cnt = tr/0-9//; # count the digits in $_
1239 tr/a-zA-Z//s; # bookkeeper -> bokeper
1241 ($HOST = $host) =~ tr/a-z/A-Z/;
1243 tr/a-zA-Z/ /cs; # change non-alphas to single space
1246 [\000-\177]; # delete 8th bit
1248 tr/\0-\xFF//CU; # translate Latin-1 to Unicode
1249 tr/\0-\x{FF}//UC; # translate Unicode to Latin-1
1251 If multiple transliterations are given for a character, only the first one is used:
1255 will transliterate any A to X.
1257 Note that because the transliteration table is built at compile time, neither
1258 the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
1259 interpolation. That means that if you want to use variables, you must use
1262 eval "tr/$oldlist/$newlist/";
1265 eval "tr/$oldlist/$newlist/, 1" or die $@;
1269 =head2 Gory details of parsing quoted constructs
1271 When presented with something which may have several different
1272 interpretations, Perl uses the principle B<DWIM> (expanded to Do What I Mean
1273 - not what I wrote) to pick up the most probable interpretation of the
1274 source. This strategy is so successful that Perl users usually do not
1275 suspect ambivalence of what they write. However, time to time Perl's ideas
1276 differ from what the author meant.
1278 The target of this section is to clarify the Perl's way of interpreting
1279 quoted constructs. The most frequent reason one may have to want to know the
1280 details discussed in this section is hairy regular expressions. However, the
1281 first steps of parsing are the same for all Perl quoting operators, so here
1282 they are discussed together.
1284 The most important detail of Perl parsing rules is the first one
1285 discussed below; when processing a quoted construct, Perl I<first>
1286 finds the end of the construct, then it interprets the contents of the
1287 construct. If you understand this rule, you may skip the rest of this
1288 section on the first reading. The other rules would
1289 contradict user's expectations much less frequently than the first one.
1291 Some of the passes discussed below are performed concurrently, but as
1292 far as results are the same, we consider them one-by-one. For different
1293 quoting constructs Perl performs different number of passes, from
1294 one to five, but they are always performed in the same order.
1298 =item Finding the end
1300 First pass is finding the end of the quoted construct, be it
1301 a multichar delimiter
1302 C<"\nEOF\n"> of C<<<EOF> construct, C</> which terminates C<qq/> construct,
1303 C<]> which terminates C<qq[> construct, or C<E<gt>> which terminates a
1304 fileglob started with C<<>.
1306 When searching for one-char non-matching delimiter, such as C</>, combinations
1307 C<\\> and C<\/> are skipped. When searching for one-char matching delimiter,
1308 such as C<]>, combinations C<\\>, C<\]> and C<\[> are skipped, and
1309 nested C<[>, C<]> are skipped as well. When searching for multichar delimiter
1310 no skipping is performed.
1312 For constructs with 3-part delimiters (C<s///> etc.) the search is
1315 During this search no attention is paid to the semantic of the construct,
1318 "$hash{"$foo/$bar"}"
1323 bar # NOT a comment, this slash / terminated m//!
1326 do not form legal quoted expressions, the quoted part ends on the first C<">
1327 and C</>, and the rest happens to be a syntax error. Note that since the slash
1328 which terminated C<m//> was followed by a C<SPACE>, the above is not C<m//x>,
1329 but rather C<m//> with no 'x' switch. So the embedded C<#> is interpreted
1332 =item Removal of backslashes before delimiters
1334 During the second pass the text between the starting delimiter and
1335 the ending delimiter is copied to a safe location, and the C<\> is
1336 removed from combinations consisting of C<\> and delimiter(s) (both starting
1337 and ending delimiter if they differ).
