.\" -*- mode: troff; coding: utf-8 -*- .\" Automatically generated by Pod::Man 5.01 (Pod::Simple 3.43) .\" .\" Standard preamble: .\" ======================================================================== .de Sp \" Vertical space (when we can't use .PP) .if t .sp .5v .if n .sp .. .de Vb \" Begin verbatim text .ft CW .nf .ne \\$1 .. .de Ve \" End verbatim text .ft R .fi .. .\" \*(C` and \*(C' are quotes in nroff, nothing in troff, for use with C<>. .ie n \{\ . ds C` "" . ds C' "" 'br\} .el\{\ . ds C` . ds C' 'br\} .\" .\" Escape single quotes in literal strings from groff's Unicode transform. .ie \n(.g .ds Aq \(aq .el .ds Aq ' .\" .\" If the F register is >0, we'll generate index entries on stderr for .\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index .\" entries marked with X<> in POD. Of course, you'll have to process the .\" output yourself in some meaningful fashion. .\" .\" Avoid warning from groff about undefined register 'F'. .de IX .. .nr rF 0 .if \n(.g .if rF .nr rF 1 .if (\n(rF:(\n(.g==0)) \{\ . if \nF \{\ . de IX . tm Index:\\$1\t\\n%\t"\\$2" .. . if !\nF==2 \{\ . nr % 0 . nr F 2 . \} . \} .\} .rr rF .\" ======================================================================== .\" .IX Title "PERLOP 1" .TH PERLOP 1 2024-01-25 "perl v5.38.2" "Perl Programmers Reference Guide" .\" For nroff, turn off justification. Always turn off hyphenation; it makes .\" way too many mistakes in technical documents. .if n .ad l .nh .SH NAME perlop \- Perl operators and precedence .IX Xref "operator" .SH DESCRIPTION .IX Header "DESCRIPTION" In Perl, the operator determines what operation is performed, independent of the type of the operands. For example \f(CW\*(C`$x\ +\ $y\*(C'\fR is always a numeric addition, and if \f(CW$x\fR or \f(CW$y\fR do not contain numbers, an attempt is made to convert them to numbers first. .PP This is in contrast to many other dynamic languages, where the operation is determined by the type of the first argument. It also means that Perl has two versions of some operators, one for numeric and one for string comparison. For example \f(CW\*(C`$x\ ==\ $y\*(C'\fR compares two numbers for equality, and \f(CW\*(C`$x\ eq\ $y\*(C'\fR compares two strings. .PP There are a few exceptions though: \f(CW\*(C`x\*(C'\fR can be either string repetition or list repetition, depending on the type of the left operand, and \f(CW\*(C`&\*(C'\fR, \f(CW\*(C`|\*(C'\fR, \f(CW\*(C`^\*(C'\fR and \f(CW\*(C`~\*(C'\fR can be either string or numeric bit operations. .SS "Operator Precedence and Associativity" .IX Xref "operator, precedence precedence associativity" .IX Subsection "Operator Precedence and Associativity" Operator precedence and associativity work in Perl more or less like they do in mathematics. .PP \&\fIOperator precedence\fR means some operators group more tightly than others. For example, in \f(CW\*(C`2 + 4 * 5\*(C'\fR, the multiplication has higher precedence, so \f(CW\*(C`4 * 5\*(C'\fR is grouped together as the right-hand operand of the addition, rather than \f(CW\*(C`2 + 4\*(C'\fR being grouped together as the left-hand operand of the multiplication. It is as if the expression were written \f(CW\*(C`2 + (4 * 5)\*(C'\fR, not \&\f(CW\*(C`(2 + 4) * 5\*(C'\fR. So the expression yields \f(CW\*(C`2 + 20 == 22\*(C'\fR, rather than \&\f(CW\*(C`6 * 5 == 30\*(C'\fR. .PP \&\fIOperator associativity\fR defines what happens if a sequence of the same operators is used one after another: usually that they will be grouped at the left or the right. For example, in \f(CW\*(C`9 \- 3 \- 2\*(C'\fR, subtraction is left associative, so \f(CW\*(C`9 \- 3\*(C'\fR is grouped together as the left-hand operand of the second subtraction, rather than \f(CW\*(C`3 \- 2\*(C'\fR being grouped together as the right-hand operand of the first subtraction. It is as if the expression were written \&\f(CW\*(C`(9 \- 3) \- 2\*(C'\fR, not \f(CW\*(C`9 \- (3 \- 2)\*(C'\fR. So the expression yields \f(CW\*(C`6 \- 2 == 4\*(C'\fR, rather than \f(CW\*(C`9 \- 1 == 8\*(C'\fR. .PP For simple operators that evaluate all their operands and then combine the values in some way, precedence and associativity (and parentheses) imply some ordering requirements on those combining operations. For example, in \f(CW2 + 4 * 5\fR, the grouping implied by precedence means that the multiplication of 4 and 5 must be performed before the addition of 2 and 20, simply because the result of that multiplication is required as one of the operands of the addition. But the order of operations is not fully determined by this: in \f(CW\*(C`2 * 2 + 4 * 5\*(C'\fR both multiplications must be performed before the addition, but the grouping does not say anything about the order in which the two multiplications are performed. In fact Perl has a general rule that the operands of an operator are evaluated in left-to-right order. A few operators such as \f(CW\*(C`&&=\*(C'\fR have special evaluation rules that can result in an operand not being evaluated at all; in general, the top-level operator in an expression has control of operand evaluation. .PP Some comparison operators, as their associativity, \fIchain\fR with some operators of the same precedence (but never with operators of different precedence). This chaining means that each comparison is performed on the two arguments surrounding it, with each interior argument taking part in two comparisons, and the comparison results are implicitly ANDed. Thus \f(CW"$x\ <\ $y\ <=\ $z"\fR behaves exactly like \f(CW"$x\ <\ $y\ &&\ $y\ <=\ $z"\fR, assuming that \f(CW"$y"\fR is as simple a scalar as it looks. The ANDing short-circuits just like \f(CW"&&"\fR does, stopping the sequence of comparisons as soon as one yields false. .PP In a chained comparison, each argument expression is evaluated at most once, even if it takes part in two comparisons, but the result of the evaluation is fetched for each comparison. (It is not evaluated at all if the short-circuiting means that it's not required for any comparisons.) This matters if the computation of an interior argument is expensive or non-deterministic. For example, .PP .Vb 1 \& if($x < expensive_sub() <= $z) { ... .Ve .PP is not entirely like .PP .Vb 1 \& if($x < expensive_sub() && expensive_sub() <= $z) { ... .Ve .PP but instead closer to .PP .Vb 2 \& my $tmp = expensive_sub(); \& if($x < $tmp && $tmp <= $z) { ... .Ve .PP in that the subroutine is only called once. However, it's not exactly like this latter code either, because the chained comparison doesn't actually involve any temporary variable (named or otherwise): there is no assignment. This doesn't make much difference where the expression is a call to an ordinary subroutine, but matters more with an lvalue subroutine, or if the argument expression yields some unusual kind of scalar by other means. For example, if the argument expression yields a tied scalar, then the expression is evaluated to produce that scalar at most once, but the value of that scalar may be fetched up to twice, once for each comparison in which it is actually used. .PP In this example, the expression is evaluated only once, and the tied scalar (the result of the expression) is fetched for each comparison that uses it. .PP .Vb 1 \& if ($x < $tied_scalar < $z) { ... .Ve .PP In the next example, the expression is evaluated only once, and the tied scalar is fetched once as part of the operation within the expression. The result of that operation is fetched for each comparison, which normally doesn't matter unless that expression result is also magical due to operator overloading. .PP .Vb 1 \& if ($x < $tied_scalar + 42 < $z) { ... .Ve .PP Some operators are instead non-associative, meaning that it is a syntax error to use a sequence of those operators of the same precedence. For example, \f(CW"$x\ ..\ $y\ ..\ $z"\fR is an error. .PP Perl operators have the following associativity and precedence, listed from highest precedence to lowest. Operators borrowed from C keep the same precedence relationship with each other, even where C's precedence is slightly screwy. (This makes learning Perl easier for C folks.) With very few exceptions, these all operate on scalar values only, not array values. .PP .Vb 10 \& left terms and list operators (leftward) \& left \-> \& nonassoc ++ \-\- \& right ** \& right ! ~ ~. \e and unary + and \- \& left =~ !~ \& left * / % x \& left + \- . \& left << >> \& nonassoc named unary operators \& nonassoc isa \& chained < > <= >= lt gt le ge \& chain/na == != eq ne <=> cmp ~~ \& left & &. \& left | |. ^ ^. \& left && \& left || // \& nonassoc .. ... \& right ?: \& right = += \-= *= etc. goto last next redo dump \& left , => \& nonassoc list operators (rightward) \& right not \& left and \& left or xor .Ve .PP In the following sections, these operators are covered in detail, in the same order in which they appear in the table above. .PP Many operators can be overloaded for objects. See overload. .SS "Terms and List Operators (Leftward)" .IX Xref "list operator operator, list term" .IX Subsection "Terms and List Operators (Leftward)" A TERM has the highest precedence in Perl. They include variables, quote and quote-like operators, any expression in parentheses, and any function whose arguments are parenthesized. Actually, there aren't really functions in this sense, just list operators and unary operators behaving as functions because you put parentheses around the arguments. These are all documented in perlfunc. .PP If any list operator (\f(CWprint()\fR, etc.) or any unary operator (\f(CWchdir()\fR, etc.) is followed by a left parenthesis as the next token, the operator and arguments within parentheses are taken to be of highest precedence, just like a normal function call. .PP In the absence of parentheses, the precedence of list operators such as \&\f(CW\*(C`print\*(C'\fR, \f(CW\*(C`sort\*(C'\fR, or \f(CW\*(C`chmod\*(C'\fR is either very high or very low depending on whether you are looking at the left side or the right side of the operator. For example, in .PP .Vb 2 \& @ary = (1, 3, sort 4, 2); \& print @ary; # prints 1324 .Ve .PP the commas on the right of the \f(CW\*(C`sort\*(C'\fR are evaluated before the \f(CW\*(C`sort\*(C'\fR, but the commas on the left are evaluated after. In other words, list operators tend to gobble up all arguments that follow, and then act like a simple TERM with regard to the preceding expression. Be careful with parentheses: .PP .Vb 3 \& # These evaluate exit before doing the print: \& print($foo, exit); # Obviously not what you want. \& print $foo, exit; # Nor is this. \& \& # These do the print before evaluating exit: \& (print $foo), exit; # This is what you want. \& print($foo), exit; # Or this. \& print ($foo), exit; # Or even this. .Ve .PP Also note that .PP .Vb 1 \& print ($foo & 255) + 1, "\en"; .Ve .PP probably doesn't do what you expect at first glance. The parentheses enclose the argument list for \f(CW\*(C`print\*(C'\fR which is evaluated (printing the result of \f(CW\*(C`$foo\ &\ 255\*(C'\fR). Then one is added to the return value of \f(CW\*(C`print\*(C'\fR (usually 1). The result is something like this: .PP .Vb 1 \& 1 + 1, "\en"; # Obviously not what you meant. .Ve .PP To do what you meant properly, you must write: .PP .Vb 1 \& print(($foo & 255) + 1, "\en"); .Ve .PP See "Named Unary Operators" for more discussion of this. .PP Also parsed as terms are the \f(CW\*(C`do\ {}\*(C'\fR and \f(CW\*(C`eval\ {}\*(C'\fR constructs, as well as subroutine and method calls, and the anonymous constructors \f(CW\*(C`[]\*(C'\fR and \f(CW\*(C`{}\*(C'\fR. .PP See also "Quote and Quote-like Operators" toward the end of this section, as well as "I/O Operators". .SS "The Arrow Operator" .IX Xref "arrow dereference ->" .IX Subsection "The Arrow Operator" "\f(CW\*(C`\->\*(C'\fR" is an infix dereference operator, just as it is in C and C++. If the right side is either a \f(CW\*(C`[...]\*(C'\fR, \f(CW\*(C`{...}\*(C'\fR, or a \&\f(CW\*(C`(...)\*(C'\fR subscript, then the left side must be either a hard or symbolic reference to an array, a hash, or a subroutine respectively. (Or technically speaking, a location capable of holding a hard reference, if it's an array or hash reference being used for assignment.) See perlreftut and perlref. .PP Otherwise, the right side is a method name or a simple scalar variable containing either the method name or a subroutine reference, and (if it is a method name) the left side must be either an object (a blessed reference) or a class name (that is, a package name). See perlobj. .PP The dereferencing cases (as opposed to method-calling cases) are somewhat extended by the \f(CW\*(C`postderef\*(C'\fR feature. For the details of that feature, consult "Postfix Dereference Syntax" in perlref. .SS "Auto-increment and Auto-decrement" .IX Xref "increment auto-increment ++ decrement auto-decrement --" .IX Subsection "Auto-increment and Auto-decrement" \&\f(CW"++"\fR and \f(CW"\-\-"\fR work as in C. That is, if placed before a variable, they increment or decrement the variable by one before returning the value, and if placed after, increment or decrement after returning the value. .PP .Vb 3 \& $i = 0; $j = 0; \& print $i++; # prints 0 \& print ++$j; # prints 1 .Ve .PP Note that just as in C, Perl doesn't define \fBwhen\fR the variable is incremented or decremented. You just know it will be done sometime before or after the value is returned. This also means that modifying a variable twice in the same statement will lead to undefined behavior. Avoid statements like: .PP .Vb 2 \& $i = $i ++; \& print ++ $i + $i ++; .Ve .PP Perl will not guarantee what the result of the above statements is. .PP The auto-increment operator has a little extra builtin magic to it. If you increment a variable that is numeric, or that has ever been used in a numeric context, you get a normal increment. If, however, the variable has been used in only string contexts since it was set, and has a value that is not the empty string and matches the pattern \&\f(CW\*(C`/^[a\-zA\-Z]*[0\-9]*\ez/\*(C'\fR, the increment is done as a string, preserving each character within its range, with carry: .PP .Vb 4 \& print ++($foo = "99"); # prints "100" \& print ++($foo = "a0"); # prints "a1" \& print ++($foo = "Az"); # prints "Ba" \& print ++($foo = "zz"); # prints "aaa" .Ve .PP \&\f(CW\*(C`undef\*(C'\fR is always treated as numeric, and in particular is changed to \f(CW0\fR before incrementing (so that a post-increment of an undef value will return \f(CW0\fR rather than \f(CW\*(C`undef\*(C'\fR). .PP The auto-decrement operator is not magical. .SS Exponentiation .IX Xref "** exponentiation power" .IX Subsection "Exponentiation" Binary \f(CW"**"\fR is the exponentiation operator. It binds even more tightly than unary minus, so \f(CW\*(C`\-2**4\*(C'\fR is \f(CW\*(C`\-(2**4)\*(C'\fR, not \f(CW\*(C`(\-2)**4\*(C'\fR. (This is implemented using C's \f(CWpow(3)\fR function, which actually works on doubles internally.) .PP Note that certain exponentiation expressions are ill-defined: these include \f(CW\*(C`0**0\*(C'\fR, \f(CW\*(C`1**Inf\*(C'\fR, and \f(CW\*(C`Inf**0\*(C'\fR. Do not expect any particular results from these special cases, the results are platform-dependent. .SS "Symbolic Unary Operators" .IX Xref "unary operator operator, unary" .IX Subsection "Symbolic Unary Operators" Unary \f(CW"!"\fR performs logical negation, that is, "not". See also \&\f(CW\*(C`not\*(C'\fR for a lower precedence version of this. .IX Xref "!" .PP Unary \f(CW"\-"\fR performs arithmetic negation if the operand is numeric, including any string that looks like a number. If the operand is an identifier, a string consisting of a minus sign concatenated with the identifier is returned. Otherwise, if the string starts with a plus or minus, a string starting with the opposite sign is returned. One effect of these rules is that \f(CW\*(C`\-bareword\*(C'\fR is equivalent to the string \f(CW"\-bareword"\fR. If, however, the string begins with a non-alphabetic character (excluding \f(CW"+"\fR or \f(CW"\-"\fR), Perl will attempt to convert the string to a numeric, and the arithmetic negation is performed. If the string cannot be cleanly converted to a numeric, Perl will give the warning \&\fBArgument "the string" isn't numeric in negation (\-) at ...\fR. .IX Xref "- negation, arithmetic" .PP Unary \f(CW"~"\fR performs bitwise negation, that is, 1's complement. For example, \f(CW\*(C`0666\ &\ ~027\*(C'\fR is 0640. (See also "Integer Arithmetic" and "Bitwise String Operators".) Note that the width of the result is platform-dependent: \f(CW\*(C`~0\*(C'\fR is 32 bits wide on a 32\-bit platform, but 64 bits wide on a 64\-bit platform, so if you are expecting a certain bit width, remember to use the \f(CW"&"\fR operator to mask off the excess bits. .IX Xref "~ negation, binary" .PP Starting in Perl 5.28, it is a fatal error to try to complement a string containing a character with an ordinal value above 255. .PP If the "bitwise" feature is enabled via \f(CW\*(C`use\ feature\ \*(Aqbitwise\*(Aq\*(C'\fR or \f(CW\*(C`use v5.28\*(C'\fR, then unary \&\f(CW"~"\fR always treats its argument as a number, and an alternate form of the operator, \f(CW"~."