3 perlop - Perl operators and precedence
7 Perl operators have the following associativity and precedence,
8 listed from highest precedence to lowest. Operators borrowed from
9 C keep the same precedence relationship with each other, even where
10 C's precedence is slightly screwy. (This makes learning Perl easier
11 for C folks.) With very few exceptions, these all operate on scalar
12 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,
68 but the commas on the left are evaluated after. In other words,
69 list operators tend to gobble up all arguments that follow, and
70 then act like a simple TERM with regard to the preceding expression.
71 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 "C<< -> >>" is an infix dereference operator, just as it is in C
99 and C++. If the right side is either a C<[...]>, C<{...}>, or a
100 C<(...)> subscript, then the left side must be either a hard or
101 symbolic reference to an array, a hash, or a subroutine respectively.
102 (Or technically speaking, a location capable of holding a hard
103 reference, if it's an array or hash reference being used for
104 assignment.) See L<perlreftut> and L<perlref>.
106 Otherwise, the right side is a method name or a simple scalar
107 variable containing either the method name or a subroutine reference,
108 and the left side must be either an object (a blessed reference)
109 or a class name (that is, a package name). See L<perlobj>.
111 =head2 Auto-increment and Auto-decrement
113 "++" and "--" work as in C. That is, if placed before a variable, they
114 increment or decrement the variable before returning the value, and if
115 placed after, increment or decrement the variable after returning the value.
117 The auto-increment operator has a little extra builtin magic to it. If
118 you increment a variable that is numeric, or that has ever been used in
119 a numeric context, you get a normal increment. If, however, the
120 variable has been used in only string contexts since it was set, and
121 has a value that is not the empty string and matches the pattern
122 C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
123 character within its range, with carry:
125 print ++($foo = '99'); # prints '100'
126 print ++($foo = 'a0'); # prints 'a1'
127 print ++($foo = 'Az'); # prints 'Ba'
128 print ++($foo = 'zz'); # prints 'aaa'
130 C<undef> is always treated as numeric, and in particular is changed
131 to C<0> before incrementing (so that a post-increment of an undef value
132 will return C<0> rather than C<undef>).
134 The auto-decrement operator is not magical.
136 =head2 Exponentiation
138 Binary "**" is the exponentiation operator. It binds even more
139 tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
140 implemented using C's pow(3) function, which actually works on doubles
143 =head2 Symbolic Unary Operators
145 Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
146 precedence version of this.
148 Unary "-" performs arithmetic negation if the operand is numeric. If
149 the operand is an identifier, a string consisting of a minus sign
150 concatenated with the identifier is returned. Otherwise, if the string
151 starts with a plus or minus, a string starting with the opposite sign
152 is returned. One effect of these rules is that C<-bareword> is equivalent
155 Unary "~" performs bitwise negation, i.e., 1's complement. For
156 example, C<0666 & ~027> is 0640. (See also L<Integer Arithmetic> and
157 L<Bitwise String Operators>.) Note that the width of the result is
158 platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
159 bits wide on a 64-bit platform, so if you are expecting a certain bit
160 width, remember use the & operator to mask off the excess bits.
162 Unary "+" has no effect whatsoever, even on strings. It is useful
163 syntactically for separating a function name from a parenthesized expression
164 that would otherwise be interpreted as the complete list of function
165 arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
167 Unary "\" creates a reference to whatever follows it. See L<perlreftut>
168 and L<perlref>. Do not confuse this behavior with the behavior of
169 backslash within a string, although both forms do convey the notion
170 of protecting the next thing from interpolation.
172 =head2 Binding Operators
174 Binary "=~" binds a scalar expression to a pattern match. Certain operations
175 search or modify the string $_ by default. This operator makes that kind
176 of operation work on some other string. The right argument is a search
177 pattern, substitution, or transliteration. The left argument is what is
178 supposed to be searched, substituted, or transliterated instead of the default
179 $_. When used in scalar context, the return value generally indicates the
180 success of the operation. Behavior in list context depends on the particular
181 operator. See L</"Regexp Quote-Like Operators"> for details.
183 If the right argument is an expression rather than a search pattern,
184 substitution, or transliteration, it is interpreted as a search pattern at run
187 Binary "!~" is just like "=~" except the return value is negated in
190 =head2 Multiplicative Operators
192 Binary "*" multiplies two numbers.
194 Binary "/" divides two numbers.
196 Binary "%" computes the modulus of two numbers. Given integer
197 operands C<$a> and C<$b>: If C<$b> is positive, then C<$a % $b> is
198 C<$a> minus the largest multiple of C<$b> that is not greater than
199 C<$a>. If C<$b> is negative, then C<$a % $b> is C<$a> minus the
200 smallest multiple of C<$b> that is not less than C<$a> (i.e. the
201 result will be less than or equal to zero).
202 Note that when C<use integer> is in scope, "%" gives you direct access
203 to the modulus operator as implemented by your C compiler. This
204 operator is not as well defined for negative operands, but it will
207 Binary "x" is the repetition operator. In scalar context or if the left
208 operand is not enclosed in parentheses, it returns a string consisting
209 of the left operand repeated the number of times specified by the right
210 operand. In list context, if the left operand is enclosed in
211 parentheses, it repeats the list.
213 print '-' x 80; # print row of dashes
215 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
217 @ones = (1) x 80; # a list of 80 1's
218 @ones = (5) x @ones; # set all elements to 5
221 =head2 Additive Operators
223 Binary "+" returns the sum of two numbers.
225 Binary "-" returns the difference of two numbers.
227 Binary "." concatenates two strings.
229 =head2 Shift Operators
231 Binary "<<" returns the value of its left argument shifted left by the
232 number of bits specified by the right argument. Arguments should be
233 integers. (See also L<Integer Arithmetic>.)
235 Binary ">>" returns the value of its left argument shifted right by
236 the number of bits specified by the right argument. Arguments should
237 be integers. (See also L<Integer Arithmetic>.)
239 Note that both "<<" and ">>" in Perl are implemented directly using
240 "<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
241 in force then signed C integers are used, else unsigned C integers are
242 used. Either way, the implementation isn't going to generate results
243 larger than the size of the integer type Perl was built with (32 bits
246 The result of overflowing the range of the integers is undefined
247 because it is undefined also in C. In other words, using 32-bit
248 integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
249 of bits is also undefined.
251 =head2 Named Unary Operators
253 The various named unary operators are treated as functions with one
254 argument, with optional parentheses. These include the filetest
255 operators, like C<-f>, C<-M>, etc. See L<perlfunc>.
257 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
258 is followed by a left parenthesis as the next token, the operator and
259 arguments within parentheses are taken to be of highest precedence,
260 just like a normal function call. For example,
261 because named unary operators are higher precedence than ||:
263 chdir $foo || die; # (chdir $foo) || die
264 chdir($foo) || die; # (chdir $foo) || die
265 chdir ($foo) || die; # (chdir $foo) || die
266 chdir +($foo) || die; # (chdir $foo) || die
268 but, because * is higher precedence than named operators:
270 chdir $foo * 20; # chdir ($foo * 20)
271 chdir($foo) * 20; # (chdir $foo) * 20
272 chdir ($foo) * 20; # (chdir $foo) * 20
273 chdir +($foo) * 20; # chdir ($foo * 20)
275 rand 10 * 20; # rand (10 * 20)
276 rand(10) * 20; # (rand 10) * 20
277 rand (10) * 20; # (rand 10) * 20
278 rand +(10) * 20; # rand (10 * 20)
280 See also L<"Terms and List Operators (Leftward)">.
282 =head2 Relational Operators
284 Binary "<" returns true if the left argument is numerically less than
287 Binary ">" returns true if the left argument is numerically greater
288 than the right argument.
290 Binary "<=" returns true if the left argument is numerically less than
291 or equal to the right argument.
293 Binary ">=" returns true if the left argument is numerically greater
294 than or equal to the right argument.
296 Binary "lt" returns true if the left argument is stringwise less than
299 Binary "gt" returns true if the left argument is stringwise greater
300 than the right argument.
302 Binary "le" returns true if the left argument is stringwise less than
303 or equal to the right argument.
305 Binary "ge" returns true if the left argument is stringwise greater
306 than or equal to the right argument.
308 =head2 Equality Operators
310 Binary "==" returns true if the left argument is numerically equal to
313 Binary "!=" returns true if the left argument is numerically not equal
314 to the right argument.