1339 The removal does not happen for multi-char delimiters.
1341 Note that the combination C<\\> is left as it was!
1343 Starting from this step no information about the delimiter(s) is used in the
1348 Next step is interpolation in the obtained delimiter-independent text.
1349 There are four different cases.
1353 =item C<<<'EOF'>, C<m''>, C<s'''>, C<tr///>, C<y///>
1355 No interpolation is performed.
1359 The only interpolation is removal of C<\> from pairs C<\\>.
1361 =item C<"">, C<``>, C<qq//>, C<qx//>, C<<file*globE<gt>>
1363 C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are converted
1364 to corresponding Perl constructs, thus C<"$foo\Qbaz$bar"> is converted to :
1366 $foo . (quotemeta("baz" . $bar));
1368 Other combinations of C<\> with following chars are substituted with
1369 appropriate expansions.
1371 Let it be stressed that I<whatever is between C<\Q> and C<\E>> is interpolated
1372 in the usual way. Say, C<"\Q\\E"> has no C<\E> inside: it has C<\Q>, C<\\>,
1373 and C<E>, thus the result is the same as for C<"\\\\E">. Generally speaking,
1374 having backslashes between C<\Q> and C<\E> may lead to counterintuitive
1375 results. So, C<"\Q\t\E"> is converted to:
1379 which is the same as C<"\\\t"> (since TAB is not alphanumerical). Note also
1385 may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
1387 Interpolated scalars and arrays are internally converted to the C<join> and
1388 C<.> Perl operations, thus C<"$foo >>> '@arr'"> becomes:
1390 $foo . " >>> '" . (join $", @arr) . "'";
1392 All the operations in the above are performed simultaneously left-to-right.
1394 Since the result of "\Q STRING \E" has all the metacharacters quoted
1395 there is no way to insert a literal C<$> or C<@> inside a C<\Q\E> pair: if
1396 protected by C<\> C<$> will be quoted to became "\\\$", if not, it is
1397 interpreted as starting an interpolated scalar.
1399 Note also that the interpolating code needs to make a decision on where the
1400 interpolated scalar ends. For instance, whether C<"a $b -E<gt> {c}"> means:
1402 "a " . $b . " -> {c}";
1408 I<Most of the time> the decision is to take the longest possible text which
1409 does not include spaces between components and contains matching
1410 braces/brackets. Since the outcome may be determined by I<voting> based
1411 on heuristic estimators, the result I<is not strictly predictable>, but
1412 is usually correct for the ambiguous cases.
1414 =item C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
1416 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> and interpolation happens
1417 (almost) as with C<qq//> constructs, but I<the substitution of C<\> followed by
1418 RE-special chars (including C<\>) is not performed>! Moreover,
1419 inside C<(?{BLOCK})>, C<(?# comment )>, and C<#>-comment of
1420 C<//x>-regular expressions no processing is performed at all.
1421 This is the first step where presence of the C<//x> switch is relevant.
1423 Interpolation has several quirks: C<$|>, C<$(> and C<$)> are not interpolated, and
1424 constructs C<$var[SOMETHING]> are I<voted> (by several different estimators)
1425 to be an array element or C<$var> followed by a RE alternative. This is
1426 the place where the notation C<${arr[$bar]}> comes handy: C</${arr[0-9]}/>
1427 is interpreted as an array element C<-9>, not as a regular expression from
1428 variable C<$arr> followed by a digit, which is the interpretation of
1429 C</$arr[0-9]/>. Since voting among different estimators may be performed,
1430 the result I<is not predictable>.
1432 It is on this step that C<\1> is converted to C<$1> in the replacement
1435 Note that absence of processing of C<\\> creates specific restrictions on the
1436 post-processed text: if the delimiter is C</>, one cannot get the combination
1437 C<\/> into the result of this step: C</> will finish the regular expression,
1438 C<\/> will be stripped to C</> on the previous step, and C<\\/> will be left
1439 as is. Since C</> is equivalent to C<\/> inside a regular expression, this
1440 does not matter unless the delimiter is a special character for the RE engine,
1441 as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>, or an alphanumeric char, as in:
1445 In the above RE, which is intentionally obfuscated for illustration, the
1446 delimiter is C<m>, the modifier is C<mx>, and after backslash-removal the
1447 RE is the same as for C<m/ ^ a s* b /mx>).
1451 This step is the last one for all the constructs except regular expressions,
1452 which are processed further.