\fR, always treats its argument as a string. So \f(CW\*(C`~0\*(C'\fR and \f(CW\*(C`~"0"\*(C'\fR will both give 2**32\-1 on 32\-bit platforms, whereas \f(CW\*(C`~.0\*(C'\fR and \f(CW\*(C`~."0"\*(C'\fR will both yield \f(CW"\exff"\fR. Until Perl 5.28, this feature produced a warning in the \f(CW"experimental::bitwise"\fR category. .PP Unary \f(CW"+"\fR has no effect whatsoever, even on strings. It is useful syntactically for separating a function name from a parenthesized expression that would otherwise be interpreted as the complete list of function arguments. (See examples above under "Terms and List Operators (Leftward)".) .IX Xref "+" .PP Unary \f(CW"\e"\fR creates references. If its operand is a single sigilled thing, it creates a reference to that object. If its operand is a parenthesised list, then it creates references to the things mentioned in the list. Otherwise it puts its operand in list context, and creates a list of references to the scalars in the list provided by the operand. See perlreftut and perlref. Do not confuse this behavior with the behavior of backslash within a string, although both forms do convey the notion of protecting the next thing from interpolation. .IX Xref "\\ reference backslash" .SS "Binding Operators" .IX Xref "binding operator, binding =~ !~" .IX Subsection "Binding Operators" Binary \f(CW"=~"\fR binds a scalar expression to a pattern match. Certain operations search or modify the string \f(CW$_\fR by default. This operator makes that kind of operation work on some other string. The right argument is a search pattern, substitution, or transliteration. The left argument is what is supposed to be searched, substituted, or transliterated instead of the default \&\f(CW$_\fR. When used in scalar context, the return value generally indicates the success of the operation. The exceptions are substitution (\f(CW\*(C`s///\*(C'\fR) and transliteration (\f(CW\*(C`y///\*(C'\fR) with the \f(CW\*(C`/r\*(C'\fR (non-destructive) option, which cause the \fBr\fReturn value to be the result of the substitution. Behavior in list context depends on the particular operator. See "Regexp Quote-Like Operators" for details and perlretut for examples using these operators. .PP If the right argument is an expression rather than a search pattern, substitution, or transliteration, it is interpreted as a search pattern at run time. Note that this means that its contents will be interpolated twice, so .PP .Vb 1 \& \*(Aq\e\e\*(Aq =~ q\*(Aq\e\e\*(Aq; .Ve .PP is not ok, as the regex engine will end up trying to compile the pattern \f(CW\*(C`\e\*(C'\fR, which it will consider a syntax error. .PP Binary \f(CW"!~"\fR is just like \f(CW"=~"\fR except the return value is negated in the logical sense. .PP Binary \f(CW"!~"\fR with a non-destructive substitution (\f(CW\*(C`s///r\*(C'\fR) or transliteration (\f(CW\*(C`y///r\*(C'\fR) is a syntax error. .SS "Multiplicative Operators" .IX Xref "operator, multiplicative" .IX Subsection "Multiplicative Operators" Binary \f(CW"*"\fR multiplies two numbers. .IX Xref "*" .PP Binary \f(CW"/"\fR divides two numbers. .IX Xref "slash" .PP Binary \f(CW"%"\fR is the modulo operator, which computes the division remainder of its first argument with respect to its second argument. Given integer operands \f(CW$m\fR and \f(CW$n\fR: If \f(CW$n\fR is positive, then \f(CW\*(C`$m\ %\ $n\*(C'\fR is \&\f(CW$m\fR minus the largest multiple of \f(CW$n\fR less than or equal to \&\f(CW$m\fR. If \f(CW$n\fR is negative, then \f(CW\*(C`$m\ %\ $n\*(C'\fR is \f(CW$m\fR minus the smallest multiple of \f(CW$n\fR that is not less than \f(CW$m\fR (that is, the result will be less than or equal to zero). If the operands \&\f(CW$m\fR and \f(CW$n\fR are floating point values and the absolute value of \&\f(CW$n\fR (that is \f(CWabs($n)\fR) is less than \f(CW\*(C`(UV_MAX\ +\ 1)\*(C'\fR, only the integer portion of \f(CW$m\fR and \f(CW$n\fR will be used in the operation (Note: here \f(CW\*(C`UV_MAX\*(C'\fR means the maximum of the unsigned integer type). If the absolute value of the right operand (\f(CWabs($n)\fR) is greater than or equal to \f(CW\*(C`(UV_MAX\ +\ 1)\*(C'\fR, \f(CW"%"\fR computes the floating-point remainder \&\f(CW$r\fR in the equation \f(CW\*(C`($r\ =\ $m\ \-\ $i*$n)\*(C'\fR where \f(CW$i\fR is a certain integer that makes \f(CW$r\fR have the same sign as the right operand \&\f(CW$n\fR (\fBnot\fR as the left operand \f(CW$m\fR like C function \f(CWfmod()\fR) and the absolute value less than that of \f(CW$n\fR. Note that when \f(CW\*(C`use\ integer\*(C'\fR is in scope, \f(CW"%"\fR gives you direct access to the modulo operator as implemented by your C compiler. This operator is not as well defined for negative operands, but it will execute faster. .IX Xref "% remainder modulo mod" .PP Binary \f(CW\*(C`x\*(C'\fR is the repetition operator. In scalar context, or if the left operand is neither enclosed in parentheses nor a \f(CW\*(C`qw//\*(C'\fR list, it performs a string repetition. In that case it supplies scalar context to the left operand, and returns a string consisting of the left operand string repeated the number of times specified by the right operand. If the \f(CW\*(C`x\*(C'\fR is in list context, and the left operand is either enclosed in parentheses or a \f(CW\*(C`qw//\*(C'\fR list, it performs a list repetition. In that case it supplies list context to the left operand, and returns a list consisting of the left operand list repeated the number of times specified by the right operand. If the right operand is zero or negative (raising a warning on negative), it returns an empty string or an empty list, depending on the context. .IX Xref "x" .PP .Vb 1 \& print \*(Aq\-\*(Aq x 80; # print row of dashes \& \& print "\et" x ($tab/8), \*(Aq \*(Aq x ($tab%8); # tab over \& \& @ones = (1) x 80; # a list of 80 1\*(Aqs \& @ones = (5) x @ones; # set all elements to 5 .Ve .SS "Additive Operators" .IX Xref "operator, additive" .IX Subsection "Additive Operators" Binary \f(CW"+"\fR returns the sum of two numbers. .IX Xref "+" .PP Binary \f(CW"\-"\fR returns the difference of two numbers. .IX Xref "-" .PP Binary \f(CW"."\fR concatenates two strings. .IX Xref "string, concatenation concatenation cat concat concatenate ." .SS "Shift Operators" .IX Xref "shift operator operator, shift << >> right shift left shift bitwise shift shl shr shift, right shift, left" .IX Subsection "Shift Operators" Binary \f(CW"<<"\fR returns the value of its left argument shifted left by the number of bits specified by the right argument. Arguments should be integers. (See also "Integer Arithmetic".) .PP Binary \f(CW">>"\fR returns the value of its left argument shifted right by the number of bits specified by the right argument. Arguments should be integers. (See also "Integer Arithmetic".) .PP If \f(CW\*(C`use\ integer\*(C'\fR (see "Integer Arithmetic") is in force then signed C integers are used (\fIarithmetic shift\fR), otherwise unsigned C integers are used (\fIlogical shift\fR), even for negative shiftees. In arithmetic right shift the sign bit is replicated on the left, in logical shift zero bits come in from the left. .PP Either way, the implementation isn't going to generate results larger than the size of the integer type Perl was built with (32 bits or 64 bits). .PP Shifting by negative number of bits means the reverse shift: left shift becomes right shift, right shift becomes left shift. This is unlike in C, where negative shift is undefined. .PP Shifting by more bits than the size of the integers means most of the time zero (all bits fall off), except that under \f(CW\*(C`use\ integer\*(C'\fR right overshifting a negative shiftee results in \-1. This is unlike in C, where shifting by too many bits is undefined. A common C behavior is "shift by modulo wordbits", so that for example .PP .Vb 1 \& 1 >> 64 == 1 >> (64 % 64) == 1 >> 0 == 1 # Common C behavior. .Ve .PP but that is completely accidental. .PP If you get tired of being subject to your platform's native integers, the \f(CW\*(C`use\ bigint\*(C'\fR pragma neatly sidesteps the issue altogether: .PP .Vb 5 \& print 20 << 20; # 20971520 \& print 20 << 40; # 5120 on 32\-bit machines, \& # 21990232555520 on 64\-bit machines \& use bigint; \& print 20 << 100; # 25353012004564588029934064107520 .Ve .SS "Named Unary Operators" .IX Xref "operator, named unary" .IX Subsection "Named Unary Operators" The various named unary operators are treated as functions with one argument, with optional parentheses. .PP If any list operator (\f(CWprint()\fR, etc.) or any unary operator (\f(CWchdir()\fR, etc.) is followed by a left parenthesis as the next token, the operator and arguments within parentheses are taken to be of highest precedence, just like a normal function call. For example, because named unary operators are higher precedence than \f(CW\*(C`||\*(C'\fR: .PP .Vb 4 \& chdir $foo || die; # (chdir $foo) || die \& chdir($foo) || die; # (chdir $foo) || die \& chdir ($foo) || die; # (chdir $foo) || die \& chdir +($foo) || die; # (chdir $foo) || die .Ve .PP but, because \f(CW"*"\fR is higher precedence than named operators: .PP .Vb 4 \& chdir $foo * 20; # chdir ($foo * 20) \& chdir($foo) * 20; # (chdir $foo) * 20 \& chdir ($foo) * 20; # (chdir $foo) * 20 \& chdir +($foo) * 20; # chdir ($foo * 20) \& \& rand 10 * 20; # rand (10 * 20) \& rand(10) * 20; # (rand 10) * 20 \& rand (10) * 20; # (rand 10) * 20 \& rand +(10) * 20; # rand (10 * 20) .Ve .PP Regarding precedence, the filetest operators, like \f(CW\*(C`\-f\*(C'\fR, \f(CW\*(C`\-M\*(C'\fR, etc. are treated like named unary operators, but they don't follow this functional parenthesis rule. That means, for example, that \f(CW\*(C`\-f($file).".bak"\*(C'\fR is equivalent to \f(CW\*(C`\-f\ "$file.bak"\*(C'\fR. .IX Xref "-X filetest operator, filetest" .PP See also "Terms and List Operators (Leftward)". .SS "Relational Operators" .IX Xref "relational operator operator, relational" .IX Subsection "Relational Operators" Perl operators that return true or false generally return values that can be safely used as numbers. For example, the relational operators in this section and the equality operators in the next one return \f(CW1\fR for true and a special version of the defined empty string, \f(CW""\fR, which counts as a zero but is exempt from warnings about improper numeric conversions, just as \f(CW"0\ but\ true"\fR is. .PP Binary \f(CW"<"\fR returns true if the left argument is numerically less than the right argument. .IX Xref "<" .PP Binary \f(CW">"\fR returns true if the left argument is numerically greater than the right argument. .IX Xref ">" .PP Binary \f(CW"<="\fR returns true if the left argument is numerically less than or equal to the right argument. .IX Xref "<=" .PP Binary \f(CW">="\fR returns true if the left argument is numerically greater than or equal to the right argument. .IX Xref ">=" .PP Binary \f(CW"lt"\fR returns true if the left argument is stringwise less than the right argument. .IX Xref "lt" .PP Binary \f(CW"gt"\fR returns true if the left argument is stringwise greater than the right argument. .IX Xref "gt" .PP Binary \f(CW"le"\fR returns true if the left argument is stringwise less than or equal to the right argument. .IX Xref "le" .PP Binary \f(CW"ge"\fR returns true if the left argument is stringwise greater than or equal to the right argument. .IX Xref "ge" .PP A sequence of relational operators, such as \f(CW"$x\ <\ $y\ <=\ $z"\fR, performs chained comparisons, in the manner described above in the section "Operator Precedence and Associativity". Beware that they do not chain with equality operators, which have lower precedence. .SS "Equality Operators" .IX Xref "equality equal equals operator, equality" .IX Subsection "Equality Operators" Binary \f(CW"=="\fR returns true if the left argument is numerically equal to the right argument. .IX Xref "==" .PP Binary \f(CW"!="\fR returns true if the left argument is numerically not equal to the right argument. .IX Xref "!=" .PP Binary \f(CW"eq"\fR returns true if the left argument is stringwise equal to the right argument. .IX Xref "eq" .PP Binary \f(CW"ne"\fR returns true if the left argument is stringwise not equal to the right argument. .IX Xref "ne" .PP A sequence of the above equality operators, such as \f(CW"$x\ ==\ $y\ ==\ $z"\fR, performs chained comparisons, in the manner described above in the section "Operator Precedence and Associativity". Beware that they do not chain with relational operators, which have higher precedence. .PP Binary \f(CW"<=>"\fR returns \-1, 0, or 1 depending on whether the left argument is numerically less than, equal to, or greater than the right argument. If your platform supports \f(CW\*(C`NaN\*(C'\fR's (not-a-numbers) as numeric values, using them with \f(CW"<=>"\fR returns undef. \f(CW\*(C`NaN\*(C'\fR is not \&\f(CW"<"\fR, \f(CW"=="\fR, \f(CW">"\fR, \f(CW"<="\fR or \f(CW">="\fR anything (even \f(CW\*(C`NaN\*(C'\fR), so those 5 return false. \f(CW\*(C`NaN\ !=\ NaN\*(C'\fR returns true, as does \f(CW\*(C`NaN\ !=\*(C'\fR\ \fIanything\ else\fR. If your platform doesn't support \f(CW\*(C`NaN\*(C'\fR's then \f(CW\*(C`NaN\*(C'\fR is just a string with numeric value 0. .IX Xref "<=> spaceship" .PP .Vb 2 \& $ perl \-le \*(Aq$x = "NaN"; print "No NaN support here" if $x == $x\*(Aq \& $ perl \-le \*(Aq$x = "NaN"; print "NaN support here" if $x != $x\*(Aq .Ve .PP (Note that the bigint, bigrat, and bignum pragmas all support \f(CW"NaN"\fR.) .PP Binary \f(CW"cmp"\fR returns \-1, 0, or 1 depending on whether the left argument is stringwise less than, equal to, or greater than the right argument. .PP Here we can see the difference between <=> and cmp, .PP .Vb 2 \& print 10 <=> 2 #prints 1 \& print 10 cmp 2 #prints \-1 .Ve .PP (likewise between gt and >, lt and <, etc.) .IX Xref "cmp" .PP Binary \f(CW"~~"\fR does a smartmatch between its arguments. Smart matching is described in the next section. .IX Xref "~~" .PP The two-sided ordering operators \f(CW"<=>"\fR and \f(CW"cmp"\fR, and the smartmatch operator \f(CW"~~"\fR, are non-associative with respect to each other and with respect to the equality operators of the same precedence. .PP \&\f(CW"lt"\fR, \f(CW"le"\fR, \f(CW"ge"\fR, \f(CW"gt"\fR and \f(CW"cmp"\fR use the collation (sort) order specified by the current \f(CW\*(C`LC_COLLATE\*(C'\fR locale if a \f(CW\*(C`use\ locale\*(C'\fR form that includes collation is in effect. See perllocale. Do not mix these with Unicode, only use them with legacy 8\-bit locale encodings. The standard \f(CW\*(C`Unicode::Collate\*(C'\fR and \&\f(CW\*(C`Unicode::Collate::Locale\*(C'\fR modules offer much more powerful solutions to collation issues. .PP For case-insensitive comparisons, look at the "fc" in perlfunc case-folding function, available in Perl v5.16 or later: .PP .Vb 1 \& if ( fc($x) eq fc($y) ) { ... } .Ve .SS "Class Instance Operator" .IX Xref "isa operator" .IX Subsection "Class Instance Operator" Binary \f(CW\*(C`isa\*(C'\fR evaluates to true when the left argument is an object instance of the class (or a subclass derived from that class) given by the right argument. If the left argument is not defined, not a blessed object instance, nor does not derive from the class given by the right argument, the operator evaluates as false. The right argument may give the class either as a bareword or a scalar expression that yields a string class name: .PP .Vb 1 \& if( $obj isa Some::Class ) { ... } \& \& if( $obj isa "Different::Class" ) { ... } \& if( $obj isa $name_of_class ) { ... } .Ve .PP This feature is available from Perl 5.31.6 onwards when enabled by \&\f(CW\*(C`use feature \*(Aqisa\*(Aq\*(C'\fR. This feature is enabled automatically by a \&\f(CW\*(C`use v5.36\*(C'\fR (or higher) declaration in the current scope. .SS "Smartmatch Operator" .IX Subsection "Smartmatch Operator" First available in Perl 5.10.1 (the 5.10.0 version behaved differently), binary \f(CW\*(C`~~\*(C'\fR does a "smartmatch" between its arguments. This is mostly used implicitly in the \f(CW\*(C`when\*(C'\fR construct described in perlsyn, although not all \f(CW\*(C`when\*(C'\fR clauses call the smartmatch operator. Unique among all of Perl's operators, the smartmatch operator can recurse. The smartmatch operator is experimental and its behavior is subject to change. .PP It is also unique in that all other Perl operators impose a context (usually string or numeric context) on their operands, autoconverting those operands to those imposed contexts. In contrast, smartmatch \&\fIinfers\fR contexts from the actual types of its operands and uses that type information to select a suitable comparison mechanism. .PP The \f(CW\*(C`~~\*(C'\fR operator compares its operands "polymorphically", determining how to compare them according to their actual types (numeric, string, array, hash, etc.). Like the equality operators with which it shares the same precedence, \f(CW\*(C`~~\*(C'\fR returns 1 for true and \f(CW""\fR for false. It is often best read aloud as "in", "inside of", or "is contained in", because the left operand is often looked for \fIinside\fR the right operand. That makes the order of the operands to the smartmatch operand often opposite that of the regular match operator. In other words, the "smaller" thing is usually placed in the left operand and the larger one in the right. .PP The behavior of a smartmatch depends on what type of things its arguments are, as determined by the following table. The first row of the table whose types apply determines the smartmatch behavior. Because what actually happens is mostly determined by the type of the second operand, the table is sorted on the right operand instead of on the left. .PP .Vb 4 \& Left Right Description and pseudocode \& =============================================================== \& Any undef check whether Any is undefined \& like: !defined Any \& \& Any Object invoke ~~ overloading on Object, or die \& \& Right operand is an ARRAY: \& \& Left Right Description and pseudocode \& =============================================================== \& ARRAY1 ARRAY2 recurse on paired elements of ARRAY1 and ARRAY2[2] \& like: (ARRAY1[0] ~~ ARRAY2[0]) \& && (ARRAY1[1] ~~ ARRAY2[1]) && ... \& HASH ARRAY any ARRAY elements exist as HASH keys \& like: grep { exists HASH\->{$_} } ARRAY \& Regexp ARRAY any ARRAY elements pattern match Regexp \& like: grep { /Regexp/ } ARRAY \& undef ARRAY undef in ARRAY \& like: grep { !defined } ARRAY \& Any ARRAY smartmatch each ARRAY element[3] \& like: grep { Any ~~ $_ } ARRAY \& \& Right operand is a HASH: \& \& Left Right Description and pseudocode \& =============================================================== \& HASH1 HASH2 all same keys in both HASHes \& like: keys HASH1 == \& grep { exists HASH2\->{$_} } keys HASH1 \& ARRAY HASH any ARRAY elements exist as HASH keys \& like: grep { exists HASH\->{$_} } ARRAY \& Regexp HASH any HASH keys pattern match Regexp \& like: grep { /Regexp/ } keys HASH \& undef HASH always false (undef cannot be a key) \& like: 0 == 1 \& Any HASH HASH key existence \& like: exists HASH\->{Any} \& \& Right operand is CODE: \& \& Left Right Description and pseudocode \& =============================================================== \& ARRAY CODE sub returns true on all ARRAY elements[1] \& like: !grep { !CODE\->($_) } ARRAY \& HASH CODE sub returns true on all HASH keys[1] \& like: !grep { !CODE\->($_) } keys HASH \& Any CODE sub passed Any returns true \& like: CODE\->(Any) \& \& Right operand is a Regexp: \& \& Left Right Description and pseudocode \& =============================================================== \& ARRAY Regexp any ARRAY elements match Regexp \& like: grep { /Regexp/ } ARRAY \& HASH Regexp any HASH keys match Regexp \& like: grep { /Regexp/ } keys HASH \& Any Regexp pattern match \& like: Any =~ /Regexp/ \& \& Other: \& \& Left Right Description and pseudocode \& =============================================================== \& Object Any invoke ~~ overloading on Object, \& or fall back to... \& \& Any Num numeric equality \& like: Any == Num \& Num nummy[4] numeric equality \& like: Num == nummy \& undef Any check whether undefined \& like: !defined(Any) \& Any Any string equality \& like: Any eq Any .Ve .PP Notes: .IP "1. Empty hashes or arrays match." 4 .IX Item "1. Empty hashes or arrays match." .PD 0 .IP "2. That is, each element smartmatches the element of the same index in the other array.[3]" 4 .IX Item "2. That is, each element smartmatches the element of the same index in the other array.[3]" .IP "3. If a circular reference is found, fall back to referential equality." 4 .IX Item "3. If a circular reference is found, fall back to referential equality." .IP "4. Either an actual number, or a string that looks like one." 4 .IX Item "4. Either an actual number, or a string that looks like one." .PD .PP The smartmatch implicitly dereferences any non-blessed hash or array reference, so the \f(CW\*(C`\fR\f(CIHASH\fR\f(CW\*(C'\fR and \f(CW\*(C`\fR\f(CIARRAY\fR\f(CW\*(C'\fR entries apply in those cases. For blessed references, the \f(CW\*(C`\fR\f(CIObject\fR\f(CW\*(C'\fR entries apply. Smartmatches involving hashes only consider hash keys, never hash values. .PP The "like" code entry is not always an exact rendition. For example, the smartmatch operator short-circuits whenever possible, but \f(CW\*(C`grep\*(C'\fR does not. Also, \f(CW\*(C`grep\*(C'\fR in scalar context returns the number of matches, but \&\f(CW\*(C`~~\*(C'\fR returns only true or false. .PP Unlike most operators, the smartmatch operator knows to treat \f(CW\*(C`undef\*(C'\fR specially: .PP .Vb 3 \& use v5.10.1; \& @array = (1, 2, 3, undef, 4, 5); \& say "some elements undefined" if undef ~~ @array; .Ve .PP Each operand is considered in a modified scalar context, the modification being that array and hash variables are passed by reference to the operator, which implicitly dereferences them. Both elements of each pair are the same: .PP .Vb 1 \& use v5.10.1; \& \& my %hash = (red => 1, blue => 2, green => 3, \& orange => 4, yellow => 5, purple => 6, \& black => 7, grey => 8, white => 9); \& \& my @array = qw(red blue green); \& \& say "some array elements in hash keys" if @array ~~ %hash; \& say "some array elements in hash keys" if \e@array ~~ \e%hash; \& \& say "red in array" if "red" ~~ @array; \& say "red in array" if "red" ~~ \e@array; \& \& say "some keys end in e" if /e$/ ~~ %hash; \& say "some keys end in e" if /e$/ ~~ \e%hash; .Ve .PP Two arrays smartmatch if each element in the first array smartmatches (that is, is "in") the corresponding element in the second array, recursively. .PP .Vb 6 \& use v5.10.1; \& my @little = qw(red blue green); \& my @bigger = ("red", "blue", [ "orange", "green" ] ); \& if (@little ~~ @bigger) { # true! \& say "little is contained in bigger"; \& } .Ve .PP Because the smartmatch operator recurses on nested arrays, this will still report that "red" is in the array. .PP .Vb 4 \& use v5.10.1; \& my @array = qw(red blue green); \& my $nested_array = [[[[[[[ @array ]]]]]]]; \& say "red in array" if "red" ~~ $nested_array; .Ve .PP If two arrays smartmatch each other, then they are deep copies of each others' values, as this example reports: .PP .Vb 3 \& use v5.12.0; \& my @a = (0, 1, 2, [3, [4, 5], 6], 7); \& my @b = (0, 1, 2, [3, [4, 5], 6], 7); \& \& if (@a ~~ @b && @b ~~ @a) { \& say "a and b are deep copies of each other"; \& } \& elsif (@a ~~ @b) { \& say "a smartmatches in b"; \& } \& elsif (@b ~~ @a) { \& say "b smartmatches in a"; \& } \& else { \& say "a and b don\*(Aqt smartmatch each other at all"; \& } .Ve .PP If you were to set \f(CW\*(C`$b[3]\ =\ 4\*(C'\fR, then instead of reporting that "a and b are deep copies of each other", it now reports that \f(CW"b smartmatches in a"\fR. That's because the corresponding position in \f(CW@a\fR contains an array that (eventually) has a 4 in it. .PP Smartmatching one hash against another reports whether both contain the same keys, no more and no less. This could be used to see whether two records have the same field names, without caring what values those fields might have. For example: .PP .Vb 3 \& use v5.10.1; \& sub make_dogtag { \& state $REQUIRED_FIELDS = { name=>1, rank=>1, serial_num=>1 }; \& \& my ($class, $init_fields) = @_; \& \& die "Must supply (only) name, rank, and serial number" \& unless $init_fields ~~ $REQUIRED_FIELDS; \& \& ... \& } .Ve .PP However, this only does what you mean if \f(CW$init_fields\fR is indeed a hash reference. The condition \f(CW\*(C`$init_fields ~~ $REQUIRED_FIELDS\*(C'\fR also allows the strings \f(CW"name"\fR, \f(CW"rank"\fR, \f(CW"serial_num"\fR as well as any array reference that contains \f(CW"name"\fR or \f(CW"rank"\fR or \f(CW"serial_num"\fR anywhere to pass through. .PP The smartmatch operator is most often used as the implicit operator of a \&\f(CW\*(C`when\*(C'\fR clause. See the section on "Switch Statements" in perlsyn. .PP \fISmartmatching of Objects\fR .IX Subsection "Smartmatching of Objects" .PP To avoid relying on an object's underlying representation, if the smartmatch's right operand is an object that doesn't overload \f(CW\*(C`~~\*(C'\fR, it raises the exception "\f(CW\*(C`Smartmatching a non\-overloaded object breaks encapsulation\*(C'\fR". That's because one has no business digging around to see whether something is "in" an object. These are all illegal on objects without a \f(CW\*(C`~~\*(C'\fR overload: .PP .Vb 3 \& %hash ~~ $object \& 42 ~~ $object \& "fred" ~~ $object .Ve .PP However, you can change the way an object is smartmatched by overloading the \f(CW\*(C`~~\*(C'\fR operator. This is allowed to extend the usual smartmatch semantics. For objects that do have an \f(CW\*(C`~~\*(C'\fR overload, see overload. .PP Using an object as the left operand is allowed, although not very useful. Smartmatching rules take precedence over overloading, so even if the object in the left operand has smartmatch overloading, this will be ignored. A left operand that is a non-overloaded object falls back on a string or numeric comparison of whatever the \f(CW\*(C`ref\*(C'\fR operator returns. That means that .PP .Vb 1 \& $object ~~ X .Ve .PP does \fInot\fR invoke the overload method with \f(CW\*(C`\fR\f(CIX\fR\f(CW\*(C'\fR as an argument. Instead the above table is consulted as normal, and based on the type of \&\f(CW\*(C`\fR\f(CIX\fR\f(CW\*(C'\fR, overloading may or may not be invoked. For simple strings or numbers, "in" becomes equivalent to this: .PP .Vb 2 \& $object ~~ $number ref($object) == $number \& $object ~~ $string ref($object) eq $string .Ve .PP For example, this reports that the handle smells IOish (but please don't really do this!): .PP .Vb 5 \& use IO::Handle; \& my $fh = IO::Handle\->new(); \& if ($fh ~~ /\ebIO\eb/) { \& say "handle smells IOish"; \& } .Ve .PP That's because it treats \f(CW$fh\fR as a string like \&\f(CW"IO::Handle=GLOB(0x8039e0)"\fR, then pattern matches against that. .SS "Bitwise And" .IX Xref "operator, bitwise, and bitwise and &" .IX Subsection "Bitwise And" Binary \f(CW"&"\fR returns its operands ANDed together bit by bit. Although no warning is currently raised, the result is not well defined when this operation is performed on operands that aren't either numbers (see "Integer Arithmetic") nor bitstrings (see "Bitwise String Operators"). .PP Note that \f(CW"&"\fR has lower priority than relational operators, so for example the parentheses are essential in a test like .PP .Vb 1 \& print "Even\en" if ($x & 1) == 0; .Ve .PP If the "bitwise" feature is enabled via \f(CW\*(C`use\ feature\ \*(Aqbitwise\*(Aq\*(C'\fR or \&\f(CW\*(C`use v5.28\*(C'\fR, then this operator always treats its operands as numbers. Before Perl 5.28 this feature produced a warning in the \&\f(CW"experimental::bitwise"\fR category. .SS "Bitwise Or and Exclusive Or" .IX Xref "operator, bitwise, or bitwise or | operator, bitwise, xor bitwise xor ^" .IX Subsection "Bitwise Or and Exclusive Or" Binary \f(CW"|"\fR returns its operands ORed together bit by bit. .PP Binary \f(CW"^"\fR returns its operands XORed together bit by bit. .PP Although no warning is currently raised, the results are not well defined when these operations are performed on operands that aren't either numbers (see "Integer Arithmetic") nor bitstrings (see "Bitwise String Operators"). .PP Note that \f(CW"|"\fR and \f(CW"^"\fR have lower priority than relational operators, so for example the parentheses are essential in a test like .PP .Vb 1 \& print "false\en" if (8 | 2) != 10; .Ve .PP If the "bitwise" feature is enabled via \f(CW\*(C`use\ feature\ \*(Aqbitwise\*(Aq\*(C'\fR or \&\f(CW\*(C`use v5.28\*(C'\fR, then this operator always treats its operands as numbers. Before Perl 5.28. this feature produced a warning in the \&\f(CW"experimental::bitwise"\fR category. .SS "C\-style Logical And" .IX Xref "&& logical and operator, logical, and" .IX Subsection "C-style Logical And" Binary \f(CW"&&"\fR performs a short-circuit logical AND operation. That is, if the left operand is false, the right operand is not even evaluated. Scalar or list context propagates down to the right operand if it is evaluated. .SS "C\-style Logical Or" .IX Xref "|| operator, logical, or" .IX Subsection "C-style Logical Or" Binary \f(CW"||"\fR performs a short-circuit logical OR operation. That is, if the left operand is true, the right operand is not even evaluated. Scalar or list context propagates down to the right operand if it is evaluated. .SS "Logical Defined-Or" .IX Xref "operator, logical, defined-or" .IX Subsection "Logical Defined-Or" Although it has no direct equivalent in C, Perl's \f(CW\*(C`//\*(C'\fR operator is related to its C\-style "or". In fact, it's exactly the same as \f(CW\*(C`||\*(C'\fR, except that it tests the left hand side's definedness instead of its truth. Thus, \&\f(CW\*(C`EXPR1\ //\ EXPR2\*(C'\fR returns the value of \f(CW\*(C`EXPR1\*(C'\fR if it's defined, otherwise, the value of \f(CW\*(C`EXPR2\*(C'\fR is returned. (\f(CW\*(C`EXPR1\*(C'\fR is evaluated in scalar context, \f(CW\*(C`EXPR2\*(C'\fR in the context of \f(CW\*(C`//\*(C'\fR itself). Usually, this is the same result as \f(CW\*(C`defined(EXPR1)\ ?\ EXPR1\ :\ EXPR2\*(C'\fR (except that the ternary-operator form can be used as a lvalue, while \f(CW\*(C`EXPR1\ //\ EXPR2\*(C'\fR cannot). This is very useful for providing default values for variables. If you actually want to test if at least one of \f(CW$x\fR and \f(CW$y\fR is defined, use \f(CW\*(C`defined($x\ //\ $y)\*(C'\fR. .PP The \f(CW\*(C`||\*(C'\fR, \f(CW\*(C`//\*(C'\fR and \f(CW\*(C`&&\*(C'\fR operators return the last value evaluated (unlike C's \f(CW\*(C`||\*(C'\fR and \f(CW\*(C`&&\*(C'\fR, which return 0 or 1). Thus, a reasonably portable way to find out the home directory might be: .PP .Vb 4 \& $home = $ENV{HOME} \& // $ENV{LOGDIR} \& // (getpwuid($<))[7] \& // die "You\*(Aqre homeless!\en"; .Ve .PP In particular, this means that you shouldn't use this for selecting between two aggregates for assignment: .PP .Vb 3 \& @a = @b || @c; # This doesn\*(Aqt do the right thing \& @a = scalar(@b) || @c; # because it really means this. \& @a = @b ? @b : @c; # This works fine, though. .Ve .PP As alternatives to \f(CW\*(C`&&\*(C'\fR and \f(CW\*(C`||\*(C'\fR when used for control flow, Perl provides the \f(CW\*(C`and\*(C'\fR and \f(CW\*(C`or\*(C'\fR operators (see below). The short-circuit behavior is identical. The precedence of \f(CW"and"\fR and \f(CW"or"\fR is much lower, however, so that you can safely use them after a list operator without the need for parentheses: .PP .Vb 2 \& unlink "alpha", "beta", "gamma" \& or gripe(), next LINE; .Ve .PP With the C\-style operators that would have been written like this: .PP .Vb 2 \& unlink("alpha", "beta", "gamma") \& || (gripe(), next LINE); .Ve .PP It would be even more readable to write that this way: .PP .Vb 4 \& unless(unlink("alpha", "beta", "gamma")) { \& gripe(); \& next LINE; \& } .Ve .PP Using \f(CW"or"\fR for assignment is unlikely to do what you want; see below. .SS "Range Operators" .IX Xref "operator, range range .. ..." .IX Subsection "Range Operators" Binary \f(CW".."\fR is the range operator, which is really two different operators depending on the context. In list context, it returns a list of values counting (up by ones) from the left value to the right value. If the left value is greater than the right value then it returns the empty list. The range operator is useful for writing \&\f(CW\*(C`foreach\ (1..10)\*(C'\fR loops and for doing slice operations on arrays. In the current implementation, no temporary array is created when the range operator is used as the expression in \f(CW\*(C`foreach\*(C'\fR loops, but older versions of Perl might burn a lot of memory when you write something like this: .PP .Vb 3 \& for (1 .. 1_000_000) { \& # code \& } .Ve .PP The range operator also works on strings, using the magical auto-increment, see below. .PP In scalar context, \f(CW".."\fR returns a boolean value. The operator is bistable, like a flip-flop, and emulates the line-range (comma) operator of \fBsed\fR, \fBawk\fR, and various editors. Each \f(CW".."