316 Binary "<=>" returns -1, 0, or 1 depending on whether the left
317 argument is numerically less than, equal to, or greater than the right
318 argument. If your platform supports NaNs (not-a-numbers) as numeric
319 values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
320 "<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
321 returns true, as does NaN != anything else. If your platform doesn't
322 support NaNs then NaN is just a string with numeric value 0.
324 perl -le '$a = NaN; print "No NaN support here" if $a == $a'
325 perl -le '$a = NaN; print "NaN support here" if $a != $a'
327 Binary "eq" returns true if the left argument is stringwise equal to
330 Binary "ne" returns true if the left argument is stringwise not equal
331 to the right argument.
333 Binary "cmp" returns -1, 0, or 1 depending on whether the left
334 argument is stringwise less than, equal to, or greater than the right
337 "lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
338 by the current locale if C<use locale> is in effect. See L<perllocale>.
342 Binary "&" returns its operands ANDed together bit by bit.
343 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
345 Note that "&" has lower priority than relational operators, so for example
346 the brackets are essential in a test like
348 print "Even\n" if ($x & 1) == 0;
350 =head2 Bitwise Or and Exclusive Or
352 Binary "|" returns its operands ORed together bit by bit.
353 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
355 Binary "^" returns its operands XORed together bit by bit.
356 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
358 Note that "|" and "^" have lower priority than relational operators, so
359 for example the brackets are essential in a test like
361 print "false\n" if (8 | 2) != 10;
363 =head2 C-style Logical And
365 Binary "&&" performs a short-circuit logical AND operation. That is,
366 if the left operand is false, the right operand is not even evaluated.
367 Scalar or list context propagates down to the right operand if it
370 =head2 C-style Logical Or
372 Binary "||" performs a short-circuit logical OR operation. That is,
373 if the left operand is true, the right operand is not even evaluated.
374 Scalar or list context propagates down to the right operand if it
377 =head2 C-style Logical Defined-Or
379 Although it has no direct equivalent in C, Perl's C<//> operator is related
380 to its C-style or. In fact, it's exactly the same as C<||>, except that it
381 tests the left hand side's definedness instead of its truth. Thus, C<$a // $b>
382 is similar to C<defined($a) || $b> (except that it returns the value of C<$a>
383 rather than the value of C<defined($a)>) and is exactly equivalent to
384 C<defined($a) ? $a : $b>. This is very useful for providing default values
385 for variables. If you actually want to test if at least one of C<$a> and C<$b> is
386 defined, use C<defined($a // $b)>.
388 The C<||>, C<//> and C<&&> operators differ from C's in that, rather than returning
389 0 or 1, they return the last value evaluated. Thus, a reasonably portable
390 way to find out the home directory might be:
392 $home = $ENV{'HOME'} // $ENV{'LOGDIR'} //
393 (getpwuid($<))[7] // die "You're homeless!\n";
395 In particular, this means that you shouldn't use this
396 for selecting between two aggregates for assignment:
398 @a = @b || @c; # this is wrong
399 @a = scalar(@b) || @c; # really meant this
400 @a = @b ? @b : @c; # this works fine, though
402 As more readable alternatives to C<&&>, C<//> and C<||> when used for
403 control flow, Perl provides C<and>, C<err> and C<or> operators (see below).
404 The short-circuit behavior is identical. The precedence of "and", "err"
405 and "or" is much lower, however, so that you can safely use them after a
406 list operator without the need for parentheses:
408 unlink "alpha", "beta", "gamma"
409 or gripe(), next LINE;
411 With the C-style operators that would have been written like this:
413 unlink("alpha", "beta", "gamma")
414 || (gripe(), next LINE);
416 Using "or" for assignment is unlikely to do what you want; see below.
418 =head2 Range Operators
420 Binary ".." is the range operator, which is really two different
421 operators depending on the context. In list context, it returns an
422 list of values counting (up by ones) from the left value to the right
423 value. If the left value is greater than the right value then it
424 returns the empty array. The range operator is useful for writing
425 C<foreach (1..10)> loops and for doing slice operations on arrays. In
426 the current implementation, no temporary array is created when the
427 range operator is used as the expression in C<foreach> loops, but older
428 versions of Perl might burn a lot of memory when you write something
431 for (1 .. 1_000_000) {
435 The range operator also works on strings, using the magical auto-increment,
438 In scalar context, ".." returns a boolean value. The operator is
439 bistable, like a flip-flop, and emulates the line-range (comma) operator
440 of B<sed>, B<awk>, and various editors. Each ".." operator maintains its
441 own boolean state. It is false as long as its left operand is false.
442 Once the left operand is true, the range operator stays true until the
443 right operand is true, I<AFTER> which the range operator becomes false
444 again. It doesn't become false till the next time the range operator is
445 evaluated. It can test the right operand and become false on the same
446 evaluation it became true (as in B<awk>), but it still returns true once.
447 If you don't want it to test the right operand till the next
448 evaluation, as in B<sed>, just use three dots ("...") instead of
449 two. In all other regards, "..." behaves just like ".." does.
451 The right operand is not evaluated while the operator is in the
452 "false" state, and the left operand is not evaluated while the
453 operator is in the "true" state. The precedence is a little lower
454 than || and &&. The value returned is either the empty string for
455 false, or a sequence number (beginning with 1) for true. The
456 sequence number is reset for each range encountered. The final
457 sequence number in a range has the string "E0" appended to it, which
458 doesn't affect its numeric value, but gives you something to search
459 for if you want to exclude the endpoint. You can exclude the
460 beginning point by waiting for the sequence number to be greater
461 than 1. If either operand of scalar ".." is a constant expression,
462 that operand is implicitly compared to the C<$.> variable, the
463 current line number. Examples:
465 As a scalar operator:
467 if (101 .. 200) { print; } # print 2nd hundred lines
468 next line if (1 .. /^$/); # skip header lines
469 s/^/> / if (/^$/ .. eof()); # quote body
471 # parse mail messages
473 $in_header = 1 .. /^$/;
474 $in_body = /^$/ .. eof();
475 # do something based on those
477 close ARGV if eof; # reset $. each file
482 for (101 .. 200) { print; } # print $_ 100 times
483 @foo = @foo[0 .. $#foo]; # an expensive no-op
484 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
486 The range operator (in list context) makes use of the magical
487 auto-increment algorithm if the operands are strings. You
490 @alphabet = ('A' .. 'Z');
492 to get all normal letters of the English alphabet, or
494 $hexdigit = (0 .. 9, 'a' .. 'f')[$num & 15];
496 to get a hexadecimal digit, or
498 @z2 = ('01' .. '31'); print $z2[$mday];
500 to get dates with leading zeros. If the final value specified is not
501 in the sequence that the magical increment would produce, the sequence
502 goes until the next value would be longer than the final value
505 =head2 Conditional Operator
507 Ternary "?:" is the conditional operator, just as in C. It works much
508 like an if-then-else. If the argument before the ? is true, the
509 argument before the : is returned, otherwise the argument after the :
510 is returned. For example:
512 printf "I have %d dog%s.\n", $n,
513 ($n == 1) ? '' : "s";
515 Scalar or list context propagates downward into the 2nd
516 or 3rd argument, whichever is selected.
518 $a = $ok ? $b : $c; # get a scalar
519 @a = $ok ? @b : @c; # get an array
520 $a = $ok ? @b : @c; # oops, that's just a count!
522 The operator may be assigned to if both the 2nd and 3rd arguments are
523 legal lvalues (meaning that you can assign to them):
525 ($a_or_b ? $a : $b) = $c;
527 Because this operator produces an assignable result, using assignments
528 without parentheses will get you in trouble. For example, this:
530 $a % 2 ? $a += 10 : $a += 2
534 (($a % 2) ? ($a += 10) : $a) += 2
538 ($a % 2) ? ($a += 10) : ($a += 2)
540 That should probably be written more simply as:
542 $a += ($a % 2) ? 10 : 2;
544 =head2 Assignment Operators
546 "=" is the ordinary assignment operator.
548 Assignment operators work as in C. That is,
556 although without duplicating any side effects that dereferencing the lvalue
557 might trigger, such as from tie(). Other assignment operators work similarly.
558 The following are recognized:
565 Although these are grouped by family, they all have the precedence
568 Unlike in C, the scalar assignment operator produces a valid lvalue.
569 Modifying an assignment is equivalent to doing the assignment and
570 then modifying the variable that was assigned to. This is useful
571 for modifying a copy of something, like this:
573 ($tmp = $global) =~ tr [A-Z] [a-z];
584 Similarly, a list assignment in list context produces the list of
585 lvalues assigned to, and a list assignment in scalar context returns
586 the number of elements produced by the expression on the right hand
587 side of the assignment.