1454 =item Interpolation of regular expressions
1456 All the previous steps were performed during the compilation of Perl code,
1457 this one happens in run time (though it may be optimized to be calculated
1458 at compile time if appropriate). After all the preprocessing performed
1459 above (and possibly after evaluation if catenation, joining, up/down-casing
1460 and C<quotemeta()>ing are involved) the resulting I<string> is passed to RE
1461 engine for compilation.
1463 Whatever happens in the RE engine is better be discussed in L<perlre>,
1464 but for the sake of continuity let us do it here.
1466 This is another step where presence of the C<//x> switch is relevant.
1467 The RE engine scans the string left-to-right, and converts it to a finite
1470 Backslashed chars are either substituted by corresponding literal
1471 strings (as with C<\{>), or generate special nodes of the finite automaton
1472 (as with C<\b>). Characters which are special to the RE engine (such as
1473 C<|>) generate corresponding nodes or groups of nodes. C<(?#...)>
1474 comments are ignored. All the rest is either converted to literal strings
1475 to match, or is ignored (as is whitespace and C<#>-style comments if
1478 Note that the parsing of the construct C<[...]> is performed using
1479 rather different rules than for the rest of the regular expression.
1480 The terminator of this construct is found using the same rules as for
1481 finding a terminator of a C<{}>-delimited construct, the only exception
1482 being that C<]> immediately following C<[> is considered as if preceded
1483 by a backslash. Similarly, the terminator of C<(?{...})> is found using
1484 the same rules as for finding a terminator of a C<{}>-delimited construct.
1486 It is possible to inspect both the string given to RE engine, and the
1487 resulting finite automaton. See arguments C<debug>/C<debugcolor>
1488 of C<use L<re>> directive, and/or B<-Dr> option of Perl in
1489 L<perlrun/Switches>.
1491 =item Optimization of regular expressions
1493 This step is listed for completeness only. Since it does not change
1494 semantics, details of this step are not documented and are subject
1495 to change. This step is performed over the finite automaton generated
1496 during the previous pass.
1498 However, in older versions of Perl C<L<split>> used to silently
1499 optimize C</^/> to mean C</^/m>. This behaviour, though present
1500 in current versions of Perl, may be deprecated in future.
1504 =head2 I/O Operators
1506 There are several I/O operators you should know about.
1508 A string enclosed by backticks (grave accents) first undergoes
1509 variable substitution just like a double quoted string. It is then
1510 interpreted as a command, and the output of that command is the value
1511 of the pseudo-literal, like in a shell. In scalar context, a single
1512 string consisting of all the output is returned. In list context,
1513 a list of values is returned, one for each line of output. (You can
1514 set C<$/> to use a different line terminator.) The command is executed
1515 each time the pseudo-literal is evaluated. The status value of the
1516 command is returned in C<$?> (see L<perlvar> for the interpretation
1517 of C<$?>). Unlike in B<csh>, no translation is done on the return
1518 data--newlines remain newlines. Unlike in any of the shells, single
1519 quotes do not hide variable names in the command from interpretation.
1520 To pass a $ through to the shell you need to hide it with a backslash.
1521 The generalized form of backticks is C<qx//>. (Because backticks
1522 always undergo shell expansion as well, see L<perlsec> for
1525 In a scalar context, evaluating a filehandle in angle brackets yields the
1526 next line from that file (newline, if any, included), or C<undef> at
1527 end-of-file. When C<$/> is set to C<undef> (i.e. file slurp mode),
1528 and the file is empty, it returns C<''> the first time, followed by
1529 C<undef> subsequently.
1531 Ordinarily you must assign the returned value to a variable, but there is one
1532 situation where an automatic assignment happens. I<If and ONLY if> the
1533 input symbol is the only thing inside the conditional of a C<while> or
1534 C<for(;;)> loop, the value is automatically assigned to the variable
1535 C<$_>. In these loop constructs, the assigned value (whether assignment
1536 is automatic or explicit) is then tested to see if it is defined.