\fR operator maintains its own boolean state, even across calls to a subroutine that contains it. It is false as long as its left operand is false. Once the left operand is true, the range operator stays true until the right operand is true, \fIAFTER\fR which the range operator becomes false again. It doesn't become false till the next time the range operator is evaluated. It can test the right operand and become false on the same evaluation it became true (as in \fBawk\fR), but it still returns true once. If you don't want it to test the right operand until the next evaluation, as in \fBsed\fR, just use three dots (\f(CW"..."\fR) instead of two. In all other regards, \f(CW"..."\fR behaves just like \f(CW".."\fR does. .PP The right operand is not evaluated while the operator is in the "false" state, and the left operand is not evaluated while the operator is in the "true" state. The precedence is a little lower than || and &&. The value returned is either the empty string for false, or a sequence number (beginning with 1) for true. The sequence number is reset for each range encountered. The final sequence number in a range has the string \f(CW"E0"\fR appended to it, which doesn't affect its numeric value, but gives you something to search for if you want to exclude the endpoint. You can exclude the beginning point by waiting for the sequence number to be greater than 1. .PP If either operand of scalar \f(CW".."\fR is a constant expression, that operand is considered true if it is equal (\f(CW\*(C`==\*(C'\fR) to the current input line number (the \f(CW$.\fR variable). .PP To be pedantic, the comparison is actually \f(CW\*(C`int(EXPR)\ ==\ int(EXPR)\*(C'\fR, but that is only an issue if you use a floating point expression; when implicitly using \f(CW$.\fR as described in the previous paragraph, the comparison is \f(CW\*(C`int(EXPR)\ ==\ int($.)\*(C'\fR which is only an issue when \f(CW$.\fR is set to a floating point value and you are not reading from a file. Furthermore, \f(CW"span"\ ..\ "spat"\fR or \f(CW\*(C`2.18\ ..\ 3.14\*(C'\fR will not do what you want in scalar context because each of the operands are evaluated using their integer representation. .PP Examples: .PP As a scalar operator: .PP .Vb 2 \& if (101 .. 200) { print; } # print 2nd hundred lines, short for \& # if ($. == 101 .. $. == 200) { print; } \& \& next LINE if (1 .. /^$/); # skip header lines, short for \& # next LINE if ($. == 1 .. /^$/); \& # (typically in a loop labeled LINE) \& \& s/^/> / if (/^$/ .. eof()); # quote body \& \& # parse mail messages \& while (<>) { \& $in_header = 1 .. /^$/; \& $in_body = /^$/ .. eof; \& if ($in_header) { \& # do something \& } else { # in body \& # do something else \& } \& } continue { \& close ARGV if eof; # reset $. each file \& } .Ve .PP Here's a simple example to illustrate the difference between the two range operators: .PP .Vb 4 \& @lines = (" \- Foo", \& "01 \- Bar", \& "1 \- Baz", \& " \- Quux"); \& \& foreach (@lines) { \& if (/0/ .. /1/) { \& print "$_\en"; \& } \& } .Ve .PP This program will print only the line containing "Bar". If the range operator is changed to \f(CW\*(C`...\*(C'\fR, it will also print the "Baz" line. .PP And now some examples as a list operator: .PP .Vb 3 \& for (101 .. 200) { print } # print $_ 100 times \& @foo = @foo[0 .. $#foo]; # an expensive no\-op \& @foo = @foo[$#foo\-4 .. $#foo]; # slice last 5 items .Ve .PP Because each operand is evaluated in integer form, \f(CW\*(C`2.18\ ..\ 3.14\*(C'\fR will return two elements in list context. .PP .Vb 1 \& @list = (2.18 .. 3.14); # same as @list = (2 .. 3); .Ve .PP The range operator in list context can make use of the magical auto-increment algorithm if both operands are strings, subject to the following rules: .IP \(bu 4 With one exception (below), if both strings look like numbers to Perl, the magic increment will not be applied, and the strings will be treated as numbers (more specifically, integers) instead. .Sp For example, \f(CW"\-2".."2"\fR is the same as \f(CW\-2..2\fR, and \&\f(CW"2.18".."3.14"\fR produces \f(CW\*(C`2, 3\*(C'\fR. .IP \(bu 4 The exception to the above rule is when the left-hand string begins with \&\f(CW0\fR and is longer than one character, in this case the magic increment \&\fIwill\fR be applied, even though strings like \f(CW"01"\fR would normally look like a number to Perl. .Sp For example, \f(CW"01".."04"\fR produces \f(CW"01", "02", "03", "04"\fR, and \&\f(CW"00".."\-1"\fR produces \f(CW"00"\fR through \f(CW"99"\fR \- this may seem surprising, but see the following rules for why it works this way. To get dates with leading zeros, you can say: .Sp .Vb 2 \& @z2 = ("01" .. "31"); \& print $z2[$mday]; .Ve .Sp If you want to force strings to be interpreted as numbers, you could say .Sp .Vb 1 \& @numbers = ( 0+$first .. 0+$last ); .Ve .Sp \&\fBNote:\fR In Perl versions 5.30 and below, \fIany\fR string on the left-hand side beginning with \f(CW"0"\fR, including the string \f(CW"0"\fR itself, would cause the magic string increment behavior. This means that on these Perl versions, \f(CW"0".."\-1"\fR would produce \f(CW"0"\fR through \f(CW"99"\fR, which was inconsistent with \f(CW\*(C`0..\-1\*(C'\fR, which produces the empty list. This also means that \f(CW"0".."9"\fR now produces a list of integers instead of a list of strings. .IP \(bu 4 If the initial value specified isn't part of a magical increment sequence (that is, a non-empty string matching \f(CW\*(C`/^[a\-zA\-Z]*[0\-9]*\ez/\*(C'\fR), only the initial value will be returned. .Sp For example, \f(CW"ax".."az"\fR produces \f(CW"ax", "ay", "az"\fR, but \&\f(CW"*x".."az"\fR produces only \f(CW"*x"\fR. .IP \(bu 4 For other initial values that are strings that do follow the rules of the magical increment, the corresponding sequence will be returned. .Sp For example, you can say .Sp .Vb 1 \& @alphabet = ("A" .. "Z"); .Ve .Sp to get all normal letters of the English alphabet, or .Sp .Vb 1 \& $hexdigit = (0 .. 9, "a" .. "f")[$num & 15]; .Ve .Sp to get a hexadecimal digit. .IP \(bu 4 If the final value specified is not in the sequence that the magical increment would produce, the sequence goes until the next value would be longer than the final value specified. If the length of the final string is shorter than the first, the empty list is returned. .Sp For example, \f(CW"a".."\-\-"\fR is the same as \f(CW"a".."zz"\fR, \f(CW"0".."xx"\fR produces \f(CW"0"\fR through \f(CW"99"\fR, and \f(CW"aaa".."\-\-"\fR returns the empty list. .PP As of Perl 5.26, the list-context range operator on strings works as expected in the scope of \f(CW"use\ feature\ \*(Aqunicode_strings"\fR. In previous versions, and outside the scope of that feature, it exhibits "The "Unicode Bug"" in perlunicode: its behavior depends on the internal encoding of the range endpoint. .PP Because the magical increment only works on non-empty strings matching \&\f(CW\*(C`/^[a\-zA\-Z]*[0\-9]*\ez/\*(C'\fR, the following will only return an alpha: .PP .Vb 2 \& use charnames "greek"; \& my @greek_small = ("\eN{alpha}" .. "\eN{omega}"); .Ve .PP To get the 25 traditional lowercase Greek letters, including both sigmas, you could use this instead: .PP .Vb 5 \& use charnames "greek"; \& my @greek_small = map { chr } ( ord("\eN{alpha}") \& .. \& ord("\eN{omega}") \& ); .Ve .PP However, because there are \fImany\fR other lowercase Greek characters than just those, to match lowercase Greek characters in a regular expression, you could use the pattern \f(CW\*(C`/(?:(?=\ep{Greek})\ep{Lower})+/\*(C'\fR (or the experimental feature \f(CW\*(C`/(?[\ \ep{Greek}\ &\ \ep{Lower}\ ])+/\*(C'\fR). .SS "Conditional Operator" .IX Xref "operator, conditional operator, ternary ternary ?:" .IX Subsection "Conditional Operator" Ternary \f(CW"?:"\fR is the conditional operator, just as in C. It works much like an if-then-else. If the argument before the \f(CW\*(C`?\*(C'\fR is true, the argument before the \f(CW\*(C`:\*(C'\fR is returned, otherwise the argument after the \&\f(CW\*(C`:\*(C'\fR is returned. For example: .PP .Vb 2 \& printf "I have %d dog%s.\en", $n, \& ($n == 1) ? "" : "s"; .Ve .PP Scalar or list context propagates downward into the 2nd or 3rd argument, whichever is selected. .PP .Vb 3 \& $x = $ok ? $y : $z; # get a scalar \& @x = $ok ? @y : @z; # get an array \& $x = $ok ? @y : @z; # oops, that\*(Aqs just a count! .Ve .PP The operator may be assigned to if both the 2nd and 3rd arguments are legal lvalues (meaning that you can assign to them): .PP .Vb 1 \& ($x_or_y ? $x : $y) = $z; .Ve .PP Because this operator produces an assignable result, using assignments without parentheses will get you in trouble. For example, this: .PP .Vb 1 \& $x % 2 ? $x += 10 : $x += 2 .Ve .PP Really means this: .PP .Vb 1 \& (($x % 2) ? ($x += 10) : $x) += 2 .Ve .PP Rather than this: .PP .Vb 1 \& ($x % 2) ? ($x += 10) : ($x += 2) .Ve .PP That should probably be written more simply as: .PP .Vb 1 \& $x += ($x % 2) ? 10 : 2; .Ve .SS "Assignment Operators" .IX Xref "assignment operator, assignment = **= += *= &= <<= &&= -= = |= >>= ||= = .= %= ^= x= &.= |.= ^.=" .IX Subsection "Assignment Operators" \&\f(CW"="\fR is the ordinary assignment operator. .PP Assignment operators work as in C. That is, .PP .Vb 1 \& $x += 2; .Ve .PP is equivalent to .PP .Vb 1 \& $x = $x + 2; .Ve .PP although without duplicating any side effects that dereferencing the lvalue might trigger, such as from \f(CWtie()\fR. Other assignment operators work similarly. The following are recognized: .PP .Vb 4 \& **= += *= &= &.= <<= &&= \& \-= /= |= |.= >>= ||= \& .= %= ^= ^.= //= \& x= .Ve .PP Although these are grouped by family, they all have the precedence of assignment. These combined assignment operators can only operate on scalars, whereas the ordinary assignment operator can assign to arrays, hashes, lists and even references. (See "Context" and "List value constructors" in perldata, and "Assigning to References" in perlref.) .PP Unlike in C, the scalar assignment operator produces a valid lvalue. Modifying an assignment is equivalent to doing the assignment and then modifying the variable that was assigned to. This is useful for modifying a copy of something, like this: .PP .Vb 1 \& ($tmp = $global) =~ tr/13579/24680/; .Ve .PP Although as of 5.14, that can be also be accomplished this way: .PP .Vb 2 \& use v5.14; \& $tmp = ($global =~ tr/13579/24680/r); .Ve .PP Likewise, .PP .Vb 1 \& ($x += 2) *= 3; .Ve .PP is equivalent to .PP .Vb 2 \& $x += 2; \& $x *= 3; .Ve .PP Similarly, a list assignment in list context produces the list of lvalues assigned to, and a list assignment in scalar context returns the number of elements produced by the expression on the right hand side of the assignment. .PP The three dotted bitwise assignment operators (\f(CW\*(C`&.=\*(C'\fR \f(CW\*(C`|.=\*(C'\fR \f(CW\*(C`^.=\*(C'\fR) are new in Perl 5.22. See "Bitwise String Operators". .SS "Comma Operator" .IX Xref "comma operator, comma ," .IX Subsection "Comma Operator" Binary \f(CW","\fR is the comma operator. In scalar context it evaluates its left argument, throws that value away, then evaluates its right argument and returns that value. This is just like C's comma operator. .PP In list context, it's just the list argument separator, and inserts both its arguments into the list. These arguments are also evaluated from left to right. .PP The \f(CW\*(C`=>\*(C'\fR operator (sometimes pronounced "fat comma") is a synonym for the comma except that it causes a word on its left to be interpreted as a string if it begins with a letter or underscore and is composed only of letters, digits and underscores. This includes operands that might otherwise be interpreted as operators, constants, single number v\-strings or function calls. If in doubt about this behavior, the left operand can be quoted explicitly. .PP Otherwise, the \f(CW\*(C`=>\*(C'\fR operator behaves exactly as the comma operator or list argument separator, according to context. .PP For example: .PP .Vb 1 \& use constant FOO => "something"; \& \& my %h = ( FOO => 23 ); .Ve .PP is equivalent to: .PP .Vb 1 \& my %h = ("FOO", 23); .Ve .PP It is \fINOT\fR: .PP .Vb 1 \& my %h = ("something", 23); .Ve .PP The \f(CW\*(C`=>\*(C'\fR operator is helpful in documenting the correspondence between keys and values in hashes, and other paired elements in lists. .PP .Vb 2 \& %hash = ( $key => $value ); \& login( $username => $password ); .Ve .PP The special quoting behavior ignores precedence, and hence may apply to \&\fIpart\fR of the left operand: .PP .Vb 1 \& print time.shift => "bbb"; .Ve .PP That example prints something like \f(CW"1314363215shiftbbb"\fR, because the \&\f(CW\*(C`=>\*(C'\fR implicitly quotes the \f(CW\*(C`shift\*(C'\fR immediately on its left, ignoring the fact that \f(CW\*(C`time.shift\*(C'\fR is the entire left operand. .SS "List Operators (Rightward)" .IX Xref "operator, list, rightward list operator" .IX Subsection "List Operators (Rightward)" On the right side of a list operator, the comma has very low precedence, such that it controls all comma-separated expressions found there. The only operators with lower precedence are the logical operators \&\f(CW"and"\fR, \f(CW"or"\fR, and \f(CW"not"\fR, which may be used to evaluate calls to list operators without the need for parentheses: .PP .Vb 2 \& open HANDLE, "< :encoding(UTF\-8)", "filename" \& or die "Can\*(Aqt open: $!\en"; .Ve .PP However, some people find that code harder to read than writing it with parentheses: .PP .Vb 2 \& open(HANDLE, "< :encoding(UTF\-8)", "filename") \& or die "Can\*(Aqt open: $!\en"; .Ve .PP in which case you might as well just use the more customary \f(CW"||"\fR operator: .PP .Vb 2 \& open(HANDLE, "< :encoding(UTF\-8)", "filename") \& || die "Can\*(Aqt open: $!\en"; .Ve .PP See also discussion of list operators in "Terms and List Operators (Leftward)". .SS "Logical Not" .IX Xref "operator, logical, not not" .IX Subsection "Logical Not" Unary \f(CW"not"\fR returns the logical negation of the expression to its right. It's the equivalent of \f(CW"!"\fR except for the very low precedence. .SS "Logical And" .IX Xref "operator, logical, and and" .IX Subsection "Logical And" Binary \f(CW"and"\fR returns the logical conjunction of the two surrounding expressions. It's equivalent to \f(CW\*(C`&&\*(C'\fR except for the very low precedence. This means that it short-circuits: the right expression is evaluated only if the left expression is true. .SS "Logical or and Exclusive Or" .IX Xref "operator, logical, or operator, logical, xor operator, logical, exclusive or or xor" .IX Subsection "Logical or and Exclusive Or" Binary \f(CW"or"\fR returns the logical disjunction of the two surrounding expressions. It's equivalent to \f(CW\*(C`||\*(C'\fR except for the very low precedence. This makes it useful for control flow: .PP .Vb 1 \& print FH $data or die "Can\*(Aqt write to FH: $!"; .Ve .PP This means that it short-circuits: the right expression is evaluated only if the left expression is false. Due to its precedence, you must be careful to avoid using it as replacement for the \f(CW\*(C`||\*(C'\fR operator. It usually works out better for flow control than in assignments: .PP .Vb 3 \& $x = $y or $z; # bug: this is wrong \& ($x = $y) or $z; # really means this \& $x = $y || $z; # better written this way .Ve .PP However, when it's a list-context assignment and you're trying to use \&\f(CW\*(C`||\*(C'\fR for control flow, you probably need \f(CW"or"\fR so that the assignment takes higher precedence. .PP .Vb 2 \& @info = stat($file) || die; # oops, scalar sense of stat! \& @info = stat($file) or die; # better, now @info gets its due .Ve .PP Then again, you could always use parentheses. .PP Binary \f(CW"xor"\fR returns the exclusive-OR of the two surrounding expressions. It cannot short-circuit (of course). .PP There is no low precedence operator for defined-OR. .SS "C Operators Missing From Perl" .IX Xref "operator, missing from perl & * typecasting (TYPE)" .IX Subsection "C Operators Missing From Perl" Here is what C has that Perl doesn't: .IP "unary &" 8 .IX Item "unary &" Address-of operator. (But see the \f(CW"\e"\fR operator for taking a reference.) .IP "unary *" 8 .IX Item "unary *" Dereference-address operator. (Perl's prefix dereferencing operators are typed: \f(CW\*(C`$\*(C'\fR, \f(CW\*(C`@\*(C'\fR, \f(CW\*(C`%\*(C'\fR, and \f(CW\*(C`&\*(C'\fR.) .IP (TYPE) 8 .IX Item "(TYPE)" Type-casting operator. .SS "Quote and Quote-like Operators" .IX Xref "operator, quote operator, quote-like q qq qx qw m qr s tr ' '' "" """" ` `` << escape sequence escape" .IX Subsection "Quote and Quote-like Operators" While we usually think of quotes as literal values, in Perl they function as operators, providing various kinds of interpolating and