589 =head2 Comma Operator
591 Binary "," is the comma operator. In scalar context it evaluates
592 its left argument, throws that value away, then evaluates its right
593 argument and returns that value. This is just like C's comma operator.
595 In list context, it's just the list argument separator, and inserts
596 both its arguments into the list.
598 The => digraph is mostly just a synonym for the comma operator. It's useful for
599 documenting arguments that come in pairs. As of release 5.001, it also forces
600 any word to the left of it to be interpreted as a string.
602 =head2 List Operators (Rightward)
604 On the right side of a list operator, it has very low precedence,
605 such that it controls all comma-separated expressions found there.
606 The only operators with lower precedence are the logical operators
607 "and", "or", and "not", which may be used to evaluate calls to list
608 operators without the need for extra parentheses:
610 open HANDLE, "filename"
611 or die "Can't open: $!\n";
613 See also discussion of list operators in L<Terms and List Operators (Leftward)>.
617 Unary "not" returns the logical negation of the expression to its right.
618 It's the equivalent of "!" except for the very low precedence.
622 Binary "and" returns the logical conjunction of the two surrounding
623 expressions. It's equivalent to && except for the very low
624 precedence. This means that it short-circuits: i.e., the right
625 expression is evaluated only if the left expression is true.
627 =head2 Logical or, Defined or, and Exclusive Or
629 Binary "or" returns the logical disjunction of the two surrounding
630 expressions. It's equivalent to || except for the very low precedence.
631 This makes it useful for control flow
633 print FH $data or die "Can't write to FH: $!";
635 This means that it short-circuits: i.e., the right expression is evaluated
636 only if the left expression is false. Due to its precedence, you should
637 probably avoid using this for assignment, only for control flow.
639 $a = $b or $c; # bug: this is wrong
640 ($a = $b) or $c; # really means this
641 $a = $b || $c; # better written this way
643 However, when it's a list-context assignment and you're trying to use
644 "||" for control flow, you probably need "or" so that the assignment
645 takes higher precedence.
647 @info = stat($file) || die; # oops, scalar sense of stat!
648 @info = stat($file) or die; # better, now @info gets its due
650 Then again, you could always use parentheses.
652 Binary "err" is equivalent to C<//>--it's just like binary "or", except it tests
653 its left argument's definedness instead of its truth. There are two ways to
654 remember "err": either because many functions return C<undef> on an B<err>or,
655 or as a sort of correction: C<$a=($b err 'default')>
657 Binary "xor" returns the exclusive-OR of the two surrounding expressions.
658 It cannot short circuit, of course.
660 =head2 C Operators Missing From Perl
662 Here is what C has that Perl doesn't:
668 Address-of operator. (But see the "\" operator for taking a reference.)
672 Dereference-address operator. (Perl's prefix dereferencing
673 operators are typed: $, @, %, and &.)
677 Type-casting operator.
681 =head2 Quote and Quote-like Operators
683 While we usually think of quotes as literal values, in Perl they
684 function as operators, providing various kinds of interpolating and
685 pattern matching capabilities. Perl provides customary quote characters
686 for these behaviors, but also provides a way for you to choose your
687 quote character for any of them. In the following table, a C<{}> represents
688 any pair of delimiters you choose.
690 Customary Generic Meaning Interpolates
695 // m{} Pattern match yes*
697 s{}{} Substitution yes*
698 tr{}{} Transliteration no (but see below)
701 * unless the delimiter is ''.
703 Non-bracketing delimiters use the same character fore and aft, but the four
704 sorts of brackets (round, angle, square, curly) will all nest, which means
713 Note, however, that this does not always work for quoting Perl code:
715 $s = q{ if($a eq "}") ... }; # WRONG
717 is a syntax error. The C<Text::Balanced> module (from CPAN, and
718 starting from Perl 5.8 part of the standard distribution) is able
721 There can be whitespace between the operator and the quoting
722 characters, except when C<#> is being used as the quoting character.
723 C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
724 operator C<q> followed by a comment. Its argument will be taken
725 from the next line. This allows you to write:
727 s {foo} # Replace foo
730 The following escape sequences are available in constructs that interpolate
731 and in transliterations.
738 \a alarm (bell) (BEL)
740 \033 octal char (ESC)
742 \x{263a} wide hex char (SMILEY)
743 \c[ control char (ESC)
744 \N{name} named Unicode character
746 The following escape sequences are available in constructs that interpolate
747 but not in transliterations.
749 \l lowercase next char
750 \u uppercase next char
753 \E end case modification
754 \Q quote non-word characters till \E
756 If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
757 C<\u> and C<\U> is taken from the current locale. See L<perllocale>.
758 If Unicode (for example, C<\N{}> or wide hex characters of 0x100 or
759 beyond) is being used, the case map used by C<\l>, C<\L>, C<\u> and
760 C<\U> is as defined by Unicode. For documentation of C<\N{name}>,
763 All systems use the virtual C<"\n"> to represent a line terminator,
764 called a "newline". There is no such thing as an unvarying, physical
765 newline character. It is only an illusion that the operating system,
766 device drivers, C libraries, and Perl all conspire to preserve. Not all
767 systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
768 on a Mac, these are reversed, and on systems without line terminator,
769 printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
770 you mean a "newline" for your system, but use the literal ASCII when you
771 need an exact character. For example, most networking protocols expect
772 and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
773 and although they often accept just C<"\012">, they seldom tolerate just
774 C<"\015">. If you get in the habit of using C<"\n"> for networking,
775 you may be burned some day.
777 For constructs that do interpolate, variables beginning with "C<$>"
778 or "C<@>" are interpolated. Subscripted variables such as C<$a[3]> or
779 C<< $href->{key}[0] >> are also interpolated, as are array and hash slices.
780 But method calls such as C<< $obj->meth >> are not.
782 Interpolating an array or slice interpolates the elements in order,
783 separated by the value of C<$">, so is equivalent to interpolating
784 C<join $", @array>. "Punctuation" arrays such as C<@+> are only
785 interpolated if the name is enclosed in braces C<@{+}>.
787 You cannot include a literal C<$> or C<@> within a C<\Q> sequence.
788 An unescaped C<$> or C<@> interpolates the corresponding variable,
789 while escaping will cause the literal string C<\$> to be inserted.
790 You'll need to write something like C<m/\Quser\E\@\Qhost/>.
792 Patterns are subject to an additional level of interpretation as a
793 regular expression. This is done as a second pass, after variables are
794 interpolated, so that regular expressions may be incorporated into the
795 pattern from the variables. If this is not what you want, use C<\Q> to
796 interpolate a variable literally.
798 Apart from the behavior described above, Perl does not expand
799 multiple levels of interpolation. In particular, contrary to the
800 expectations of shell programmers, back-quotes do I<NOT> interpolate
801 within double quotes, nor do single quotes impede evaluation of
802 variables when used within double quotes.
804 =head2 Regexp Quote-Like Operators
806 Here are the quote-like operators that apply to pattern
807 matching and related activities.
813 This is just like the C</pattern/> search, except that it matches only
814 once between calls to the reset() operator. This is a useful
815 optimization when you want to see only the first occurrence of
816 something in each file of a set of files, for instance. Only C<??>
817 patterns local to the current package are reset.
821 # blank line between header and body
824 reset if eof; # clear ?? status for next file
827 This usage is vaguely deprecated, which means it just might possibly
828 be removed in some distant future version of Perl, perhaps somewhere
829 around the year 2168.
831 =item m/PATTERN/cgimosx
833 =item /PATTERN/cgimosx
835 Searches a string for a pattern match, and in scalar context returns
836 true if it succeeds, false if it fails. If no string is specified
837 via the C<=~> or C<!~> operator, the $_ string is searched. (The
838 string specified with C<=~> need not be an lvalue--it may be the
839 result of an expression evaluation, but remember the C<=~> binds
840 rather tightly.) See also L<perlre>. See L<perllocale> for
841 discussion of additional considerations that apply when C<use locale>
846 c Do not reset search position on a failed match when /g is in effect.
847 g Match globally, i.e., find all occurrences.
848 i Do case-insensitive pattern matching.
849 m Treat string as multiple lines.
850 o Compile pattern only once.
851 s Treat string as single line.
852 x Use extended regular expressions.