1537 The defined test avoids problems where line has a string value
1538 that would be treated as false by perl e.g. "" or "0" with no trailing
1539 newline. (This may seem like an odd thing to you, but you'll use the
1540 construct in almost every Perl script you write.) Anyway, the following
1541 lines are equivalent to each other:
1543 while (defined($_ = <STDIN>)) { print; }
1544 while ($_ = <STDIN>) { print; }
1545 while (<STDIN>) { print; }
1546 for (;<STDIN>;) { print; }
1547 print while defined($_ = <STDIN>);
1548 print while ($_ = <STDIN>);
1549 print while <STDIN>;
1551 and this also behaves similarly, but avoids the use of $_ :
1553 while (my $line = <STDIN>) { print $line }
1555 If you really mean such values to terminate the loop they should be
1556 tested for explicitly:
1558 while (($_ = <STDIN>) ne '0') { ... }
1559 while (<STDIN>) { last unless $_; ... }
1561 In other boolean contexts, C<E<lt>I<filehandle>E<gt>> without explicit C<defined>
1562 test or comparison will solicit a warning if C<-w> is in effect.
1564 The filehandles STDIN, STDOUT, and STDERR are predefined. (The
1565 filehandles C<stdin>, C<stdout>, and C<stderr> will also work except in
1566 packages, where they would be interpreted as local identifiers rather
1567 than global.) Additional filehandles may be created with the open()
1568 function. See L<perlfunc/open> for details on this.
1570 If a E<lt>FILEHANDLEE<gt> is used in a context that is looking for a list, a
1571 list consisting of all the input lines is returned, one line per list
1572 element. It's easy to make a I<LARGE> data space this way, so use with
1575 E<lt>FILEHANDLEE<gt> may also be spelt readline(FILEHANDLE). See
1576 L<perlfunc/readline>.
1578 The null filehandle E<lt>E<gt> is special and can be used to emulate the
1579 behavior of B<sed> and B<awk>. Input from E<lt>E<gt> comes either from
1580 standard input, or from each file listed on the command line. Here's
1581 how it works: the first time E<lt>E<gt> is evaluated, the @ARGV array is
1582 checked, and if it is empty, C<$ARGV[0]> is set to "-", which when opened
1583 gives you standard input. The @ARGV array is then processed as a list
1584 of filenames. The loop
1587 ... # code for each line
1590 is equivalent to the following Perl-like pseudo code:
1592 unshift(@ARGV, '-') unless @ARGV;
1593 while ($ARGV = shift) {
1596 ... # code for each line
1600 except that it isn't so cumbersome to say, and will actually work. It
1601 really does shift array @ARGV and put the current filename into variable
1602 $ARGV. It also uses filehandle I<ARGV> internally--E<lt>E<gt> is just a
1603 synonym for E<lt>ARGVE<gt>, which is magical. (The pseudo code above
1604 doesn't work because it treats E<lt>ARGVE<gt> as non-magical.)
1606 You can modify @ARGV before the first E<lt>E<gt> as long as the array ends up
1607 containing the list of filenames you really want. Line numbers (C<$.>)
1608 continue as if the input were one big happy file. (But see example
1609 under C<eof> for how to reset line numbers on each file.)
1611 If you want to set @ARGV to your own list of files, go right ahead.
1612 This sets @ARGV to all plain text files if no @ARGV was given:
1614 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
1616 You can even set them to pipe commands. For example, this automatically
1617 filters compressed arguments through B<gzip>:
1619 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
1621 If you want to pass switches into your script, you can use one of the
1622 Getopts modules or put a loop on the front like this:
1624 while ($_ = $ARGV[0], /^-/) {
1627 if (/^-D(.*)/) { $debug = $1 }
1628 if (/^-v/) { $verbose++ }
1629 # ... # other switches
1633 # ... # code for each line
1636 The E<lt>E<gt> symbol will return C<undef> for end-of-file only once.
1637 If you call it again after this it will assume you are processing another
1638 @ARGV list, and if you haven't set @ARGV, will input from STDIN.