pattern matching capabilities. Perl provides customary quote characters for these behaviors, but also provides a way for you to choose your quote character for any of them. In the following table, a \f(CW\*(C`{}\*(C'\fR represents any pair of delimiters you choose. .PP .Vb 11 \& Customary Generic Meaning Interpolates \& \*(Aq\*(Aq q{} Literal no \& "" qq{} Literal yes \& \`\` qx{} Command yes* \& qw{} Word list no \& // m{} Pattern match yes* \& qr{} Pattern yes* \& s{}{} Substitution yes* \& tr{}{} Transliteration no (but see below) \& y{}{} Transliteration no (but see below) \& <{key}[0]\*(C'\fR are also interpolated, as are array and hash slices. But method calls such as \f(CW\*(C`$obj\->meth\*(C'\fR are not. .PP Interpolating an array or slice interpolates the elements in order, separated by the value of \f(CW$"\fR, so is equivalent to interpolating \&\f(CW\*(C`join\ $",\ @array\*(C'\fR. "Punctuation" arrays such as \f(CW\*(C`@*\*(C'\fR are usually interpolated only if the name is enclosed in braces \f(CW\*(C`@{*}\*(C'\fR, but the arrays \f(CW@_\fR, \f(CW\*(C`@+\*(C'\fR, and \f(CW\*(C`@\-\*(C'\fR are interpolated even without braces. .PP For double-quoted strings, the quoting from \f(CW\*(C`\eQ\*(C'\fR is applied after interpolation and escapes are processed. .PP .Vb 1 \& "abc\eQfoo\etbar$s\eExyz" .Ve .PP is equivalent to .PP .Vb 1 \& "abc" . quotemeta("foo\etbar$s") . "xyz" .Ve .PP For the pattern of regex operators (\f(CW\*(C`qr//\*(C'\fR, \f(CW\*(C`m//\*(C'\fR and \f(CW\*(C`s///\*(C'\fR), the quoting from \f(CW\*(C`\eQ\*(C'\fR is applied after interpolation is processed, but before escapes are processed. This allows the pattern to match literally (except for \f(CW\*(C`$\*(C'\fR and \f(CW\*(C`@\*(C'\fR). For example, the following matches: .PP .Vb 1 \& \*(Aq\es\et\*(Aq =~ /\eQ\es\et/ .Ve .PP Because \f(CW\*(C`$\*(C'\fR or \f(CW\*(C`@\*(C'\fR trigger interpolation, you'll need to use something like \f(CW\*(C`/\eQuser\eE\e@\eQhost/\*(C'\fR to match them literally. .PP Patterns are subject to an additional level of interpretation as a regular expression. This is done as a second pass, after variables are interpolated, so that regular expressions may be incorporated into the pattern from the variables. If this is not what you want, use \f(CW\*(C`\eQ\*(C'\fR to interpolate a variable literally. .PP Apart from the behavior described above, Perl does not expand multiple levels of interpolation. In particular, contrary to the expectations of shell programmers, back-quotes do \fINOT\fR interpolate within double quotes, nor do single quotes impede evaluation of variables when used within double quotes. .SS "Regexp Quote-Like Operators" .IX Xref "operator, regexp" .IX Subsection "Regexp Quote-Like Operators" Here are the quote-like operators that apply to pattern matching and related activities. .ie n .IP """qr/\fISTRING\fR/msixpodualn""" 8 .el .IP \f(CWqr/\fR\f(CISTRING\fR\f(CW/msixpodualn\fR 8 .IX Xref "qr i m o s x p" .IX Item "qr/STRING/msixpodualn" This operator quotes (and possibly compiles) its \fISTRING\fR as a regular expression. \fISTRING\fR is interpolated the same way as \fIPATTERN\fR in \f(CW\*(C`m/\fR\f(CIPATTERN\fR\f(CW/\*(C'\fR. If \f(CW"\*(Aq"\fR is used as the delimiter, no variable interpolation is done. Returns a Perl value which may be used instead of the corresponding \f(CW\*(C`/\fR\f(CISTRING\fR\f(CW/msixpodualn\*(C'\fR expression. The returned value is a normalized version of the original pattern. It magically differs from a string containing the same characters: \f(CWref(qr/x/)\fR returns "Regexp"; however, dereferencing it is not well defined (you currently get the normalized version of the original pattern, but this may change). .Sp For example, .Sp .Vb 3 \& $rex = qr/my.STRING/is; \& print $rex; # prints (?si\-xm:my.STRING) \& s/$rex/foo/; .Ve .Sp is equivalent to .Sp .Vb 1 \& s/my.STRING/foo/is; .Ve .Sp The result may be used as a subpattern in a match: .Sp .Vb 5 \& $re = qr/$pattern/; \& $string =~ /foo${re}bar/; # can be interpolated in other \& # patterns \& $string =~ $re; # or used standalone \& $string =~ /$re/; # or this way .Ve .Sp Since Perl may compile the pattern at the moment of execution of the \f(CWqr()\fR operator, using \f(CWqr()\fR may have speed advantages in some situations, notably if the result of \f(CWqr()\fR is used standalone: .Sp .Vb 11 \& sub match { \& my $patterns = shift; \& my @compiled = map qr/$_/i, @$patterns; \& grep { \& my $success = 0; \& foreach my $pat (@compiled) { \& $success = 1, last if /$pat/; \& } \& $success; \& } @_; \& } .Ve .Sp Precompilation of the pattern into an internal representation at the moment of \f(CWqr()\fR avoids the need to recompile the pattern every time a match \f(CW\*(C`/$pat/\*(C'\fR is attempted. (Perl has many other internal optimizations, but none would be triggered in the above example if we did not use \f(CWqr()\fR operator.) .Sp Options (specified by the following modifiers) are: .Sp .Vb 10 \& m Treat string as multiple lines. \& s Treat string as single line. (Make . match a newline) \& i Do case\-insensitive pattern matching. \& x Use extended regular expressions; specifying two \& x\*(Aqs means \et and the SPACE character are ignored within \& square\-bracketed character classes \& p When matching preserve a copy of the matched string so \& that ${^PREMATCH}, ${^MATCH}, ${^POSTMATCH} will be \& defined (ignored starting in v5.20 as these are always \& defined starting in that release) \& o Compile pattern only once. \& a ASCII\-restrict: Use ASCII for \ed, \es, \ew and [[:posix:]] \& character classes; specifying two a\*(Aqs adds the further \& restriction that no ASCII character will match a \& non\-ASCII one under /i. \& l Use the current run\-time locale\*(Aqs rules. \& u Use Unicode rules. \& d Use Unicode or native charset, as in 5.12 and earlier. \& n Non\-capture mode. Don\*(Aqt let () fill in $1, $2, etc... .Ve .Sp If a precompiled pattern is embedded in a larger pattern then the effect of \f(CW"msixpluadn"\fR will be propagated appropriately. The effect that the \&\f(CW\*(C`/o\*(C'\fR modifier has is not propagated, being restricted to those patterns explicitly using it. .Sp The \f(CW\*(C`/a\*(C'\fR, \f(CW\*(C`/d\*(C'\fR, \f(CW\*(C`/l\*(C'\fR, and \f(CW\*(C`/u\*(C'\fR modifiers (added in Perl 5.14) control the character set rules, but \f(CW\*(C`/a\*(C'\fR is the only one you are likely to want to specify explicitly; the other three are selected automatically by various pragmas. .Sp See perlre for additional information on valid syntax for \fISTRING\fR, and for a detailed look at the semantics of regular expressions. In particular, all modifiers except the largely obsolete \f(CW\*(C`/o\*(C'\fR are further explained in "Modifiers" in perlre. \f(CW\*(C`/o\*(C'\fR is described in the next section. .ie n .IP """m/\fIPATTERN\fR/msixpodualngc""" 8 .el .IP \f(CWm/\fR\f(CIPATTERN\fR\f(CW/msixpodualngc\fR 8 .IX Xref "m operator, match regexp, options regexp regex, options regex m s i x p o g c" .IX Item "m/PATTERN/msixpodualngc" .PD 0 .ie n .IP """/\fIPATTERN\fR/msixpodualngc""" 8 .el .IP \f(CW/\fR\f(CIPATTERN\fR\f(CW/msixpodualngc\fR 8 .IX Item "/PATTERN/msixpodualngc" .PD Searches a string for a pattern match, and in scalar context returns true if it succeeds, false if it fails. If no string is specified via the \f(CW\*(C`=~\*(C'\fR or \f(CW\*(C`!~\*(C'\fR operator, the \f(CW$_\fR string is searched. (The string specified with \f(CW\*(C`=~\*(C'\fR need not be an lvalue\-\-it may be the result of an expression evaluation, but remember the \f(CW\*(C`=~\*(C'\fR binds rather tightly.) See also perlre. .Sp Options are as described in \f(CW\*(C`qr//\*(C'\fR above; in addition, the following match process modifiers are available: .Sp .Vb 3 \& g Match globally, i.e., find all occurrences. \& c Do not reset search position on a failed match when /g is \& in effect. .Ve .Sp If \f(CW"/"\fR is the delimiter then the initial \f(CW\*(C`m\*(C'\fR is optional. With the \f(CW\*(C`m\*(C'\fR you can use any pair of non-whitespace (ASCII) characters as delimiters. This is particularly useful for matching path names that contain \f(CW"/"\fR, to avoid LTS (leaning toothpick syndrome). If \f(CW"?"\fR is the delimiter, then a match-only-once rule applies, described in \f(CW\*(C`m?\fR\f(CIPATTERN\fR\f(CW?\*(C'\fR below. If \f(CW"\*(Aq"\fR (single quote) is the delimiter, no variable interpolation is performed on the \fIPATTERN\fR. When using a delimiter character valid in an identifier, whitespace is required after the \f(CW\*(C`m\*(C'\fR. .Sp \&\fIPATTERN\fR may contain variables, which will be interpolated every time the pattern search is evaluated, except for when the delimiter is a single quote. (Note that \f(CW$(\fR, \f(CW$)\fR, and \&\f(CW$|\fR are not interpolated because they look like end-of-string tests.) Perl will not recompile the pattern unless an interpolated variable that it contains changes. You can force Perl to skip the test and never recompile by adding a \f(CW\*(C`/o\*(C'\fR (which stands for "once") after the trailing delimiter. Once upon a time, Perl would recompile regular expressions unnecessarily, and this modifier was useful to tell it not to do so, in the interests of speed. But now, the only reasons to use \f(CW\*(C`/o\*(C'\fR are one of: .RS 8 .IP 1. 4 The variables are thousands of characters long and you know that they don't change, and you need to wring out the last little bit of speed by having Perl skip testing for that. (There is a maintenance penalty for doing this, as mentioning \f(CW\*(C`/o\*(C'\fR constitutes a promise that you won't change the variables in the pattern. If you do change them, Perl won't even notice.) .IP 2. 4 you want the pattern to use the initial values of the variables regardless of whether they change or not. (But there are saner ways of accomplishing this than using \f(CW\*(C`/o\*(C'\fR.) .IP 3. 4 If the pattern contains embedded code, such as .Sp .Vb 3 \& use re \*(Aqeval\*(Aq; \& $code = \*(Aqfoo(?{ $x })\*(Aq; \& /$code/ .Ve .Sp then perl will recompile each time, even though the pattern string hasn't changed, to ensure that the current value of \f(CW$x\fR is seen each time. Use \f(CW\*(C`/o\*(C'\fR if you want to avoid this. .RE .RS 8 .Sp The bottom line is that using \f(CW\*(C`/o\*(C'\fR is almost never a good idea. .RE .ie n .IP "The empty pattern ""//""" 8 .el .IP "The empty pattern \f(CW//\fR" 8 .IX Item "The empty pattern //" If the \fIPATTERN\fR evaluates to the empty string, the last \&\fIsuccessfully\fR matched regular expression is used instead. In this case, only the \f(CW\*(C`g\*(C'\fR and \f(CW\*(C`c\*(C'\fR flags on the empty pattern are honored; the other flags are taken from the original pattern. If no match has previously succeeded, this will (silently) act instead as a genuine empty pattern (which will always match). Using a user supplied string as a pattern has the risk that if the string is empty that it triggers the "last successful match" behavior, which can be very confusing. In such cases you are recommended to replace \f(CW\*(C`m/$pattern/\*(C'\fR with \&\f(CW\*(C`m/(?:$pattern)/\*(C'\fR to avoid this behavior. .Sp The last successful pattern may be accessed as a variable via \&\f(CW\*(C`${^LAST_SUCCESSFUL_PATTERN}\*(C'\fR. Matching against it, or the empty pattern should have the same effect, with the exception that when there is no last successful pattern the empty pattern will silently match, whereas using the \f(CW\*(C`${^LAST_SUCCESSFUL_PATTERN}\*(C'\fR variable will produce undefined warnings (if warnings are enabled). You can check \&\f(CWdefined(${^LAST_SUCCESSFUL_PATTERN})\fR to test if there is a "last successful match" in the current scope. .Sp Note that it's possible to confuse Perl into thinking \f(CW\*(C`//\*(C'\fR (the empty regex) is really \f(CW\*(C`//\*(C'\fR (the defined-or operator). Perl is usually pretty good about this, but some pathological cases might trigger this, such as \&\f(CW\*(C`$x///\*(C'\fR (is that \f(CW\*(C`($x)\ /\ (//)\*(C'\fR or \f(CW\*(C`$x\ //\ /\*(C'\fR?) and \f(CW\*(C`print\ $fh\ //\*(C'\fR (\f(CW\*(C`print\ $fh(//\*(C'\fR or \f(CW\*(C`print($fh\ //\*(C'\fR?). In all of these examples, Perl will assume you meant defined-or. If you meant the empty regex, just use parentheses or spaces to disambiguate, or even prefix the empty regex with an \f(CW\*(C`m\*(C'\fR (so \f(CW\*(C`//\*(C'\fR becomes \f(CW\*(C`m//\*(C'\fR). .IP "Matching in list context" 8 .IX Item "Matching in list context" If the \f(CW\*(C`/g\*(C'\fR option is not used, \f(CW\*(C`m//\*(C'\fR in list context returns a list consisting of the subexpressions matched by the parentheses in the pattern, that is, (\f(CW$1\fR, \f(CW$2\fR, \f(CW$3\fR...) (Note that here \f(CW$1\fR etc. are also set). When there are no parentheses in the pattern, the return value is the list \f(CW\*(C`(1)\*(C'\fR for success. With or without parentheses, an empty list is returned upon failure. .Sp Examples: .Sp .Vb 2 \& open(TTY, "+ =~ /^y/i && foo(); # do foo if desired \& \& if (/Version: *([0\-9.]*)/) { $version = $1; } \& \& next if m#^/usr/spool/uucp#; \& \& # poor man\*(Aqs grep \& $arg = shift; \& while (<>) { \& print if /$arg/o; # compile only once (no longer needed!) \& } \& \& if (($F1, $F2, $Etc) = ($foo =~ /^(\eS+)\es+(\eS+)\es*(.*)/)) .Ve .Sp This last example splits \f(CW$foo\fR into the first two words and the remainder of the line, and assigns those three fields to \f(CW$F1\fR, \f(CW$F2\fR, and \&\f(CW$Etc\fR. The conditional is true if any variables were assigned; that is, if the pattern matched. .Sp The \f(CW\*(C`/g\*(C'\fR modifier specifies global pattern matching\-\-that is, matching as many times as possible within the string. How it behaves depends on the context. In list context, it returns a list of the substrings matched by any capturing parentheses in the regular expression. If there are no parentheses, it returns a list of all the matched strings, as if there were parentheses around the whole pattern. .Sp In scalar context, each execution of \f(CW\*(C`m//g\*(C'\fR finds the next match, returning true if it matches, and false if there is no further match. The position after the last match can be read or set using the \f(CWpos()\fR function; see "pos" in perlfunc. A failed match normally resets the search position to the beginning of the string, but you can avoid that by adding the \f(CW\*(C`/c\*(C'\fR modifier (for example, \f(CW\*(C`m//gc\*(C'\fR). Modifying the target string also resets the search position. .ie n .IP """\eG \fIassertion\fR""" 8 .el .IP "\f(CW\eG \fR\f(CIassertion\fR\f(CW\fR" 8 .IX Item "G assertion" You can intermix \f(CW\*(C`m//g\*(C'\fR matches with \f(CW\*(C`m/\eG.../g\*(C'\fR, where \f(CW\*(C`\eG\*(C'\fR is a zero-width assertion that matches the exact position where the previous \f(CW\*(C`m//g\*(C'\fR, if any, left off. Without the \f(CW\*(C`/g\*(C'\fR modifier, the \&\f(CW\*(C`\eG\*(C'\fR assertion still anchors at \f(CWpos()\fR as it was at the start of the operation (see "pos" in perlfunc), but the match is of course only attempted once. Using \f(CW\*(C`\eG\*(C'\fR without \f(CW\*(C`/g\*(C'\fR on a target string that has not previously had a \f(CW\*(C`/g\*(C'\fR match applied to it is the same as using the \f(CW\*(C`\eA\*(C'\fR assertion to match the beginning of the string. Note also that, currently, \f(CW\*(C`\eG\*(C'\fR is only properly supported when anchored at the very beginning of the pattern. .Sp Examples: .Sp .Vb 2 \& # list context \& ($one,$five,$fifteen) = (\`uptime\` =~ /(\ed+\e.\ed+)/g); \& \& # scalar context \& local $/ = ""; \& while ($paragraph = <>) { \& while ($paragraph =~ /\ep{Ll}[\*(Aq")]*[.!?]+[\*(Aq")]*\es/g) { \& $sentences++; \& } \& } \& say $sentences; .Ve .Sp Here's another way to check for sentences in a paragraph: .Sp .Vb 10 \& my $sentence_rx = qr{ \& (?: (?<= ^ ) | (?<= \es ) ) # after start\-of\-string or \& # whitespace \& \ep{Lu} # capital letter \& .*? # a bunch of anything \& (?<= \eS ) # that ends in non\- \& # whitespace \& (?) { \& say "NEW PARAGRAPH"; \& my $count = 0; \& while ($paragraph =~ /($sentence_rx)/g) { \& printf "\etgot sentence %d: <%s>\en", ++$count, $1; \& } \& } .Ve .Sp Here's how to use \f(CW\*(C`m//gc\*(C'\fR with \f(CW\*(C`\eG\*(C'\fR: .Sp .Vb 10 \& $_ = "ppooqppqq"; \& while ($i++ < 2) { \& print "1: \*(Aq"; \& print $1 while /(o)/gc; print "\*(Aq, pos=", pos, "\en"; \& print "2: \*(Aq"; \& print $1 if /\eG(q)/gc; print "\*(Aq, pos=", pos, "\en"; \& print "3: \*(Aq"; \& print $1 while /(p)/gc; print "\*(Aq, pos=", pos, "\en"; \& } \& print "Final: \*(Aq$1\*(Aq, pos=",pos,"\en" if /\eG(.)