854 If "/" is the delimiter then the initial C<m> is optional. With the C<m>
855 you can use any pair of non-alphanumeric, non-whitespace characters
856 as delimiters. This is particularly useful for matching path names
857 that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
858 the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
859 If "'" is the delimiter, no interpolation is performed on the PATTERN.
861 PATTERN may contain variables, which will be interpolated (and the
862 pattern recompiled) every time the pattern search is evaluated, except
863 for when the delimiter is a single quote. (Note that C<$(>, C<$)>, and
864 C<$|> are not interpolated because they look like end-of-string tests.)
865 If you want such a pattern to be compiled only once, add a C</o> after
866 the trailing delimiter. This avoids expensive run-time recompilations,
867 and is useful when the value you are interpolating won't change over
868 the life of the script. However, mentioning C</o> constitutes a promise
869 that you won't change the variables in the pattern. If you change them,
870 Perl won't even notice. See also L<"qr/STRING/imosx">.
872 If the PATTERN evaluates to the empty string, the last
873 I<successfully> matched regular expression is used instead. In this
874 case, only the C<g> and C<c> flags on the empty pattern is honoured -
875 the other flags are taken from the original pattern. If no match has
876 previously succeeded, this will (silently) act instead as a genuine
877 empty pattern (which will always match).
879 Note that it's possible to confuse Perl into thinking C<//> (the empty
880 regex) is really C<//> (the defined-or operator). Perl is usually pretty
881 good about this, but some pathological cases might trigger this, such as
882 C<$a///> (is that C<($a) / (//)> or C<$a // />?) and C<print $fh //>
883 (C<print $fh(//> or C<print($fh //>?). In all of these examples, Perl
884 will assume you meant defined-or. If you meant the empty regex, just
885 use parentheses or spaces to disambiguate, or even prefix the empty
886 regex with an C<m> (so C<//> becomes C<m//>).
888 If the C</g> option is not used, C<m//> in list context returns a
889 list consisting of the subexpressions matched by the parentheses in the
890 pattern, i.e., (C<$1>, C<$2>, C<$3>...). (Note that here C<$1> etc. are
891 also set, and that this differs from Perl 4's behavior.) When there are
892 no parentheses in the pattern, the return value is the list C<(1)> for
893 success. With or without parentheses, an empty list is returned upon
898 open(TTY, '/dev/tty');
899 <TTY> =~ /^y/i && foo(); # do foo if desired
901 if (/Version: *([0-9.]*)/) { $version = $1; }
903 next if m#^/usr/spool/uucp#;
908 print if /$arg/o; # compile only once
911 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
913 This last example splits $foo into the first two words and the
914 remainder of the line, and assigns those three fields to $F1, $F2, and
915 $Etc. The conditional is true if any variables were assigned, i.e., if
918 The C</g> modifier specifies global pattern matching--that is,
919 matching as many times as possible within the string. How it behaves
920 depends on the context. In list context, it returns a list of the
921 substrings matched by any capturing parentheses in the regular
922 expression. If there are no parentheses, it returns a list of all
923 the matched strings, as if there were parentheses around the whole
926 In scalar context, each execution of C<m//g> finds the next match,
927 returning true if it matches, and false if there is no further match.
928 The position after the last match can be read or set using the pos()
929 function; see L<perlfunc/pos>. A failed match normally resets the
930 search position to the beginning of the string, but you can avoid that
931 by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
932 string also resets the search position.
934 You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
935 zero-width assertion that matches the exact position where the previous
936 C<m//g>, if any, left off. Without the C</g> modifier, the C<\G> assertion
937 still anchors at pos(), but the match is of course only attempted once.
938 Using C<\G> without C</g> on a target string that has not previously had a
939 C</g> match applied to it is the same as using the C<\A> assertion to match
940 the beginning of the string. Note also that, currently, C<\G> is only
941 properly supported when anchored at the very beginning of the pattern.
946 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
950 while (defined($paragraph = <>)) {
951 while ($paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
955 print "$sentences\n";
957 # using m//gc with \G
961 print $1 while /(o)/gc; print "', pos=", pos, "\n";
963 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
965 print $1 while /(p)/gc; print "', pos=", pos, "\n";
967 print "Final: '$1', pos=",pos,"\n" if /\G(.)/;
969 The last example should print:
979 Notice that the final match matched C<q> instead of C<p>, which a match
980 without the C<\G> anchor would have done. Also note that the final match
981 did not update C<pos> -- C<pos> is only updated on a C</g> match. If the
982 final match did indeed match C<p>, it's a good bet that you're running an
983 older (pre-5.6.0) Perl.
985 A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
986 combine several regexps like this to process a string part-by-part,
987 doing different actions depending on which regexp matched. Each
988 regexp tries to match where the previous one leaves off.
991 $url = new URI::URL "http://www/"; die if $url eq "xXx";
995 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
996 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
997 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
998 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
999 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
1000 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
1001 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
1002 print ". That's all!\n";
1005 Here is the output (split into several lines):
1007 line-noise lowercase line-noise lowercase UPPERCASE line-noise
1008 UPPERCASE line-noise lowercase line-noise lowercase line-noise
1009 lowercase lowercase line-noise lowercase lowercase line-noise
1010 MiXeD line-noise. That's all!
1016 A single-quoted, literal string. A backslash represents a backslash
1017 unless followed by the delimiter or another backslash, in which case
1018 the delimiter or backslash is interpolated.
1020 $foo = q!I said, "You said, 'She said it.'"!;
1021 $bar = q('This is it.');
1022 $baz = '\n'; # a two-character string
1028 A double-quoted, interpolated string.
1031 (*** The previous line contains the naughty word "$1".\n)
1032 if /\b(tcl|java|python)\b/i; # :-)
1033 $baz = "\n"; # a one-character string
1035 =item qr/STRING/imosx
1037 This operator quotes (and possibly compiles) its I<STRING> as a regular
1038 expression. I<STRING> is interpolated the same way as I<PATTERN>
1039 in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1040 is done. Returns a Perl value which may be used instead of the
1041 corresponding C</STRING/imosx> expression.
1045 $rex = qr/my.STRING/is;
1052 The result may be used as a subpattern in a match:
1055 $string =~ /foo${re}bar/; # can be interpolated in other patterns
1056 $string =~ $re; # or used standalone
1057 $string =~ /$re/; # or this way
1059 Since Perl may compile the pattern at the moment of execution of qr()
1060 operator, using qr() may have speed advantages in some situations,
1061 notably if the result of qr() is used standalone:
1064 my $patterns = shift;
1065 my @compiled = map qr/$_/i, @$patterns;
1068 foreach my $pat (@compiled) {
1069 $success = 1, last if /$pat/;
1075 Precompilation of the pattern into an internal representation at
1076 the moment of qr() avoids a need to recompile the pattern every
1077 time a match C</$pat/> is attempted. (Perl has many other internal
1078 optimizations, but none would be triggered in the above example if
1079 we did not use qr() operator.)
1083 i Do case-insensitive pattern matching.
1084 m Treat string as multiple lines.
1085 o Compile pattern only once.
1086 s Treat string as single line.
1087 x Use extended regular expressions.
1089 See L<perlre> for additional information on valid syntax for STRING, and
1090 for a detailed look at the semantics of regular expressions.
1096 A string which is (possibly) interpolated and then executed as a
1097 system command with C</bin/sh> or its equivalent. Shell wildcards,
1098 pipes, and redirections will be honored. The collected standard
1099 output of the command is returned; standard error is unaffected. In
1100 scalar context, it comes back as a single (potentially multi-line)
1101 string, or undef if the command failed. In list context, returns a
1102 list of lines (however you've defined lines with $/ or
1103 $INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
1105 Because backticks do not affect standard error, use shell file descriptor
1106 syntax (assuming the shell supports this) if you care to address this.
1107 To capture a command's STDERR and STDOUT together:
1109 $output = `cmd 2>&1`;
1111 To capture a command's STDOUT but discard its STDERR:
1113 $output = `cmd 2>/dev/null`;
1115 To capture a command's STDERR but discard its STDOUT (ordering is
1118 $output = `cmd 2>&1 1>/dev/null`;
1120 To exchange a command's STDOUT and STDERR in order to capture the STDERR
1121 but leave its STDOUT to come out the old STDERR:
1123 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1125 To read both a command's STDOUT and its STDERR separately, it's easiest
1126 and safest to redirect them separately to files, and then read from those
1127 files when the program is done:
1129 system("program args 1>/tmp/program.stdout 2>/tmp/program.stderr");
1131 Using single-quote as a delimiter protects the command from Perl's
1132 double-quote interpolation, passing it on to the shell instead:
1134 $perl_info = qx(ps $$); # that's Perl's $$
1135 $shell_info = qx'ps $$'; # that's the new shell's $$
1137 How that string gets evaluated is entirely subject to the command
1138 interpreter on your system. On most platforms, you will have to protect
1139 shell metacharacters if you want them treated literally. This is in
1140 practice difficult to do, as it's unclear how to escape which characters.