1640 If the string inside the angle brackets is a reference to a scalar
1641 variable (e.g., E<lt>$fooE<gt>), then that variable contains the name of the
1642 filehandle to input from, or its typeglob, or a reference to the same. For example:
1647 If what's within the angle brackets is neither a filehandle nor a simple
1648 scalar variable containing a filehandle name, typeglob, or typeglob
1649 reference, it is interpreted as a filename pattern to be globbed, and
1650 either a list of filenames or the next filename in the list is returned,
1651 depending on context. This distinction is determined on syntactic
1652 grounds alone. That means C<E<lt>$xE<gt>> is always a readline from
1653 an indirect handle, but C<E<lt>$hash{key}E<gt>> is always a glob.
1654 That's because $x is a simple scalar variable, but C<$hash{key}> is
1655 not--it's a hash element.
1657 One level of double-quote interpretation is done first, but you can't
1658 say C<E<lt>$fooE<gt>> because that's an indirect filehandle as explained
1659 in the previous paragraph. (In older versions of Perl, programmers
1660 would insert curly brackets to force interpretation as a filename glob:
1661 C<E<lt>${foo}E<gt>>. These days, it's considered cleaner to call the
1662 internal function directly as C<glob($foo)>, which is probably the right
1663 way to have done it in the first place.) Example:
1671 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
1677 In fact, it's currently implemented that way. (Which means it will not
1678 work on filenames with spaces in them unless you have csh(1) on your
1679 machine.) Of course, the shortest way to do the above is:
1683 Because globbing invokes a shell, it's often faster to call readdir() yourself
1684 and do your own grep() on the filenames. Furthermore, due to its current
1685 implementation of using a shell, the glob() routine may get "Arg list too
1686 long" errors (unless you've installed tcsh(1L) as F</bin/csh>).
1688 A glob evaluates its (embedded) argument only when it is starting a new
1689 list. All values must be read before it will start over. In a list
1690 context this isn't important, because you automatically get them all
1691 anyway. In scalar context, however, the operator returns the next value
1692 each time it is called, or a C<undef> value if you've just run out. As
1693 for filehandles an automatic C<defined> is generated when the glob
1694 occurs in the test part of a C<while> or C<for> - because legal glob returns
1695 (e.g. a file called F<0>) would otherwise terminate the loop.
1696 Again, C<undef> is returned only once. So if you're expecting a single value
1697 from a glob, it is much better to say
1699 ($file) = <blurch*>;
1705 because the latter will alternate between returning a filename and
1708 It you're trying to do variable interpolation, it's definitely better
1709 to use the glob() function, because the older notation can cause people
1710 to become confused with the indirect filehandle notation.
1712 @files = glob("$dir/*.[ch]");
1713 @files = glob($files[$i]);
1715 =head2 Constant Folding
1717 Like C, Perl does a certain amount of expression evaluation at
1718 compile time, whenever it determines that all arguments to an
1719 operator are static and have no side effects. In particular, string
1720 concatenation happens at compile time between literals that don't do
1721 variable substitution. Backslash interpretation also happens at
1722 compile time. You can say
1724 'Now is the time for all' . "\n" .
1725 'good men to come to.'
1727 and this all reduces to one string internally. Likewise, if
1730 foreach $file (@filenames) {
1731 if (-s $file > 5 + 100 * 2**16) { }
1734 the compiler will precompute the number that
1735 expression represents so that the interpreter
1738 =head2 Bitwise String Operators
1740 Bitstrings of any size may be manipulated by the bitwise operators
1743 If the operands to a binary bitwise op are strings of different sizes,
1744 B<|> and B<^> ops will act as if the shorter operand had additional
1745 zero bits on the right, while the B<&> op will act as if the longer
1746 operand were truncated to the length of the shorter. Note that the
1747 granularity for such extension or truncation is one or more I<bytes>.