/; .Ve .Sp The last example should print: .Sp .Vb 7 \& 1: \*(Aqoo\*(Aq, pos=4 \& 2: \*(Aqq\*(Aq, pos=5 \& 3: \*(Aqpp\*(Aq, pos=7 \& 1: \*(Aq\*(Aq, pos=7 \& 2: \*(Aqq\*(Aq, pos=8 \& 3: \*(Aq\*(Aq, pos=8 \& Final: \*(Aqq\*(Aq, pos=8 .Ve .Sp Notice that the final match matched \f(CW\*(C`q\*(C'\fR instead of \f(CW\*(C`p\*(C'\fR, which a match without the \f(CW\*(C`\eG\*(C'\fR anchor would have done. Also note that the final match did not update \f(CW\*(C`pos\*(C'\fR. \f(CW\*(C`pos\*(C'\fR is only updated on a \f(CW\*(C`/g\*(C'\fR match. If the final match did indeed match \f(CW\*(C`p\*(C'\fR, it's a good bet that you're running an ancient (pre\-5.6.0) version of Perl. .Sp A useful idiom for \f(CW\*(C`lex\*(C'\fR\-like scanners is \f(CW\*(C`/\eG.../gc\*(C'\fR. You can combine several regexps like this to process a string part-by-part, doing different actions depending on which regexp matched. Each regexp tries to match where the previous one leaves off. .Sp .Vb 4 \& $_ = <<\*(AqEOL\*(Aq; \& $url = URI::URL\->new( "http://example.com/" ); \& die if $url eq "xXx"; \& EOL \& \& LOOP: { \& print(" digits"), redo LOOP if /\eG\ed+\eb[,.;]?\es*/gc; \& print(" lowercase"), redo LOOP \& if /\eG\ep{Ll}+\eb[,.;]?\es*/gc; \& print(" UPPERCASE"), redo LOOP \& if /\eG\ep{Lu}+\eb[,.;]?\es*/gc; \& print(" Capitalized"), redo LOOP \& if /\eG\ep{Lu}\ep{Ll}+\eb[,.;]?\es*/gc; \& print(" MiXeD"), redo LOOP if /\eG\epL+\eb[,.;]?\es*/gc; \& print(" alphanumeric"), redo LOOP \& if /\eG[\ep{Alpha}\epN]+\eb[,.;]?\es*/gc; \& print(" line\-noise"), redo LOOP if /\eG\eW+/gc; \& print ". That\*(Aqs all!\en"; \& } .Ve .Sp Here is the output (split into several lines): .Sp .Vb 4 \& line\-noise lowercase line\-noise UPPERCASE line\-noise UPPERCASE \& line\-noise lowercase line\-noise lowercase line\-noise lowercase \& lowercase line\-noise lowercase lowercase line\-noise lowercase \& lowercase line\-noise MiXeD line\-noise. That\*(Aqs all! .Ve .ie n .IP """m?\fIPATTERN\fR?msixpodualngc""" 8 .el .IP \f(CWm?\fR\f(CIPATTERN\fR\f(CW?msixpodualngc\fR 8 .IX Xref "? operator, match-once" .IX Item "m?PATTERN?msixpodualngc" This is just like the \f(CW\*(C`m/\fR\f(CIPATTERN\fR\f(CW/\*(C'\fR search, except that it matches only once between calls to the \f(CWreset()\fR operator. This is a useful optimization when you want to see only the first occurrence of something in each file of a set of files, for instance. Only \f(CW\*(C`m??\*(C'\fR patterns local to the current package are reset. .Sp .Vb 7 \& while (<>) { \& if (m?^$?) { \& # blank line between header and body \& } \& } continue { \& reset if eof; # clear m?? status for next file \& } .Ve .Sp Another example switched the first "latin1" encoding it finds to "utf8" in a pod file: .Sp .Vb 1 \& s//utf8/ if m? ^ =encoding \eh+ \eK latin1 ?x; .Ve .Sp The match-once behavior is controlled by the match delimiter being \&\f(CW\*(C`?\*(C'\fR; with any other delimiter this is the normal \f(CW\*(C`m//\*(C'\fR operator. .Sp In the past, the leading \f(CW\*(C`m\*(C'\fR in \f(CW\*(C`m?\fR\f(CIPATTERN\fR\f(CW?\*(C'\fR was optional, but omitting it would produce a deprecation warning. As of v5.22.0, omitting it produces a syntax error. If you encounter this construct in older code, you can just add \&\f(CW\*(C`m\*(C'\fR. .ie n .IP """s/\fIPATTERN\fR/\fIREPLACEMENT\fR/msixpodualngcer""" 8 .el .IP \f(CWs/\fR\f(CIPATTERN\fR\f(CW/\fR\f(CIREPLACEMENT\fR\f(CW/msixpodualngcer\fR 8 .IX Xref "s substitute substitution replace regexp, replace regexp, substitute m s i x p o g c e r" .IX Item "s/PATTERN/REPLACEMENT/msixpodualngcer" Searches a string for a pattern, and if found, replaces that pattern with the replacement text and returns the number of substitutions made. Otherwise it returns false (a value that is both an empty string (\f(CW""\fR) and numeric zero (\f(CW0\fR) as described in "Relational Operators"). .Sp If the \f(CW\*(C`/r\*(C'\fR (non-destructive) option is used then it runs the substitution on a copy of the string and instead of returning the number of substitutions, it returns the copy whether or not a substitution occurred. The original string is never changed when \&\f(CW\*(C`/r\*(C'\fR is used. The copy will always be a plain string, even if the input is an object or a tied variable. .Sp If no string is specified via the \f(CW\*(C`=~\*(C'\fR or \f(CW\*(C`!~\*(C'\fR operator, the \f(CW$_\fR variable is searched and modified. Unless the \f(CW\*(C`/r\*(C'\fR option is used, the string specified must be a scalar variable, an array element, a hash element, or an assignment to one of those; that is, some sort of scalar lvalue. .Sp If the delimiter chosen is a single quote, no variable interpolation is done on either the \fIPATTERN\fR or the \fIREPLACEMENT\fR. Otherwise, if the \&\fIPATTERN\fR contains a \f(CW\*(C`$\*(C'\fR that looks like a variable rather than an end-of-string test, the variable will be interpolated into the pattern at run-time. If you want the pattern compiled only once the first time the variable is interpolated, use the \f(CW\*(C`/o\*(C'\fR option. If the pattern evaluates to the empty string, the last successfully executed regular expression is used instead. See perlre for further explanation on these. .Sp Options are as with \f(CW\*(C`m//\*(C'\fR with the addition of the following replacement specific options: .Sp .Vb 5 \& e Evaluate the right side as an expression. \& ee Evaluate the right side as a string then eval the \& result. \& r Return substitution and leave the original string \& untouched. .Ve .Sp Any non-whitespace delimiter may replace the slashes. Add space after the \f(CW\*(C`s\*(C'\fR when using a character allowed in identifiers. If single quotes are used, no interpretation is done on the replacement string (the \f(CW\*(C`/e\*(C'\fR modifier overrides this, however). Note that Perl treats backticks as normal delimiters; the replacement text is not evaluated as a command. If the \fIPATTERN\fR is delimited by bracketing quotes, the \fIREPLACEMENT\fR has its own pair of quotes, which may or may not be bracketing quotes, for example, \&\f(CW\*(C`s(foo)(bar)\*(C'\fR or \f(CW\*(C`s/bar/\*(C'\fR. A \f(CW\*(C`/e\*(C'\fR will cause the replacement portion to be treated as a full-fledged Perl expression and evaluated right then and there. It is, however, syntax checked at compile-time. A second \f(CW\*(C`e\*(C'\fR modifier will cause the replacement portion to be \f(CW\*(C`eval\*(C'\fRed before being run as a Perl expression. .Sp Examples: .Sp .Vb 1 \& s/\ebgreen\eb/mauve/g; # don\*(Aqt change wintergreen \& \& $path =~ s|/usr/bin|/usr/local/bin|; \& \& s/Login: $foo/Login: $bar/; # run\-time pattern \& \& ($foo = $bar) =~ s/this/that/; # copy first, then \& # change \& ($foo = "$bar") =~ s/this/that/; # convert to string, \& # copy, then change \& $foo = $bar =~ s/this/that/r; # Same as above using /r \& $foo = $bar =~ s/this/that/r \& =~ s/that/the other/r; # Chained substitutes \& # using /r \& @foo = map { s/this/that/r } @bar # /r is very useful in \& # maps \& \& $count = ($paragraph =~ s/Mister\eb/Mr./g); # get change\-cnt \& \& $_ = \*(Aqabc123xyz\*(Aq; \& s/\ed+/$&*2/e; # yields \*(Aqabc246xyz\*(Aq \& s/\ed+/sprintf("%5d",$&)/e; # yields \*(Aqabc 246xyz\*(Aq \& s/\ew/$& x 2/eg; # yields \*(Aqaabbcc 224466xxyyzz\*(Aq \& \& s/%(.)/$percent{$1}/g; # change percent escapes; no /e \& s/%(.)/$percent{$1} || $&/ge; # expr now, so /e \& s/^=(\ew+)/pod($1)/ge; # use function call \& \& $_ = \*(Aqabc123xyz\*(Aq; \& $x = s/abc/def/r; # $x is \*(Aqdef123xyz\*(Aq and \& # $_ remains \*(Aqabc123xyz\*(Aq. \& \& # expand variables in $_, but dynamics only, using \& # symbolic dereferencing \& s/\e$(\ew+)/${$1}/g; \& \& # Add one to the value of any numbers in the string \& s/(\ed+)/1 + $1/eg; \& \& # Titlecase words in the last 30 characters only (presuming \& # that the substring doesn\*(Aqt start in the middle of a word) \& substr($str, \-30) =~ s/\eb(\ep{Alpha})(\ep{Alpha}*)\eb/\eu$1\eL$2/g; \& \& # This will expand any embedded scalar variable \& # (including lexicals) in $_ : First $1 is interpolated \& # to the variable name, and then evaluated \& s/(\e$\ew+)/$1/eeg; \& \& # Delete (most) C comments. \& $program =~ s { \& /\e* # Match the opening delimiter. \& .*? # Match a minimal number of characters. \& \e*/ # Match the closing delimiter. \& } []gsx; \& \& s/^\es*(.*?)\es*$/$1/; # trim whitespace in $_, \& # expensively \& \& for ($variable) { # trim whitespace in $variable, \& # cheap \& s/^\es+//; \& s/\es+$//; \& } \& \& s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields \& \& $foo !~ s/A/a/g; # Lowercase all A\*(Aqs in $foo; return \& # 0 if any were found and changed; \& # otherwise return 1 .Ve .Sp Note the use of \f(CW\*(C`$\*(C'\fR instead of \f(CW\*(C`\e\*(C'\fR in the last example. Unlike \&\fBsed\fR, we use the \e<\fIdigit\fR> form only in the left hand side. Anywhere else it's $<\fIdigit\fR>. .Sp Occasionally, you can't use just a \f(CW\*(C`/g\*(C'\fR to get all the changes to occur that you might want. Here are two common cases: .Sp .Vb 2 \& # put commas in the right places in an integer \& 1 while s/(\ed)(\ed\ed\ed)(?!\ed)/$1,$2/g; \& \& # expand tabs to 8\-column spacing \& 1 while s/\et+/\*(Aq \*(Aq x (length($&)*8 \- length($\`)%8)/e; .Ve .Sp While \f(CW\*(C`s///\*(C'\fR accepts the \f(CW\*(C`/c\*(C'\fR flag, it has no effect beyond producing a warning if warnings are enabled. .IX Xref " c" .SS "Quote-Like Operators" .IX Xref "operator, quote-like" .IX Subsection "Quote-Like Operators" .ie n .IP """q/\fISTRING\fR/""" 4 .el .IP \f(CWq/\fR\f(CISTRING\fR\f(CW/\fR 4 .IX Xref "q quote, single ' ''" .IX Item "q/STRING/" .PD 0 .ie n .IP \*(Aq\fISTRING\fR\*(Aq 4 .el .IP \f(CW\*(Aq\fR\f(CISTRING\fR\f(CW\*(Aq\fR 4 .IX Item "STRING" .PD A single-quoted, literal string. A backslash represents a backslash unless followed by the delimiter or another backslash, in which case the delimiter or backslash is interpolated. .Sp .Vb 3 \& $foo = q!I said, "You said, \*(AqShe said it.\*(Aq"!; \& $bar = q(\*(AqThis is it.\*(Aq); \& $baz = \*(Aq\en\*(Aq; # a two\-character string .Ve .ie n .IP """qq/\fISTRING\fR/""" 4 .el .IP \f(CWqq/\fR\f(CISTRING\fR\f(CW/\fR 4 .IX Xref "qq quote, double "" """"" .IX Item "qq/STRING/" .PD 0 .ie n .IP """\fISTRING\fR""" 4 .el .IP "\f(CW""\fR\f(CISTRING\fR\f(CW""\fR" 4 .IX Item """STRING""" .PD A double-quoted, interpolated string. .Sp .Vb 4 \& $_ .= qq \& (*** The previous line contains the naughty word "$1".\en) \& if /\eb(tcl|java|python)\eb/i; # :\-) \& $baz = "\en"; # a one\-character string .Ve .ie n .IP """qx/\fISTRING\fR/""" 4 .el .IP \f(CWqx/\fR\f(CISTRING\fR\f(CW/\fR 4 .IX Xref "qx ` `` backtick" .IX Item "qx/STRING/" .PD 0 .ie n .IP \`\fISTRING\fR\` 4 .el .IP \f(CW\`\fR\f(CISTRING\fR\f(CW\`\fR 4 .IX Item "STRING" .PD A string which is (possibly) interpolated and then executed as a system command, via \fI/bin/sh\fR or its equivalent if required. Shell wildcards, pipes, and redirections will be honored. Similarly to \&\f(CW\*(C`system\*(C'\fR, if the string contains no shell metacharacters then it will executed directly. The collected standard output of the command is returned; standard error is unaffected. In scalar context, it comes back as a single (potentially multi-line) string, or \f(CW\*(C`undef\*(C'\fR if the shell (or command) could not be started. In list context, returns a list of lines (however you've defined lines with \f(CW$/\fR or \&\f(CW$INPUT_RECORD_SEPARATOR\fR), or an empty list if the shell (or command) could not be started. .Sp Because backticks do not affect standard error, use shell file descriptor syntax (assuming the shell supports this) if you care to address this. To capture a command's STDERR and STDOUT together: .Sp .Vb 1 \& $output = \`cmd 2>&1\`; .Ve .Sp To capture a command's STDOUT but discard its STDERR: .Sp .Vb 1 \& $output = \`cmd 2>/dev/null\`; .Ve .Sp To capture a command's STDERR but discard its STDOUT (ordering is important here): .Sp .Vb 1 \& $output = \`cmd 2>&1 1>/dev/null\`; .Ve .Sp To exchange a command's STDOUT and STDERR in order to capture the STDERR but leave its STDOUT to come out the old STDERR: .Sp .Vb 1 \& $output = \`cmd 3>&1 1>&2 2>&3 3>&\-\`; .Ve .Sp To read both a command's STDOUT and its STDERR separately, it's easiest to redirect them separately to files, and then read from those files when the program is done: .Sp .Vb 1 \& system("program args 1>program.stdout 2>program.stderr"); .Ve .Sp The STDIN filehandle used by the command is inherited from Perl's STDIN. For example: .Sp .Vb 3 \& open(SPLAT, "stuff") || die "can\*(Aqt open stuff: $!"; \& open(STDIN, "<&SPLAT") || die "can\*(Aqt dupe SPLAT: $!"; \& print STDOUT \`sort\`; .Ve .Sp will print the sorted contents of the file named \fI"stuff"\fR. .Sp Using single-quote as a delimiter protects the command from Perl's double-quote interpolation, passing it on to the shell instead: .Sp .Vb 2 \& $perl_info = qx(ps $$); # that\*(Aqs Perl\*(Aqs $$ \& $shell_info = qx\*(Aqps $$\*(Aq; # that\*(Aqs the new shell\*(Aqs $$ .Ve .Sp How that string gets evaluated is entirely subject to the command interpreter on your system. On most platforms, you will have to protect shell metacharacters if you want them treated literally. This is in practice difficult to do, as it's unclear how to escape which characters. See perlsec for a clean and safe example of a manual \f(CWfork()\fR and \f(CWexec()\fR to emulate backticks safely. .Sp On some platforms (notably DOS-like ones), the shell may not be capable of dealing with multiline commands, so putting newlines in the string may not get you what you want. You may be able to evaluate multiple commands in a single line by separating them with the command separator character, if your shell supports that (for example, \f(CW\*(C`;\*(C'\fR on many Unix shells and \f(CW\*(C`&\*(C'\fR on the Windows NT \f(CW\*(C`cmd\*(C'\fR shell). .Sp Perl will attempt to flush all files opened for output before starting the child process, but this may not be supported on some platforms (see perlport). To be safe, you may need to set \&\f(CW$|\fR (\f(CW$AUTOFLUSH\fR in \f(CW\*(C`English\*(C'\fR) or call the \f(CWautoflush()\fR method of \&\f(CW\*(C`IO::Handle\*(C'\fR on any open handles. .Sp Beware that some command shells may place restrictions on the length of the command line. You must ensure your strings don't exceed this limit after any necessary interpolations. See the platform-specific release notes for more details about your particular environment. .Sp Using this operator can lead to programs that are difficult to port, because the shell commands called vary between systems, and may in fact not be present at all. As one example, the \f(CW\*(C`type\*(C'\fR command under the POSIX shell is very different from the \f(CW\*(C`type\*(C'\fR command under DOS. That doesn't mean you should go out of your way to avoid backticks when they're the right way to get something done. Perl was made to be a glue language, and one of the things it glues together is commands. Just understand what you're getting yourself into. .Sp Like \f(CW\*(C`system\*(C'\fR, backticks put the child process exit code in \f(CW$?\fR. If you'd like to manually inspect failure, you can check all possible failure modes by inspecting \f(CW$?\fR like this: .Sp .Vb 10 \& if ($? == \-1) { \& print "failed to execute: $!\en"; \& } \& elsif ($? & 127) { \& printf "child died with signal %d, %s coredump\en", \& ($? & 127), ($? & 128) ? \*(Aqwith\*(Aq : \*(Aqwithout\*(Aq; \& } \& else { \& printf "child exited with value %d\en", $? >> 8; \& } .Ve .Sp Use the open pragma to control the I/O layers used when reading the output of the command, for example: .Sp .Vb 2 \& use open IN => ":encoding(UTF\-8)"; \& my $x = \`cmd\-producing\-utf\-8\`; .Ve .Sp \&\f(CW\*(C`qx//\*(C'\fR can also be called like a function with "readpipe" in perlfunc. .Sp See "I/O Operators" for more discussion. .ie n .IP """qw/\fISTRING\fR/""" 4 .el .IP \f(CWqw/\fR\f(CISTRING\fR\f(CW/\fR 4 .IX Xref "qw quote, list quote, words" .IX Item "qw/STRING/" Evaluates to a list of the words extracted out of \fISTRING\fR, using embedded whitespace as the word delimiters. It can be understood as being roughly equivalent to: .Sp .Vb 1 \& split(" ", q/STRING/); .Ve .Sp the differences being that it only splits on ASCII whitespace, generates a real list at compile time, and in scalar context it returns the last element in the list. So this expression: .Sp .Vb 1 \& qw(foo bar baz) .Ve .Sp is semantically equivalent to the list: .Sp .Vb 1 \& "foo", "bar", "baz" .Ve .Sp Some frequently seen examples: .Sp .Vb 2 \& use POSIX qw( setlocale localeconv ) \& @EXPORT = qw( foo bar baz ); .Ve .Sp A common mistake is to try to separate the words with commas or to put comments into a multi-line \f(CW\*(C`qw\*(C'\fR\-string. For this reason, the \&\f(CW\*(C`use\ warnings\*(C'\fR pragma and the \fB\-w\fR switch (that is, the \f(CW$^W\fR variable) produces warnings if the \fISTRING\fR contains the \f(CW","\fR or the \f(CW"#"\fR character. .ie n .IP """tr/\fISEARCHLIST\fR/\fIREPLACEMENTLIST\fR/cdsr""" 4 .el .IP \f(CWtr/\fR\f(CISEARCHLIST\fR\f(CW/\fR\f(CIREPLACEMENTLIST\fR\f(CW/cdsr\fR 4 .IX Xref "tr y transliterate c d s" .IX Item "tr/SEARCHLIST/REPLACEMENTLIST/cdsr" .PD 0 .ie n .IP """y/\fISEARCHLIST\fR/\fIREPLACEMENTLIST\fR/cdsr""" 4 .el .IP \f(CWy/\fR\f(CISEARCHLIST\fR\f(CW/\fR\f(CIREPLACEMENTLIST\fR\f(CW/cdsr\fR 4 .IX Item "y/SEARCHLIST/REPLACEMENTLIST/cdsr" .PD Transliterates all occurrences of the characters found (or not found if the \f(CW\*(C`/c\*(C'\fR modifier is specified) in the search list with the positionally corresponding character in the replacement list, possibly deleting some, depending on the modifiers specified. It returns the number of characters replaced or deleted. If no string is specified via the \f(CW\*(C`=~\*(C'\fR or \f(CW\*(C`!