1141 See L<perlsec> for a clean and safe example of a manual fork() and exec()
1142 to emulate backticks safely.
1144 On some platforms (notably DOS-like ones), the shell may not be
1145 capable of dealing with multiline commands, so putting newlines in
1146 the string may not get you what you want. You may be able to evaluate
1147 multiple commands in a single line by separating them with the command
1148 separator character, if your shell supports that (e.g. C<;> on many Unix
1149 shells; C<&> on the Windows NT C<cmd> shell).
1151 Beginning with v5.6.0, Perl will attempt to flush all files opened for
1152 output before starting the child process, but this may not be supported
1153 on some platforms (see L<perlport>). To be safe, you may need to set
1154 C<$|> ($AUTOFLUSH in English) or call the C<autoflush()> method of
1155 C<IO::Handle> on any open handles.
1157 Beware that some command shells may place restrictions on the length
1158 of the command line. You must ensure your strings don't exceed this
1159 limit after any necessary interpolations. See the platform-specific
1160 release notes for more details about your particular environment.
1162 Using this operator can lead to programs that are difficult to port,
1163 because the shell commands called vary between systems, and may in
1164 fact not be present at all. As one example, the C<type> command under
1165 the POSIX shell is very different from the C<type> command under DOS.
1166 That doesn't mean you should go out of your way to avoid backticks
1167 when they're the right way to get something done. Perl was made to be
1168 a glue language, and one of the things it glues together is commands.
1169 Just understand what you're getting yourself into.
1171 See L<"I/O Operators"> for more discussion.
1175 Evaluates to a list of the words extracted out of STRING, using embedded
1176 whitespace as the word delimiters. It can be understood as being roughly
1179 split(' ', q/STRING/);
1181 the differences being that it generates a real list at compile time, and
1182 in scalar context it returns the last element in the list. So
1187 is semantically equivalent to the list:
1191 Some frequently seen examples:
1193 use POSIX qw( setlocale localeconv )
1194 @EXPORT = qw( foo bar baz );
1196 A common mistake is to try to separate the words with comma or to
1197 put comments into a multi-line C<qw>-string. For this reason, the
1198 C<use warnings> pragma and the B<-w> switch (that is, the C<$^W> variable)
1199 produces warnings if the STRING contains the "," or the "#" character.
1201 =item s/PATTERN/REPLACEMENT/egimosx
1203 Searches a string for a pattern, and if found, replaces that pattern
1204 with the replacement text and returns the number of substitutions
1205 made. Otherwise it returns false (specifically, the empty string).
1207 If no string is specified via the C<=~> or C<!~> operator, the C<$_>
1208 variable is searched and modified. (The string specified with C<=~> must
1209 be scalar variable, an array element, a hash element, or an assignment
1210 to one of those, i.e., an lvalue.)
1212 If the delimiter chosen is a single quote, no interpolation is
1213 done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1214 PATTERN contains a $ that looks like a variable rather than an
1215 end-of-string test, the variable will be interpolated into the pattern
1216 at run-time. If you want the pattern compiled only once the first time
1217 the variable is interpolated, use the C</o> option. If the pattern
1218 evaluates to the empty string, the last successfully executed regular
1219 expression is used instead. See L<perlre> for further explanation on these.
1220 See L<perllocale> for discussion of additional considerations that apply
1221 when C<use locale> is in effect.
1225 e Evaluate the right side as an expression.
1226 g Replace globally, i.e., all occurrences.
1227 i Do case-insensitive pattern matching.
1228 m Treat string as multiple lines.
1229 o Compile pattern only once.
1230 s Treat string as single line.
1231 x Use extended regular expressions.
1233 Any non-alphanumeric, non-whitespace delimiter may replace the
1234 slashes. If single quotes are used, no interpretation is done on the
1235 replacement string (the C</e> modifier overrides this, however). Unlike
1236 Perl 4, Perl 5 treats backticks as normal delimiters; the replacement
1237 text is not evaluated as a command. If the
1238 PATTERN is delimited by bracketing quotes, the REPLACEMENT has its own
1239 pair of quotes, which may or may not be bracketing quotes, e.g.,
1240 C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
1241 replacement portion to be treated as a full-fledged Perl expression
1242 and evaluated right then and there. It is, however, syntax checked at
1243 compile-time. A second C<e> modifier will cause the replacement portion
1244 to be C<eval>ed before being run as a Perl expression.
1248 s/\bgreen\b/mauve/g; # don't change wintergreen
1250 $path =~ s|/usr/bin|/usr/local/bin|;
1252 s/Login: $foo/Login: $bar/; # run-time pattern
1254 ($foo = $bar) =~ s/this/that/; # copy first, then change
1256 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-count
1259 s/\d+/$&*2/e; # yields 'abc246xyz'
1260 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
1261 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1263 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
1264 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
1265 s/^=(\w+)/&pod($1)/ge; # use function call
1267 # expand variables in $_, but dynamics only, using
1268 # symbolic dereferencing
1271 # Add one to the value of any numbers in the string
1274 # This will expand any embedded scalar variable
1275 # (including lexicals) in $_ : First $1 is interpolated
1276 # to the variable name, and then evaluated
1279 # Delete (most) C comments.
1281 /\* # Match the opening delimiter.
1282 .*? # Match a minimal number of characters.
1283 \*/ # Match the closing delimiter.
1286 s/^\s*(.*?)\s*$/$1/; # trim white space in $_, expensively
1288 for ($variable) { # trim white space in $variable, cheap
1293 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
1295 Note the use of $ instead of \ in the last example. Unlike
1296 B<sed>, we use the \<I<digit>> form in only the left hand side.
1297 Anywhere else it's $<I<digit>>.
1299 Occasionally, you can't use just a C</g> to get all the changes
1300 to occur that you might want. Here are two common cases:
1302 # put commas in the right places in an integer
1303 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
1305 # expand tabs to 8-column spacing
1306 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
1308 =item tr/SEARCHLIST/REPLACEMENTLIST/cds
1310 =item y/SEARCHLIST/REPLACEMENTLIST/cds
1312 Transliterates all occurrences of the characters found in the search list
1313 with the corresponding character in the replacement list. It returns
1314 the number of characters replaced or deleted. If no string is
1315 specified via the =~ or !~ operator, the $_ string is transliterated. (The
1316 string specified with =~ must be a scalar variable, an array element, a
1317 hash element, or an assignment to one of those, i.e., an lvalue.)
1319 A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
1320 does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
1321 For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1322 SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1323 its own pair of quotes, which may or may not be bracketing quotes,
1324 e.g., C<tr[A-Z][a-z]> or C<tr(+\-*/)/ABCD/>.
1326 Note that C<tr> does B<not> do regular expression character classes
1327 such as C<\d> or C<[:lower:]>. The <tr> operator is not equivalent to
1328 the tr(1) utility. If you want to map strings between lower/upper
1329 cases, see L<perlfunc/lc> and L<perlfunc/uc>, and in general consider
1330 using the C<s> operator if you need regular expressions.
1332 Note also that the whole range idea is rather unportable between
1333 character sets--and even within character sets they may cause results
1334 you probably didn't expect. A sound principle is to use only ranges
1335 that begin from and end at either alphabets of equal case (a-e, A-E),
1336 or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1337 character sets in full.
1341 c Complement the SEARCHLIST.
1342 d Delete found but unreplaced characters.
1343 s Squash duplicate replaced characters.
1345 If the C</c> modifier is specified, the SEARCHLIST character set
1346 is complemented. If the C</d> modifier is specified, any characters
1347 specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
1348 (Note that this is slightly more flexible than the behavior of some
1349 B<tr> programs, which delete anything they find in the SEARCHLIST,
1350 period.) If the C</s> modifier is specified, sequences of characters
1351 that were transliterated to the same character are squashed down
1352 to a single instance of the character.
1354 If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1355 exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1356 than the SEARCHLIST, the final character is replicated till it is long
1357 enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
1358 This latter is useful for counting characters in a class or for
1359 squashing character sequences in a class.