1749 # ASCII-based examples
1750 print "j p \n" ^ " a h"; # prints "JAPH\n"
1751 print "JA" | " ph\n"; # prints "japh\n"
1752 print "japh\nJunk" & '_____'; # prints "JAPH\n";
1753 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
1755 If you are intending to manipulate bitstrings, you should be certain that
1756 you're supplying bitstrings: If an operand is a number, that will imply
1757 a B<numeric> bitwise operation. You may explicitly show which type of
1758 operation you intend by using C<""> or C<0+>, as in the examples below.
1760 $foo = 150 | 105 ; # yields 255 (0x96 | 0x69 is 0xFF)
1761 $foo = '150' | 105 ; # yields 255
1762 $foo = 150 | '105'; # yields 255
1763 $foo = '150' | '105'; # yields string '155' (under ASCII)
1765 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
1766 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
1768 See L<perlfunc/vec> for information on how to manipulate individual bits
1771 =head2 Integer Arithmetic
1773 By default Perl assumes that it must do most of its arithmetic in
1774 floating point. But by saying
1778 you may tell the compiler that it's okay to use integer operations
1779 from here to the end of the enclosing BLOCK. An inner BLOCK may
1780 countermand this by saying
1784 which lasts until the end of that BLOCK.
1786 The bitwise operators ("&", "|", "^", "~", "<<", and ">>") always
1787 produce integral results. (But see also L<Bitwise String Operators>.)
1788 However, C<use integer> still has meaning
1789 for them. By default, their results are interpreted as unsigned
1790 integers. However, if C<use integer> is in effect, their results are
1791 interpreted as signed integers. For example, C<~0> usually evaluates
1792 to a large integral value. However, C<use integer; ~0> is -1 on twos-complement machines.
1794 =head2 Floating-point Arithmetic
1796 While C<use integer> provides integer-only arithmetic, there is no
1797 similar ways to provide rounding or truncation at a certain number of
1798 decimal places. For rounding to a certain number of digits, sprintf()
1799 or printf() is usually the easiest route.
1801 Floating-point numbers are only approximations to what a mathematician
1802 would call real numbers. There are infinitely more reals than floats,
1803 so some corners must be cut. For example:
1805 printf "%.20g\n", 123456789123456789;
1806 # produces 123456789123456784
1808 Testing for exact equality of floating-point equality or inequality is
1809 not a good idea. Here's a (relatively expensive) work-around to compare
1810 whether two floating-point numbers are equal to a particular number of
1811 decimal places. See Knuth, volume II, for a more robust treatment of
1815 my ($X, $Y, $POINTS) = @_;
1817 $tX = sprintf("%.${POINTS}g", $X);
1818 $tY = sprintf("%.${POINTS}g", $Y);
1822 The POSIX module (part of the standard perl distribution) implements
1823 ceil(), floor(), and a number of other mathematical and trigonometric
1824 functions. The Math::Complex module (part of the standard perl
1825 distribution) defines a number of mathematical functions that can also
1826 work on real numbers. Math::Complex not as efficient as POSIX, but
1827 POSIX can't work with complex numbers.
1829 Rounding in financial applications can have serious implications, and
1830 the rounding method used should be specified precisely. In these
1831 cases, it probably pays not to trust whichever system rounding is
1832 being used by Perl, but to instead implement the rounding function you
1835 =head2 Bigger Numbers
1837 The standard Math::BigInt and Math::BigFloat modules provide
1838 variable precision arithmetic and overloaded operators.
1839 At the cost of some space and considerable speed, they
1840 avoid the normal pitfalls associated with limited-precision
1844 $x = Math::BigInt->new('123456789123456789');
1847 # prints +15241578780673678515622620750190521