~\*(C'\fR operator, the \f(CW$_\fR string is transliterated. .Sp For \fBsed\fR devotees, \f(CW\*(C`y\*(C'\fR is provided as a synonym for \f(CW\*(C`tr\*(C'\fR. .Sp If the \f(CW\*(C`/r\*(C'\fR (non-destructive) option is present, a new copy of the string is made and its characters transliterated, and this copy is returned no matter whether it was modified or not: the original string is always left unchanged. The new copy is always a plain string, even if the input string is an object or a tied variable. .Sp Unless the \f(CW\*(C`/r\*(C'\fR option is used, the string specified with \f(CW\*(C`=~\*(C'\fR must be a scalar variable, an array element, a hash element, or an assignment to one of those; in other words, an lvalue. .Sp The characters delimitting \fISEARCHLIST\fR and \fIREPLACEMENTLIST\fR can be any printable character, not just forward slashes. If they are single quotes (\f(CW\*(C`tr\*(Aq\fR\f(CISEARCHLIST\fR\f(CW\*(Aq\fR\f(CIREPLACEMENTLIST\fR\f(CW\*(Aq\*(C'\fR), the only interpolation is removal of \f(CW\*(C`\e\*(C'\fR from pairs of \f(CW\*(C`\e\e\*(C'\fR; so hyphens are interpreted literally rather than specifying a character range. .Sp Otherwise, a character range may be specified with a hyphen, so \&\f(CW\*(C`tr/A\-J/0\-9/\*(C'\fR does the same replacement as \&\f(CW\*(C`tr/ACEGIBDFHJ/0246813579/\*(C'\fR. .Sp If the \fISEARCHLIST\fR is delimited by bracketing quotes, the \&\fIREPLACEMENTLIST\fR must have its own pair of quotes, which may or may not be bracketing quotes; for example, \f(CW\*(C`tr(aeiouy)(yuoiea)\*(C'\fR or \&\f(CW\*(C`tr[+\e\-*/]"ABCD"\*(C'\fR. This final example shows a way to visually clarify what is going on for people who are more familiar with regular expression patterns than with \f(CW\*(C`tr\*(C'\fR, and who may think forward slash delimiters imply that \f(CW\*(C`tr\*(C'\fR is more like a regular expression pattern than it actually is. (Another option might be to use \f(CW\*(C`tr[...][...]\*(C'\fR.) .Sp \&\f(CW\*(C`tr\*(C'\fR isn't fully like bracketed character classes, just (significantly) more like them than it is to full patterns. For example, characters appearing more than once in either list behave differently here than in patterns, and \f(CW\*(C`tr\*(C'\fR lists do not allow backslashed character classes such as \f(CW\*(C`\ed\*(C'\fR or \f(CW\*(C`\epL\*(C'\fR, nor variable interpolation, so \f(CW"$"\fR and \f(CW"@"\fR are always treated as literals. .Sp The allowed elements are literals plus \f(CW\*(C`\e\*(Aq\*(C'\fR (meaning a single quote). If the delimiters aren't single quotes, also allowed are any of the escape sequences accepted in double-quoted strings. Escape sequence details are in the table near the beginning of this section. .Sp A hyphen at the beginning or end, or preceded by a backslash is also always considered a literal. Precede a delimiter character with a backslash to allow it. .Sp The \f(CW\*(C`tr\*(C'\fR operator is not equivalent to the \f(CWtr(1)\fR utility. \&\f(CW\*(C`tr[a\-z][A\-Z]\*(C'\fR will uppercase the 26 letters "a" through "z", but for case changing not confined to ASCII, use \f(CW\*(C`lc\*(C'\fR, \&\f(CW\*(C`uc\*(C'\fR, \f(CW\*(C`lcfirst\*(C'\fR, \&\f(CW\*(C`ucfirst\*(C'\fR (all documented in perlfunc), or the substitution operator \&\f(CW\*(C`s/\fR\f(CIPATTERN\fR\f(CW/\fR\f(CIREPLACEMENT\fR\f(CW/\*(C'\fR (with \f(CW\*(C`\eU\*(C'\fR, \f(CW\*(C`\eu\*(C'\fR, \f(CW\*(C`\eL\*(C'\fR, and \f(CW\*(C`\el\*(C'\fR string-interpolation escapes in the \&\fIREPLACEMENT\fR portion). .Sp Most ranges are unportable between character sets, but certain ones signal Perl to do special handling to make them portable. There are two classes of portable ranges. The first are any subsets of the ranges \&\f(CW\*(C`A\-Z\*(C'\fR, \f(CW\*(C`a\-z\*(C'\fR, and \f(CW\*(C`0\-9\*(C'\fR, when expressed as literal characters. .Sp .Vb 1 \& tr/h\-k/H\-K/ .Ve .Sp capitalizes the letters \f(CW"h"\fR, \f(CW"i"\fR, \f(CW"j"\fR, and \f(CW"k"\fR and nothing else, no matter what the platform's character set is. In contrast, all of .Sp .Vb 3 \& tr/\ex68\-\ex6B/\ex48\-\ex4B/ \& tr/h\-\ex6B/H\-\ex4B/ \& tr/\ex68\-k/\ex48\-K/ .Ve .Sp do the same capitalizations as the previous example when run on ASCII platforms, but something completely different on EBCDIC ones. .Sp The second class of portable ranges is invoked when one or both of the range's end points are expressed as \f(CW\*(C`\eN{...}\*(C'\fR .Sp .Vb 1 \& $string =~ tr/\eN{U+20}\-\eN{U+7E}//d; .Ve .Sp removes from \f(CW$string\fR all the platform's characters which are equivalent to any of Unicode U+0020, U+0021, ... U+007D, U+007E. This is a portable range, and has the same effect on every platform it is run on. In this example, these are the ASCII printable characters. So after this is run, \f(CW$string\fR has only controls and characters which have no ASCII equivalents. .Sp But, even for portable ranges, it is not generally obvious what is included without having to look things up in the manual. A sound principle is to use only ranges that both begin from, and end at, either ASCII alphabetics of equal case (\f(CW\*(C`b\-e\*(C'\fR, \f(CW\*(C`B\-E\*(C'\fR), or digits (\f(CW\*(C`1\-4\*(C'\fR). Anything else is unclear (and unportable unless \f(CW\*(C`\eN{...}\*(C'\fR is used). If in doubt, spell out the character sets in full. .Sp Options: .Sp .Vb 5 \& c Complement the SEARCHLIST. \& d Delete found but unreplaced characters. \& r Return the modified string and leave the original string \& untouched. \& s Squash duplicate replaced characters. .Ve .Sp If the \f(CW\*(C`/d\*(C'\fR modifier is specified, any characters specified by \&\fISEARCHLIST\fR not found in \fIREPLACEMENTLIST\fR are deleted. (Note that this is slightly more flexible than the behavior of some \fBtr\fR programs, which delete anything they find in the \fISEARCHLIST\fR, period.) .Sp If the \f(CW\*(C`/s\*(C'\fR modifier is specified, sequences of characters, all in a row, that were transliterated to the same character are squashed down to a single instance of that character. .Sp .Vb 2 \& my $a = "aaabbbca"; \& $a =~ tr/ab/dd/s; # $a now is "dcd" .Ve .Sp If the \f(CW\*(C`/d\*(C'\fR modifier is used, the \fIREPLACEMENTLIST\fR is always interpreted exactly as specified. Otherwise, if the \fIREPLACEMENTLIST\fR is shorter than the \fISEARCHLIST\fR, the final character, if any, is replicated until it is long enough. There won't be a final character if and only if the \&\fIREPLACEMENTLIST\fR is empty, in which case \fIREPLACEMENTLIST\fR is copied from \fISEARCHLIST\fR. An empty \fIREPLACEMENTLIST\fR is useful for counting characters in a class, or for squashing character sequences in a class. .Sp .Vb 4 \& tr/abcd// tr/abcd/abcd/ \& tr/abcd/AB/ tr/abcd/ABBB/ \& tr/abcd//d s/[abcd]//g \& tr/abcd/AB/d (tr/ab/AB/ + s/[cd]//g) \- but run together .Ve .Sp If the \f(CW\*(C`/c\*(C'\fR modifier is specified, the characters to be transliterated are the ones NOT in \fISEARCHLIST\fR, that is, it is complemented. If \&\f(CW\*(C`/d\*(C'\fR and/or \f(CW\*(C`/s\*(C'\fR are also specified, they apply to the complemented \&\fISEARCHLIST\fR. Recall, that if \fIREPLACEMENTLIST\fR is empty (except under \f(CW\*(C`/d\*(C'\fR) a copy of \fISEARCHLIST\fR is used instead. That copy is made after complementing under \f(CW\*(C`/c\*(C'\fR. \fISEARCHLIST\fR is sorted by code point order after complementing, and any \fIREPLACEMENTLIST\fR is applied to that sorted result. This means that under \f(CW\*(C`/c\*(C'\fR, the order of the characters specified in \fISEARCHLIST\fR is irrelevant. This can lead to different results on EBCDIC systems if \fIREPLACEMENTLIST\fR contains more than one character, hence it is generally non-portable to use \f(CW\*(C`/c\*(C'\fR with such a \fIREPLACEMENTLIST\fR. .Sp Another way of describing the operation is this: If \f(CW\*(C`/c\*(C'\fR is specified, the \fISEARCHLIST\fR is sorted by code point order, then complemented. If \fIREPLACEMENTLIST\fR is empty and \f(CW\*(C`/d\*(C'\fR is not specified, \fIREPLACEMENTLIST\fR is replaced by a copy of \fISEARCHLIST\fR (as modified under \f(CW\*(C`/c\*(C'\fR), and these potentially modified lists are used as the basis for what follows. Any character in the target string that isn't in \fISEARCHLIST\fR is passed through unchanged. Every other character in the target string is replaced by the character in \&\fIREPLACEMENTLIST\fR that positionally corresponds to its mate in \&\fISEARCHLIST\fR, except that under \f(CW\*(C`/s\*(C'\fR, the 2nd and following characters are squeezed out in a sequence of characters in a row that all translate to the same character. If \fISEARCHLIST\fR is longer than \&\fIREPLACEMENTLIST\fR, characters in the target string that match a character in \fISEARCHLIST\fR that doesn't have a correspondence in \&\fIREPLACEMENTLIST\fR are either deleted from the target string if \f(CW\*(C`/d\*(C'\fR is specified; or replaced by the final character in \fIREPLACEMENTLIST\fR if \&\f(CW\*(C`/d\*(C'\fR isn't specified. .Sp Some examples: .Sp .Vb 1 \& $ARGV[1] =~ tr/A\-Z/a\-z/; # canonicalize to lower case ASCII \& \& $cnt = tr/*/*/; # count the stars in $_ \& $cnt = tr/*//; # same thing \& \& $cnt = $sky =~ tr/*/*/; # count the stars in $sky \& $cnt = $sky =~ tr/*//; # same thing \& \& $cnt = $sky =~ tr/*//c; # count all the non\-stars in $sky \& $cnt = $sky =~ tr/*/*/c; # same, but transliterate each non\-star \& # into a star, leaving the already\-stars \& # alone. Afterwards, everything in $sky \& # is a star. \& \& $cnt = tr/0\-9//; # count the ASCII digits in $_ \& \& tr/a\-zA\-Z//s; # bookkeeper \-> bokeper \& tr/o/o/s; # bookkeeper \-> bokkeeper \& tr/oe/oe/s; # bookkeeper \-> bokkeper \& tr/oe//s; # bookkeeper \-> bokkeper \& tr/oe/o/s; # bookkeeper \-> bokkopor \& \& ($HOST = $host) =~ tr/a\-z/A\-Z/; \& $HOST = $host =~ tr/a\-z/A\-Z/r; # same thing \& \& $HOST = $host =~ tr/a\-z/A\-Z/r # chained with s///r \& =~ s/:/ \-p/r; \& \& tr/a\-zA\-Z/ /cs; # change non\-alphas to single space \& \& @stripped = map tr/a\-zA\-Z/ /csr, @original; \& # /r with map \& \& tr [\e200\-\e377] \& [\e000\-\e177]; # wickedly delete 8th bit \& \& $foo !~ tr/A/a/ # transliterate all the A\*(Aqs in $foo to \*(Aqa\*(Aq, \& # return 0 if any were found and changed. \& # Otherwise return 1 .Ve .Sp If multiple transliterations are given for a character, only the first one is used: .Sp .Vb 1 \& tr/AAA/XYZ/ .Ve .Sp will transliterate any A to X. .Sp Because the transliteration table is built at compile time, neither the \fISEARCHLIST\fR nor the \fIREPLACEMENTLIST\fR are subjected to double quote interpolation. That means that if you want to use variables, you must use an \f(CWeval()\fR: .Sp .Vb 2 \& eval "tr/$oldlist/$newlist/"; \& die $@ if $@; \& \& eval "tr/$oldlist/$newlist/, 1" or die $@; .Ve .ie n .IP """<<\fIEOF\fR""" 4 .el .IP \f(CW<<\fR\f(CIEOF\fR\f(CW\fR 4 .IX Xref "here-doc heredoc here-document <<" .IX Item "<\*(C'\fR). If the starting delimiter is an unpaired character like \f(CW\*(C`/\*(C'\fR or a closing punctuation, the ending delimiter is the same as the starting delimiter. Therefore a \f(CW\*(C`/\*(C'\fR terminates a \f(CW\*(C`qq//\*(C'\fR construct, while a \f(CW\*(C`]\*(C'\fR terminates both \f(CW\*(C`qq[]\*(C'\fR and \f(CW\*(C`qq]]\*(C'\fR constructs. .Sp When searching for single-character delimiters, escaped delimiters and \f(CW\*(C`\e\e\*(C'\fR are skipped. For example, while searching for terminating \f(CW\*(C`/\*(C'\fR, combinations of \f(CW\*(C`\e\e\*(C'\fR and \f(CW\*(C`\e/\*(C'\fR are skipped. If the delimiters are bracketing, nested pairs are also skipped. For example, while searching for a closing \f(CW\*(C`]\*(C'\fR paired with the opening \f(CW\*(C`[\*(C'\fR, combinations of \f(CW\*(C`\e\e\*(C'\fR, \f(CW\*(C`\e]\*(C'\fR, and \f(CW\*(C`\e[\*(C'\fR are all skipped, and nested \f(CW\*(C`[\*(C'\fR and \f(CW\*(C`]\*(C'\fR are skipped as well. However, when backslashes are used as the delimiters (like \f(CW\*(C`qq\e\e\*(C'\fR and \&\f(CW\*(C`tr\e\e\e\*(C'\fR), nothing is skipped. During the search for the end, backslashes that escape delimiters or other backslashes are removed (exactly speaking, they are not copied to the safe location). .Sp For constructs with three-part delimiters (\f(CW\*(C`s///\*(C'\fR, \f(CW\*(C`y///\*(C'\fR, and \&\f(CW\*(C`tr///\*(C'\fR), the search is repeated once more. If the first delimiter is not an opening punctuation, the three delimiters must be the same, such as \f(CW\*(C`s!!!\*(C'\fR and \f(CW\*(C`tr)))\*(C'\fR, in which case the second delimiter terminates the left part and starts the right part at once. If the left part is delimited by bracketing punctuation (that is \f(CW\*(C`()\*(C'\fR, \&\f(CW\*(C`[]\*(C'\fR, \f(CW\*(C`{}\*(C'\fR, or \f(CW\*(C`<>\*(C'\fR), the right part needs another pair of delimiters such as \f(CW\*(C`s(){}\*(C'\fR and \f(CW\*(C`tr[]//\*(C'\fR. In these cases, whitespace and comments are allowed between the two parts, although the comment must follow at least one whitespace character; otherwise a character expected as the start of the comment may be regarded as the starting delimiter of the right part. .Sp During this search no attention is paid to the semantics of the construct. Thus: .Sp .Vb 1 \& "$hash{"$foo/$bar"}" .Ve .Sp or: .Sp .Vb 3 \& m/ \& bar # NOT a comment, this slash / terminated m//! \& /x .Ve .Sp do not form legal quoted expressions. The quoted part ends on the first \f(CW\*(C`"\*(C'\fR and \f(CW\*(C`/\*(C'\fR, and the rest happens to be a syntax error. Because the slash that terminated \f(CW\*(C`m//\*(C'\fR was followed by a \f(CW\*(C`SPACE\*(C'\fR, the example above is not \f(CW\*(C`m//x\*(C'\fR, but rather \f(CW\*(C`m//\*(C'\fR with no \f(CW\*(C`/x\*(C'\fR modifier. So the embedded \f(CW\*(C`#\*(C'\fR is interpreted as a literal \f(CW\*(C`#\*(C'\fR. .Sp Also no attention is paid to \f(CW\*(C`\ec\e\*(C'\fR (multichar control char syntax) during this search. Thus the second \f(CW\*(C`\e\*(C'\fR in \f(CW\*(C`qq/\ec\e/\*(C'\fR is interpreted as a part of \f(CW\*(C`\e/\*(C'\fR, and the following \f(CW\*(C`/\*(C'\fR is not recognized as a delimiter. Instead, use \f(CW\*(C`\e034\*(C'\fR or \f(CW\*(C`\ex1c\*(C'\fR at the end of quoted constructs. .IP Interpolation 4 .IX Xref "interpolation" .IX Item "Interpolation" The next step is interpolation in the text obtained, which is now delimiter-independent. There are multiple cases. .RS 4 .ie n .IP """<<\*(AqEOF\*(Aq""" 4 .el .IP \f(CW<<\*(AqEOF\*(Aq\fR 4 .IX Item "<"", ""<<""EOF""""" 4 .el .IP "\f(CW""""\fR, \f(CW\`\`\fR, \f(CWqq//\fR, \f(CWqx//\fR, \f(CW\fR, \f(CW<<""EOF""\fR" 4 .IX Item """"", , qq//, qx//, , <<""EOF""" \&\f(CW\*(C`\eQ\*(C'\fR, \f(CW\*(C`\eU\*(C'\fR, \f(CW\*(C`\eu\*(C'\fR, \f(CW\*(C`\eL\*(C'\fR, \f(CW\*(C`\el\*(C'\fR, \f(CW\*(C`\eF\*(C'\fR (possibly paired with \f(CW\*(C`\eE\*(C'\fR) are converted to corresponding Perl constructs. Thus, \f(CW"$foo\eQbaz$bar"\fR is converted to \f(CW\*(C`$foo\ .\ (quotemeta("baz"\ .\ $bar))\*(C'\fR internally. The other escape sequences such as \f(CW\*(C`\e200\*(C'\fR and \f(CW\*(C`\et\*(C'\fR and backslashed characters such as \f(CW\*(C`\e\e\*(C'\fR and \f(CW\*(C`\e\-\*(C'\fR are replaced with appropriate expansions. .Sp Let it be stressed that \fIwhatever falls between \fR\f(CI\*(C`\eQ\*(C'\fR\fI and \fR\f(CI\*(C`\eE\*(C'\fR is interpolated in the usual way. Something like \f(CW"\eQ\e\eE"\fR has no \f(CW\*(C`\eE\*(C'\fR inside. Instead, it has \f(CW\*(C`\eQ\*(C'\fR, \f(CW\*(C`\e\e\*(C'\fR, and \f(CW\*(C`E\*(C'\fR, so the result is the same as for \f(CW"\e\e\e\eE"\fR. As a general rule, backslashes between \f(CW\*(C`\eQ\*(C'\fR and \f(CW\*(C`\eE\*(C'\fR may lead to counterintuitive results. So, \&\f(CW"\eQ\et\eE"\fR is converted to \f(CWquotemeta("\et")\fR, which is the same as \f(CW"\e\e\et"\fR (since TAB is not alphanumeric). Note also that: .Sp .Vb 2 \& $str = \*(Aq\et\*(Aq; \& return "\eQ$str"; .Ve .Sp may be closer to the conjectural \fIintention\fR of the writer of \f(CW"\eQ\et\eE"\fR. .Sp Interpolated scalars and arrays are converted internally to the \f(CW\*(C`join\*(C'\fR and \&\f(CW"."