1363 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1365 $cnt = tr/*/*/; # count the stars in $_
1367 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
1369 $cnt = tr/0-9//; # count the digits in $_
1371 tr/a-zA-Z//s; # bookkeeper -> bokeper
1373 ($HOST = $host) =~ tr/a-z/A-Z/;
1375 tr/a-zA-Z/ /cs; # change non-alphas to single space
1378 [\000-\177]; # delete 8th bit
1380 If multiple transliterations are given for a character, only the
1385 will transliterate any A to X.
1387 Because the transliteration table is built at compile time, neither
1388 the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
1389 interpolation. That means that if you want to use variables, you
1392 eval "tr/$oldlist/$newlist/";
1395 eval "tr/$oldlist/$newlist/, 1" or die $@;
1399 A line-oriented form of quoting is based on the shell "here-document"
1400 syntax. Following a C<< << >> you specify a string to terminate
1401 the quoted material, and all lines following the current line down to
1402 the terminating string are the value of the item. The terminating
1403 string may be either an identifier (a word), or some quoted text. If
1404 quoted, the type of quotes you use determines the treatment of the
1405 text, just as in regular quoting. An unquoted identifier works like
1406 double quotes. There must be no space between the C<< << >> and
1407 the identifier, unless the identifier is quoted. (If you put a space it
1408 will be treated as a null identifier, which is valid, and matches the first
1409 empty line.) The terminating string must appear by itself (unquoted and
1410 with no surrounding whitespace) on the terminating line.
1413 The price is $Price.
1416 print << "EOF"; # same as above
1417 The price is $Price.
1420 print << `EOC`; # execute commands
1425 print <<"foo", <<"bar"; # you can stack them
1431 myfunc(<< "THIS", 23, <<'THAT');
1438 Just don't forget that you have to put a semicolon on the end
1439 to finish the statement, as Perl doesn't know you're not going to
1447 If you want your here-docs to be indented with the
1448 rest of the code, you'll need to remove leading whitespace
1449 from each line manually:
1451 ($quote = <<'FINIS') =~ s/^\s+//gm;
1452 The Road goes ever on and on,
1453 down from the door where it began.
1456 If you use a here-doc within a delimited construct, such as in C<s///eg>,
1457 the quoted material must come on the lines following the final delimiter.
1472 If the terminating identifier is on the last line of the program, you
1473 must be sure there is a newline after it; otherwise, Perl will give the
1474 warning B<Can't find string terminator "END" anywhere before EOF...>.
1476 Additionally, the quoting rules for the identifier are not related to
1477 Perl's quoting rules -- C<q()>, C<qq()>, and the like are not supported
1478 in place of C<''> and C<"">, and the only interpolation is for backslashing
1479 the quoting character:
1481 print << "abc\"def";
1485 Finally, quoted strings cannot span multiple lines. The general rule is
1486 that the identifier must be a string literal. Stick with that, and you
1491 =head2 Gory details of parsing quoted constructs
1493 When presented with something that might have several different
1494 interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
1495 principle to pick the most probable interpretation. This strategy
1496 is so successful that Perl programmers often do not suspect the
1497 ambivalence of what they write. But from time to time, Perl's
1498 notions differ substantially from what the author honestly meant.
1500 This section hopes to clarify how Perl handles quoted constructs.
1501 Although the most common reason to learn this is to unravel labyrinthine
1502 regular expressions, because the initial steps of parsing are the
1503 same for all quoting operators, they are all discussed together.
1505 The most important Perl parsing rule is the first one discussed
1506 below: when processing a quoted construct, Perl first finds the end
1507 of that construct, then interprets its contents. If you understand
1508 this rule, you may skip the rest of this section on the first
1509 reading. The other rules are likely to contradict the user's
1510 expectations much less frequently than this first one.
1512 Some passes discussed below are performed concurrently, but because
1513 their results are the same, we consider them individually. For different
1514 quoting constructs, Perl performs different numbers of passes, from
1515 one to five, but these passes are always performed in the same order.
1519 =item Finding the end
1521 The first pass is finding the end of the quoted construct, whether
1522 it be a multicharacter delimiter C<"\nEOF\n"> in the C<<<EOF>
1523 construct, a C</> that terminates a C<qq//> construct, a C<]> which
1524 terminates C<qq[]> construct, or a C<< > >> which terminates a
1525 fileglob started with C<< < >>.
1527 When searching for single-character non-pairing delimiters, such
1528 as C</>, combinations of C<\\> and C<\/> are skipped. However,
1529 when searching for single-character pairing delimiter like C<[>,
1530 combinations of C<\\>, C<\]>, and C<\[> are all skipped, and nested
1531 C<[>, C<]> are skipped as well. When searching for multicharacter
1532 delimiters, nothing is skipped.
1534 For constructs with three-part delimiters (C<s///>, C<y///>, and
1535 C<tr///>), the search is repeated once more.
1537 During this search no attention is paid to the semantics of the construct.
1540 "$hash{"$foo/$bar"}"
1545 bar # NOT a comment, this slash / terminated m//!
1548 do not form legal quoted expressions. The quoted part ends on the
1549 first C<"> and C</>, and the rest happens to be a syntax error.
1550 Because the slash that terminated C<m//> was followed by a C<SPACE>,
1551 the example above is not C<m//x>, but rather C<m//> with no C</x>
1552 modifier. So the embedded C<#> is interpreted as a literal C<#>.
1554 =item Removal of backslashes before delimiters
1556 During the second pass, text between the starting and ending
1557 delimiters is copied to a safe location, and the C<\> is removed
1558 from combinations consisting of C<\> and delimiter--or delimiters,
1559 meaning both starting and ending delimiters will should these differ.
1560 This removal does not happen for multi-character delimiters.
1561 Note that the combination C<\\> is left intact, just as it was.
1563 Starting from this step no information about the delimiters is
1568 The next step is interpolation in the text obtained, which is now
1569 delimiter-independent. There are four different cases.
1573 =item C<<<'EOF'>, C<m''>, C<s'''>, C<tr///>, C<y///>
1575 No interpolation is performed.
1579 The only interpolation is removal of C<\> from pairs C<\\>.
1581 =item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>
1583 C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
1584 converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
1585 is converted to C<$foo . (quotemeta("baz" . $bar))> internally.
1586 The other combinations are replaced with appropriate expansions.
1588 Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
1589 is interpolated in the usual way. Something like C<"\Q\\E"> has
1590 no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
1591 result is the same as for C<"\\\\E">. As a general rule, backslashes
1592 between C<\Q> and C<\E> may lead to counterintuitive results. So,
1593 C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
1594 as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
1599 may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
1601 Interpolated scalars and arrays are converted internally to the C<join> and
1602 C<.> catenation operations. Thus, C<"$foo XXX '@arr'"> becomes:
1604 $foo . " XXX '" . (join $", @arr) . "'";
1606 All operations above are performed simultaneously, left to right.
1608 Because the result of C<"\Q STRING \E"> has all metacharacters
1609 quoted, there is no way to insert a literal C<$> or C<@> inside a
1610 C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to became
1611 C<"\\\$">; if not, it is interpreted as the start of an interpolated
1614 Note also that the interpolation code needs to make a decision on
1615 where the interpolated scalar ends. For instance, whether
1616 C<< "a $b -> {c}" >> really means:
1618 "a " . $b . " -> {c}";
1624 Most of the time, the longest possible text that does not include
1625 spaces between components and which contains matching braces or
1626 brackets. because the outcome may be determined by voting based
1627 on heuristic estimators, the result is not strictly predictable.
1628 Fortunately, it's usually correct for ambiguous cases.
1630 =item C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
1632 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
1633 happens (almost) as with C<qq//> constructs, but the substitution
1634 of C<\> followed by RE-special chars (including C<\>) is not
1635 performed. Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
1636 a C<#>-comment in a C<//x>-regular expression, no processing is
1637 performed whatsoever. This is the first step at which the presence
1638 of the C<//x> modifier is relevant.
1640 Interpolation has several quirks: C<$|>, C<$(>, and C<$)> are not
1641 interpolated, and constructs C<$var[SOMETHING]> are voted (by several
1642 different estimators) to be either an array element or C<$var>
1643 followed by an RE alternative. This is where the notation
1644 C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
1645 array element C<-9>, not as a regular expression from the variable
1646 C<$arr> followed by a digit, which would be the interpretation of
1647 C</$arr[0-9]/>. Since voting among different estimators may occur,
1648 the result is not predictable.
1650 It is at this step that C<\1> is begrudgingly converted to C<$1> in
1651 the replacement text of C<s///> to correct the incorrigible
1652 I<sed> hackers who haven't picked up the saner idiom yet. A warning
1653 is emitted if the C<use warnings> pragma or the B<-w> command-line flag
1654 (that is, the C<$^W> variable) was set.
1656 The lack of processing of C<\\> creates specific restrictions on
1657 the post-processed text. If the delimiter is C</>, one cannot get
1658 the combination C<\/> into the result of this step. C</> will
1659 finish the regular expression, C<\/> will be stripped to C</> on
1660 the previous step, and C<\\/> will be left as is. Because C</> is
1661 equivalent to C<\/> inside a regular expression, this does not
1662 matter unless the delimiter happens to be character special to the
1663 RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
1664 alphanumeric char, as in:
1668 In the RE above, which is intentionally obfuscated for illustration, the
1669 delimiter is C<m>, the modifier is C<mx>, and after backslash-removal the
1670 RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
1671 reason you're encouraged to restrict your delimiters to non-alphanumeric,
1672 non-whitespace choices.
1676 This step is the last one for all constructs except regular expressions,
1677 which are processed further.
1679 =item Interpolation of regular expressions
1681 Previous steps were performed during the compilation of Perl code,
1682 but this one happens at run time--although it may be optimized to
1683 be calculated at compile time if appropriate. After preprocessing
1684 described above, and possibly after evaluation if catenation,
1685 joining, casing translation, or metaquoting are involved, the
1686 resulting I<string> is passed to the RE engine for compilation.
1688 Whatever happens in the RE engine might be better discussed in L<perlre>,
1689 but for the sake of continuity, we shall do so here.
1691 This is another step where the presence of the C<//x> modifier is
1692 relevant. The RE engine scans the string from left to right and
1693 converts it to a finite automaton.
1695 Backslashed characters are either replaced with corresponding
1696 literal strings (as with C<\{>), or else they generate special nodes
1697 in the finite automaton (as with C<\b>). Characters special to the
1698 RE engine (such as C<|>) generate corresponding nodes or groups of
1699 nodes. C<(?#...)> comments are ignored. All the rest is either
1700 converted to literal strings to match, or else is ignored (as is
1701 whitespace and C<#>-style comments if C<//x> is present).
1703 Parsing of the bracketed character class construct, C<[...]>, is
1704 rather different than the rule used for the rest of the pattern.
1705 The terminator of this construct is found using the same rules as
1706 for finding the terminator of a C<{}>-delimited construct, the only
1707 exception being that C<]> immediately following C<[> is treated as
1708 though preceded by a backslash. Similarly, the terminator of
1709 C<(?{...})> is found using the same rules as for finding the
1710 terminator of a C<{}>-delimited construct.
1712 It is possible to inspect both the string given to RE engine and the
1713 resulting finite automaton. See the arguments C<debug>/C<debugcolor>
1714 in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
1715 switch documented in L<perlrun/"Command Switches">.
1717 =item Optimization of regular expressions
1719 This step is listed for completeness only. Since it does not change
1720 semantics, details of this step are not documented and are subject
1721 to change without notice. This step is performed over the finite
1722 automaton that was generated during the previous pass.
1724 It is at this stage that C<split()> silently optimizes C</^/> to
1729 =head2 I/O Operators
1731 There are several I/O operators you should know about.
1733 A string enclosed by backticks (grave accents) first undergoes
1734 double-quote interpolation. It is then interpreted as an external
1735 command, and the output of that command is the value of the
1736 backtick string, like in a shell. In scalar context, a single string
1737 consisting of all output is returned. In list context, a list of
1738 values is returned, one per line of output. (You can set C<$/> to use
1739 a different line terminator.) The command is executed each time the
1740 pseudo-literal is evaluated. The status value of the command is
1741 returned in C<$?> (see L<perlvar> for the interpretation of C<$?>).
1742 Unlike in B<csh>, no translation is done on the return data--newlines
1743 remain newlines. Unlike in any of the shells, single quotes do not
1744 hide variable names in the command from interpretation. To pass a
1745 literal dollar-sign through to the shell you need to hide it with a
1746 backslash. The generalized form of backticks is C<qx//>. (Because
1747 backticks always undergo shell expansion as well, see L<perlsec> for
1750 In scalar context, evaluating a filehandle in angle brackets yields
1751 the next line from that file (the newline, if any, included), or
1752 C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
1753 (sometimes known as file-slurp mode) and the file is empty, it
1754 returns C<''> the first time, followed by C<undef> subsequently.
1756 Ordinarily you must assign the returned value to a variable, but
1757 there is one situation where an automatic assignment happens. If
1758 and only if the input symbol is the only thing inside the conditional
1759 of a C<while> statement (even if disguised as a C<for(;;)> loop),
1760 the value is automatically assigned to the global variable $_,
1761 destroying whatever was there previously. (This may seem like an
1762 odd thing to you, but you'll use the construct in almost every Perl
1763 script you write.) The $_ variable is not implicitly localized.
1764 You'll have to put a C<local $_;> before the loop if you want that
1767 The following lines are equivalent:
1769 while (defined($_ = <STDIN>)) { print; }
1770 while ($_ = <STDIN>) { print; }
1771 while (<STDIN>) { print; }
1772 for (;<STDIN>;) { print; }
1773 print while defined($_ = <STDIN>);
1774 print while ($_ = <STDIN>);
1775 print while <STDIN>;
1777 This also behaves similarly, but avoids $_ :
1779 while (my $line = <STDIN>) { print $line }
1781 In these loop constructs, the assigned value (whether assignment
1782 is automatic or explicit) is then tested to see whether it is
1783 defined. The defined test avoids problems where line has a string
1784 value that would be treated as false by Perl, for example a "" or
1785 a "0" with no trailing newline. If you really mean for such values
1786 to terminate the loop, they should be tested for explicitly:
1788 while (($_ = <STDIN>) ne '0') { ... }
1789 while (<STDIN>) { last unless $_; ... }
1791 In other boolean contexts, C<< <I<filehandle>> >> without an
1792 explicit C<defined> test or comparison elicit a warning if the
1793 C<use warnings> pragma or the B<-w>
1794 command-line switch (the C<$^W> variable) is in effect.
1796 The filehandles STDIN, STDOUT, and STDERR are predefined. (The
1797 filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
1798 in packages, where they would be interpreted as local identifiers
1799 rather than global.) Additional filehandles may be created with
1800 the open() function, amongst others. See L<perlopentut> and
1801 L<perlfunc/open> for details on this.
1803 If a <FILEHANDLE> is used in a context that is looking for
1804 a list, a list comprising all input lines is returned, one line per
1805 list element. It's easy to grow to a rather large data space this
1806 way, so use with care.
1808 <FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
1809 See L<perlfunc/readline>.
1811 The null filehandle <> is special: it can be used to emulate the
1812 behavior of B<sed> and B<awk>. Input from <> comes either from
1813 standard input, or from each file listed on the command line. Here's
1814 how it works: the first time <> is evaluated, the @ARGV array is
1815 checked, and if it is empty, C<$ARGV[0]> is set to "-", which when opened
1816 gives you standard input. The @ARGV array is then processed as a list
1817 of filenames. The loop
1820 ... # code for each line
1823 is equivalent to the following Perl-like pseudo code:
1825 unshift(@ARGV, '-') unless @ARGV;
1826 while ($ARGV = shift) {
1829 ... # code for each line
1833 except that it isn't so cumbersome to say, and will actually work.
1834 It really does shift the @ARGV array and put the current filename
1835 into the $ARGV variable. It also uses filehandle I<ARGV>
1836 internally--<> is just a synonym for <ARGV>, which
1837 is magical. (The pseudo code above doesn't work because it treats
1838 <ARGV> as non-magical.)
1840 You can modify @ARGV before the first <> as long as the array ends up
1841 containing the list of filenames you really want. Line numbers (C<$.>)
1842 continue as though the input were one big happy file. See the example
1843 in L<perlfunc/eof> for how to reset line numbers on each file.
1845 If you want to set @ARGV to your own list of files, go right ahead.
1846 This sets @ARGV to all plain text files if no @ARGV was given:
1848 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
1850 You can even set them to pipe commands. For example, this automatically
1851 filters compressed arguments through B<gzip>:
1853 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
1855 If you want to pass switches into your script, you can use one of the
1856 Getopts modules or put a loop on the front like this:
1858 while ($_ = $ARGV[0], /^-/) {
1861 if (/^-D(.*)/) { $debug = $1 }
1862 if (/^-v/) { $verbose++ }
1863 # ... # other switches
1867 # ... # code for each line
1870 The <> symbol will return C<undef> for end-of-file only once.
1871 If you call it again after this, it will assume you are processing another
1872 @ARGV list, and if you haven't set @ARGV, will read input from STDIN.
1874 If what the angle brackets contain is a simple scalar variable (e.g.,
1875 <$foo>), then that variable contains the name of the
1876 filehandle to input from, or its typeglob, or a reference to the
1882 If what's within the angle brackets is neither a filehandle nor a simple
1883 scalar variable containing a filehandle name, typeglob, or typeglob
1884 reference, it is interpreted as a filename pattern to be globbed, and
1885 either a list of filenames or the next filename in the list is returned,
1886 depending on context. This distinction is determined on syntactic
1887 grounds alone. That means C<< <$x> >> is always a readline() from
1888 an indirect handle, but C<< <$hash{key}> >> is always a glob().
1889 That's because $x is a simple scalar variable, but C<$hash{key}> is
1890 not--it's a hash element.
1892 One level of double-quote interpretation is done first, but you can't
1893 say C<< <$foo> >> because that's an indirect filehandle as explained
1894 in the previous paragraph. (In older versions of Perl, programmers
1895 would insert curly brackets to force interpretation as a filename glob:
1896 C<< <${foo}> >>. These days, it's considered cleaner to call the
1897 internal function directly as C<glob($foo)>, which is probably the right
1898 way to have done it in the first place.) For example:
1904 is roughly equivalent to:
1906 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
1912 except that the globbing is actually done internally using the standard
1913 C<File::Glob> extension. Of course, the shortest way to do the above is:
1917 A (file)glob evaluates its (embedded) argument only when it is
1918 starting a new list. All values must be read before it will start
1919 over. In list context, this isn't important because you automatically
1920 get them all anyway. However, in scalar context the operator returns
1921 the next value each time it's called, or C<undef> when the list has
1922 run out. As with filehandle reads, an automatic C<defined> is
1923 generated when the glob occurs in the test part of a C<while>,
1924 because legal glob returns (e.g. a file called F<0>) would otherwise
1925 terminate the loop. Again, C<undef> is returned only once. So if
1926 you're expecting a single value from a glob, it is much better to
1929 ($file) = <blurch*>;
1935 because the latter will alternate between returning a filename and
1938 If you're trying to do variable interpolation, it's definitely better
1939 to use the glob() function, because the older notation can cause people
1940 to become confused with the indirect filehandle notation.
1942 @files = glob("$dir/*.[ch]");
1943 @files = glob($files[$i]);
1945 =head2 Constant Folding
1947 Like C, Perl does a certain amount of expression evaluation at
1948 compile time whenever it determines that all arguments to an
1949 operator are static and have no side effects. In particular, string
1950 concatenation happens at compile time between literals that don't do
1951 variable substitution. Backslash interpolation also happens at
1952 compile time. You can say
1954 'Now is the time for all' . "\n" .
1955 'good men to come to.'
1957 and this all reduces to one string internally. Likewise, if
1960 foreach $file (@filenames) {
1961 if (-s $file > 5 + 100 * 2**16) { }
1964 the compiler will precompute the number which that expression
1965 represents so that the interpreter won't have to.
1967 =head2 Bitwise String Operators
1969 Bitstrings of any size may be manipulated by the bitwise operators
1972 If the operands to a binary bitwise op are strings of different
1973 sizes, B<|> and B<^> ops act as though the shorter operand had
1974 additional zero bits on the right, while the B<&> op acts as though
1975 the longer operand were truncated to the length of the shorter.
1976 The granularity for such extension or truncation is one or more
1979 # ASCII-based examples
1980 print "j p \n" ^ " a h"; # prints "JAPH\n"
1981 print "JA" | " ph\n"; # prints "japh\n"
1982 print "japh\nJunk" & '_____'; # prints "JAPH\n";
1983 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
1985 If you are intending to manipulate bitstrings, be certain that
1986 you're supplying bitstrings: If an operand is a number, that will imply
1987 a B<numeric> bitwise operation. You may explicitly show which type of
1988 operation you intend by using C<""> or C<0+>, as in the examples below.
1990 $foo = 150 | 105 ; # yields 255 (0x96 | 0x69 is 0xFF)
1991 $foo = '150' | 105 ; # yields 255
1992 $foo = 150 | '105'; # yields 255
1993 $foo = '150' | '105'; # yields string '155' (under ASCII)
1995 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
1996 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
1998 See L<perlfunc/vec> for information on how to manipulate individual bits
2001 =head2 Integer Arithmetic
2003 By default, Perl assumes that it must do most of its arithmetic in
2004 floating point. But by saying
2008 you may tell the compiler that it's okay to use integer operations
2009 (if it feels like it) from here to the end of the enclosing BLOCK.
2010 An inner BLOCK may countermand this by saying
2014 which lasts until the end of that BLOCK. Note that this doesn't
2015 mean everything is only an integer, merely that Perl may use integer
2016 operations if it is so inclined. For example, even under C<use
2017 integer>, if you take the C<sqrt(2)>, you'll still get C<1.4142135623731>
2020 Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
2021 and ">>") always produce integral results. (But see also
2022 L<Bitwise String Operators>.) However, C<use integer> still has meaning for
2023 them. By default, their results are interpreted as unsigned integers, but
2024 if C<use integer> is in effect, their results are interpreted
2025 as signed integers. For example, C<~0> usually evaluates to a large
2026 integral value. However, C<use integer; ~0> is C<-1> on twos-complement
2029 =head2 Floating-point Arithmetic
2031 While C<use integer> provides integer-only arithmetic, there is no
2032 analogous mechanism to provide automatic rounding or truncation to a
2033 certain number of decimal places. For rounding to a certain number
2034 of digits, sprintf() or printf() is usually the easiest route.
2037 Floating-point numbers are only approximations to what a mathematician
2038 would call real numbers. There are infinitely more reals than floats,
2039 so some corners must be cut. For example:
2041 printf "%.20g\n", 123456789123456789;
2042 # produces 123456789123456784
2044 Testing for exact equality of floating-point equality or inequality is
2045 not a good idea. Here's a (relatively expensive) work-around to compare
2046 whether two floating-point numbers are equal to a particular number of
2047 decimal places. See Knuth, volume II, for a more robust treatment of
2051 my ($X, $Y, $POINTS) = @_;
2053 $tX = sprintf("%.${POINTS}g", $X);
2054 $tY = sprintf("%.${POINTS}g", $Y);
2058 The POSIX module (part of the standard perl distribution) implements
2059 ceil(), floor(), and other mathematical and trigonometric functions.
2060 The Math::Complex module (part of the standard perl distribution)
2061 defines mathematical functions that work on both the reals and the
2062 imaginary numbers. Math::Complex not as efficient as POSIX, but
2063 POSIX can't work with complex numbers.
2065 Rounding in financial applications can have serious implications, and
2066 the rounding method used should be specified precisely. In these
2067 cases, it probably pays not to trust whichever system rounding is
2068 being used by Perl, but to instead implement the rounding function you
2071 =head2 Bigger Numbers
2073 The standard Math::BigInt and Math::BigFloat modules provide
2074 variable-precision arithmetic and overloaded operators, although
2075 they're currently pretty slow. At the cost of some space and
2076 considerable speed, they avoid the normal pitfalls associated with
2077 limited-precision representations.
2080 $x = Math::BigInt->new('123456789123456789');
2083 # prints +15241578780673678515622620750190521
2085 There are several modules that let you calculate with (bound only by
2086 memory and cpu-time) unlimited or fixed precision. There are also
2087 some non-standard modules that provide faster implementations via
2088 external C libraries.
2090 Here is a short, but incomplete summary:
2092 Math::Fraction big, unlimited fractions like 9973 / 12967
2093 Math::String treat string sequences like numbers
2094 Math::FixedPrecision calculate with a fixed precision
2095 Math::Currency for currency calculations
2096 Bit::Vector manipulate bit vectors fast (uses C)
2097 Math::BigIntFast Bit::Vector wrapper for big numbers
2098 Math::Pari provides access to the Pari C library
2099 Math::BigInteger uses an external C library
2100 Math::Cephes uses external Cephes C library (no big numbers)
2101 Math::Cephes::Fraction fractions via the Cephes library
2102 Math::GMP another one using an external C library