\fR catenation operations. Thus, \f(CW"$foo\ XXX\ \*(Aq@arr\*(Aq"\fR becomes: .Sp .Vb 1 \& $foo . " XXX \*(Aq" . (join $", @arr) . "\*(Aq"; .Ve .Sp All operations above are performed simultaneously, left to right. .Sp Because the result of \f(CW"\eQ\ \fR\f(CISTRING\fR\f(CW\ \eE"\fR has all metacharacters quoted, there is no way to insert a literal \f(CW\*(C`$\*(C'\fR or \f(CW\*(C`@\*(C'\fR inside a \&\f(CW\*(C`\eQ\eE\*(C'\fR pair. If protected by \f(CW\*(C`\e\*(C'\fR, \f(CW\*(C`$\*(C'\fR will be quoted to become \&\f(CW"\e\e\e$"\fR; if not, it is interpreted as the start of an interpolated scalar. .Sp Note also that the interpolation code needs to make a decision on where the interpolated scalar ends. For instance, whether \&\f(CW"a\ $x\ \->\ {c}"\fR really means: .Sp .Vb 1 \& "a " . $x . " \-> {c}"; .Ve .Sp or: .Sp .Vb 1 \& "a " . $x \-> {c}; .Ve .Sp Most of the time, the longest possible text that does not include spaces between components and which contains matching braces or brackets. because the outcome may be determined by voting based on heuristic estimators, the result is not strictly predictable. Fortunately, it's usually correct for ambiguous cases. .ie n .IP "The replacement of ""s///""" 4 .el .IP "The replacement of \f(CWs///\fR" 4 .IX Item "The replacement of s///" Processing of \f(CW\*(C`\eQ\*(C'\fR, \f(CW\*(C`\eU\*(C'\fR, \f(CW\*(C`\eu\*(C'\fR, \f(CW\*(C`\eL\*(C'\fR, \f(CW\*(C`\el\*(C'\fR, \f(CW\*(C`\eF\*(C'\fR and interpolation happens as with \f(CW\*(C`qq//\*(C'\fR constructs. .Sp It is at this step that \f(CW\*(C`\e1\*(C'\fR is begrudgingly converted to \f(CW$1\fR in the replacement text of \f(CW\*(C`s///\*(C'\fR, in order to correct the incorrigible \&\fIsed\fR hackers who haven't picked up the saner idiom yet. A warning is emitted if the \f(CW\*(C`use\ warnings\*(C'\fR pragma or the \fB\-w\fR command-line flag (that is, the \f(CW$^W\fR variable) was set. .ie n .IP """RE"" in ""m?RE?"", ""/RE/"", ""m/RE/"", ""s/RE/foo/""," 4 .el .IP "\f(CWRE\fR in \f(CWm?RE?\fR, \f(CW/RE/\fR, \f(CWm/RE/\fR, \f(CWs/RE/foo/\fR," 4 .IX Item "RE in m?RE?, /RE/, m/RE/, s/RE/foo/," Processing of \f(CW\*(C`\eQ\*(C'\fR, \f(CW\*(C`\eU\*(C'\fR, \f(CW\*(C`\eu\*(C'\fR, \f(CW\*(C`\eL\*(C'\fR, \f(CW\*(C`\el\*(C'\fR, \f(CW\*(C`\eF\*(C'\fR, \f(CW\*(C`\eE\*(C'\fR, and interpolation happens (almost) as with \f(CW\*(C`qq//\*(C'\fR constructs. .Sp Processing of \f(CW\*(C`\eN{...}\*(C'\fR is also done here, and compiled into an intermediate form for the regex compiler. (This is because, as mentioned below, the regex compilation may be done at execution time, and \f(CW\*(C`\eN{...}\*(C'\fR is a compile-time construct.) .Sp However any other combinations of \f(CW\*(C`\e\*(C'\fR followed by a character are not substituted but only skipped, in order to parse them as regular expressions at the following step. As \f(CW\*(C`\ec\*(C'\fR is skipped at this step, \f(CW\*(C`@\*(C'\fR of \f(CW\*(C`\ec@\*(C'\fR in RE is possibly treated as an array symbol (for example \f(CW@foo\fR), even though the same text in \f(CW\*(C`qq//\*(C'\fR gives interpolation of \f(CW\*(C`\ec@\*(C'\fR. .Sp Code blocks such as \f(CW\*(C`(?{BLOCK})\*(C'\fR are handled by temporarily passing control back to the perl parser, in a similar way that an interpolated array subscript expression such as \f(CW"foo$array[1+f("[xyz")]bar"\fR would be. .Sp Moreover, inside \f(CW\*(C`(?{BLOCK})\*(C'\fR, \f(CW\*(C`(?#\ comment\ )\*(C'\fR, and a \f(CW\*(C`#\*(C'\fR\-comment in a \f(CW\*(C`/x\*(C'\fR\-regular expression, no processing is performed whatsoever. This is the first step at which the presence of the \f(CW\*(C`/x\*(C'\fR modifier is relevant. .Sp Interpolation in patterns has several quirks: \f(CW$|\fR, \f(CW$(\fR, \f(CW$)\fR, \f(CW\*(C`@+\*(C'\fR and \f(CW\*(C`@\-\*(C'\fR are not interpolated, and constructs \f(CW$var[SOMETHING]\fR are voted (by several different estimators) to be either an array element or \f(CW$var\fR followed by an RE alternative. This is where the notation \&\f(CW\*(C`${arr[$bar]}\*(C'\fR comes handy: \f(CW\*(C`/${arr[0\-9]}/\*(C'\fR is interpreted as array element \f(CW\-9\fR, not as a regular expression from the variable \&\f(CW$arr\fR followed by a digit, which would be the interpretation of \&\f(CW\*(C`/$arr[0\-9]/\*(C'\fR. Since voting among different estimators may occur, the result is not predictable. .Sp The lack of processing of \f(CW\*(C`\e\e\*(C'\fR creates specific restrictions on the post-processed text. If the delimiter is \f(CW\*(C`/\*(C'\fR, one cannot get the combination \f(CW\*(C`\e/\*(C'\fR into the result of this step. \f(CW\*(C`/\*(C'\fR will finish the regular expression, \f(CW\*(C`\e/\*(C'\fR will be stripped to \f(CW\*(C`/\*(C'\fR on the previous step, and \f(CW\*(C`\e\e/\*(C'\fR will be left as is. Because \f(CW\*(C`/\*(C'\fR is equivalent to \f(CW\*(C`\e/\*(C'\fR inside a regular expression, this does not matter unless the delimiter happens to be character special to the RE engine, such as in \f(CW\*(C`s*foo*bar*\*(C'\fR, \f(CW\*(C`m[foo]\*(C'\fR, or \f(CW\*(C`m?foo?\*(C'\fR; or an alphanumeric char, as in: .Sp .Vb 1 \& m m ^ a \es* b mmx; .Ve .Sp In the RE above, which is intentionally obfuscated for illustration, the delimiter is \f(CW\*(C`m\*(C'\fR, the modifier is \f(CW\*(C`mx\*(C'\fR, and after delimiter-removal the RE is the same as for \f(CW\*(C`m/\ ^\ a\ \es*\ b\ /mx\*(C'\fR. There's more than one reason you're encouraged to restrict your delimiters to non-alphanumeric, non-whitespace choices. .RE .RS 4 .Sp This step is the last one for all constructs except regular expressions, which are processed further. .RE .IP "Parsing regular expressions" 4 .IX Xref "regexp, parse" .IX Item "Parsing regular expressions" Previous steps were performed during the compilation of Perl code, but this one happens at run time, although it may be optimized to be calculated at compile time if appropriate. After preprocessing described above, and possibly after evaluation if concatenation, joining, casing translation, or metaquoting are involved, the resulting \fIstring\fR is passed to the RE engine for compilation. .Sp Whatever happens in the RE engine might be better discussed in perlre, but for the sake of continuity, we shall do so here. .Sp This is another step where the presence of the \f(CW\*(C`/x\*(C'\fR modifier is relevant. The RE engine scans the string from left to right and converts it into a finite automaton. .Sp Backslashed characters are either replaced with corresponding literal strings (as with \f(CW\*(C`\e{\*(C'\fR), or else they generate special nodes in the finite automaton (as with \f(CW\*(C`\eb\*(C'\fR). Characters special to the RE engine (such as \f(CW\*(C`|\*(C'\fR) generate corresponding nodes or groups of nodes. \f(CW\*(C`(?#...)\*(C'\fR comments are ignored. All the rest is either converted to literal strings to match, or else is ignored (as is whitespace and \f(CW\*(C`#\*(C'\fR\-style comments if \f(CW\*(C`/x\*(C'\fR is present). .Sp Parsing of the bracketed character class construct, \f(CW\*(C`[...]\*(C'\fR, is rather different than the rule used for the rest of the pattern. The terminator of this construct is found using the same rules as for finding the terminator of a \f(CW\*(C`{}\*(C'\fR\-delimited construct, the only exception being that \f(CW\*(C`]\*(C'\fR immediately following \f(CW\*(C`[\*(C'\fR is treated as though preceded by a backslash. .Sp The terminator of runtime \f(CW\*(C`(?{...})\*(C'\fR is found by temporarily switching control to the perl parser, which should stop at the point where the logically balancing terminating \f(CW\*(C`}\*(C'\fR is found. .Sp It is possible to inspect both the string given to RE engine and the resulting finite automaton. See the arguments \f(CW\*(C`debug\*(C'\fR/\f(CW\*(C`debugcolor\*(C'\fR in the \f(CW\*(C`use\ re\*(C'\fR pragma, as well as Perl's \fB\-Dr\fR command-line switch documented in "Command Switches" in perlrun. .IP "Optimization of regular expressions" 4 .IX Xref "regexp, optimization" .IX Item "Optimization of regular expressions" This step is listed for completeness only. Since it does not change semantics, details of this step are not documented and are subject to change without notice. This step is performed over the finite automaton that was generated during the previous pass. .Sp It is at this stage that \f(CWsplit()\fR silently optimizes \f(CW\*(C`/^/\*(C'\fR to mean \f(CW\*(C`/^/m\*(C'\fR. .SS "I/O Operators" .IX Xref "operator, i o operator, io io while filehandle <> <<>> @ARGV" .IX Subsection "I/O Operators" There are several I/O operators you should know about. .PP A string enclosed by backticks (grave accents) first undergoes double-quote interpolation. It is then interpreted as an external command, and the output of that command is the value of the backtick string, like in a shell. In scalar context, a single string consisting of all output is returned. In list context, a list of values is returned, one per line of output. (You can set \f(CW$/\fR to use a different line terminator.) The command is executed each time the pseudo-literal is evaluated. The status value of the command is returned in \f(CW$?\fR (see perlvar for the interpretation of \f(CW$?\fR). Unlike in \fBcsh\fR, no translation is done on the return data\-\-newlines remain newlines. Unlike in any of the shells, single quotes do not hide variable names in the command from interpretation. To pass a literal dollar-sign through to the shell you need to hide it with a backslash. The generalized form of backticks is \f(CW\*(C`qx//\*(C'\fR, or you can call the "readpipe" in perlfunc function. (Because backticks always undergo shell expansion as well, see perlsec for security concerns.) .IX Xref "qx ` `` backtick glob" .PP In scalar context, evaluating a filehandle in angle brackets yields the next line from that file (the newline, if any, included), or \&\f(CW\*(C`undef\*(C'\fR at end-of-file or on error. When \f(CW$/\fR is set to \f(CW\*(C`undef\*(C'\fR (sometimes known as file-slurp mode) and the file is empty, it returns \f(CW\*(Aq\*(Aq\fR the first time, followed by \f(CW\*(C`undef\*(C'\fR subsequently. .PP Ordinarily you must assign the returned value to a variable, but there is one situation where an automatic assignment happens. If and only if the input symbol is the only thing inside the conditional of a \f(CW\*(C`while\*(C'\fR statement (even if disguised as a \f(CWfor(;;)\fR loop), the value is automatically assigned to the global variable \f(CW$_\fR, destroying whatever was there previously. (This may seem like an odd thing to you, but you'll use the construct in almost every Perl script you write.) The \f(CW$_\fR variable is not implicitly localized. You'll have to put a \f(CW\*(C`local\ $_;\*(C'\fR before the loop if you want that to happen. Furthermore, if the input symbol or an explicit assignment of the input symbol to a scalar is used as a \f(CW\*(C`while\*(C'\fR/\f(CW\*(C`for\*(C'\fR condition, then the condition actually tests for definedness of the expression's value, not for its regular truth value. .PP Thus the following lines are equivalent: .PP .Vb 7 \& while (defined($_ = )) { print; } \& while ($_ = ) { print; } \& while () { print; } \& for (;;) { print; } \& print while defined($_ = ); \& print while ($_ = ); \& print while ; .Ve .PP This also behaves similarly, but assigns to a lexical variable instead of to \f(CW$_\fR: .PP .Vb 1 \& while (my $line = ) { print $line } .Ve .PP In these loop constructs, the assigned value (whether assignment is automatic or explicit) is then tested to see whether it is defined. The defined test avoids problems where the line has a string value that would be treated as false by Perl; for example a "" or a \f(CW"0"\fR with no trailing newline. If you really mean for such values to terminate the loop, they should be tested for explicitly: .PP .Vb 2 \& while (($_ = ) ne \*(Aq0\*(Aq) { ... } \& while () { last unless $_; ... } .Ve .PP In other boolean contexts, \f(CW\*(C`<\fR\f(CIFILEHANDLE\fR\f(CW>\*(C'\fR without an explicit \f(CW\*(C`defined\*(C'\fR test or comparison elicits a warning if the \&\f(CW\*(C`use\ warnings\*(C'\fR pragma or the \fB\-w\fR command-line switch (the \f(CW$^W\fR variable) is in effect. .PP The filehandles STDIN, STDOUT, and STDERR are predefined. (The filehandles \f(CW\*(C`stdin\*(C'\fR, \f(CW\*(C`stdout\*(C'\fR, and \f(CW\*(C`stderr\*(C'\fR will also work except in packages, where they would be interpreted as local identifiers rather than global.) Additional filehandles may be created with the \f(CWopen()\fR function, amongst others. See perlopentut and "open" in perlfunc for details on this. .IX Xref "stdin stdout sterr" .PP If a \f(CW\*(C`<\fR\f(CIFILEHANDLE\fR\f(CW>\*(C'\fR is used in a context that is looking for a list, a list comprising all input lines is returned, one line per list element. It's easy to grow to a rather large data space this way, so use with care. .PP \&\f(CW\*(C`<\fR\f(CIFILEHANDLE\fR\f(CW>\*(C'\fR may also be spelled \f(CWreadline(*\fR\f(CIFILEHANDLE\fR\f(CW)\fR. See "readline" in perlfunc. .PP The null filehandle \f(CW\*(C`<>\*(C'\fR (sometimes called the diamond operator) is special: it can be used to emulate the behavior of \fBsed\fR and \fBawk\fR, and any other Unix filter program that takes a list of filenames, doing the same to each line of input from all of them. Input from \f(CW\*(C`<>\*(C'\fR comes either from standard input, or from each file listed on the command line. Here's how it works: the first time \f(CW\*(C`<>\*(C'\fR is evaluated, the \f(CW@ARGV\fR array is checked, and if it is empty, \f(CW$ARGV[0]\fR is set to \f(CW"\-"\fR, which when opened gives you standard input. The \f(CW@ARGV\fR array is then processed as a list of filenames. The loop .PP .Vb 3 \& while (<>) { \& ... # code for each line \& } .Ve .PP is equivalent to the following Perl-like pseudo code: .PP .Vb 7 \& unshift(@ARGV, \*(Aq\-\*(Aq) unless @ARGV; \& while ($ARGV = shift) { \& open(ARGV, $ARGV); \& while () { \& ... # code for each line \& } \& } .Ve .PP except that it isn't so cumbersome to say, and will actually work. It really does shift the \f(CW@ARGV\fR array and put the current filename into the \f(CW$ARGV\fR variable. It also uses filehandle \fIARGV\fR internally. \f(CW\*(C`<>\*(C'\fR is just a synonym for \f(CW\*(C`\*(C'\fR, which is magical. (The pseudo code above doesn't work because it treats \&\f(CW\*(C`\*(C'\fR as non-magical.) .PP Since the null filehandle uses the two argument form of "open" in perlfunc it interprets special characters, so if you have a script like this: .PP .Vb 3 \& while (<>) { \& print; \& } .Ve .PP and call it with \f(CW\*(C`perl\ dangerous.pl\ \*(Aqrm\ \-rfv\ *|\*(Aq\*(C'\fR, it actually opens a pipe, executes the \f(CW\*(C`rm\*(C'\fR command and reads \f(CW\*(C`rm\*(C'\fR's output from that pipe. If you want all items in \f(CW@ARGV\fR to be interpreted as file names, you can use the module \f(CW\*(C`ARGV::readonly\*(C'\fR from CPAN, or use the double diamond bracket: .PP .Vb 3 \& while (<<>>) { \& print; \& } .Ve .PP Using double angle brackets inside of a while causes the open to use the three argument form (with the second argument being \f(CW\*(C`<\*(C'\fR), so all arguments in \f(CW\*(C`ARGV\*(C'\fR are treated as literal filenames (including \f(CW"\-"\fR). (Note that for convenience, if you use \f(CW\*(C`<<>>\*(C'\fR and if \f(CW@ARGV\fR is empty, it will still read from the standard input.) .PP You can modify \f(CW@ARGV\fR before the first \f(CW\*(C`<>\*(C'\fR as long as the array ends up containing the list of filenames you really want. Line numbers (\f(CW$.\fR) continue as though the input were one big happy file. See the example in "eof" in perlfunc for how to reset line numbers on each file. .PP If you want to set \f(CW@ARGV\fR to your own list of files, go right ahead. This sets \f(CW@ARGV\fR to all plain text files if no \f(CW@ARGV\fR was given: .PP .Vb 1 \& @ARGV = grep { \-f && \-T } glob(\*(Aq*\*(Aq) unless @ARGV; .Ve .PP You can even set them to pipe commands. For example, this automatically filters compressed arguments through \fBgzip\fR: .PP .Vb 1 \& @ARGV = map { /\e.(gz|Z)$/ ? "gzip \-dc < $_ |" : $_ } @ARGV; .Ve .PP If you want to pass switches into your script, you can use one of the \&\f(CW\*(C`Getopts\*(C'\fR modules or put a loop on the front like this: .PP .Vb 7 \& while ($_ = $ARGV[0], /^\-/) { \& shift; \& last if /^\-\-$/; \& if (/^\-D(.*)/) { $debug = $1 } \& if (/^\-v/) { $verbose++ } \& # ... # other switches \& } \& \& while (<>) { \& # ... # code for each line \& } .Ve .PP The \f(CW\*(C`<>\*(C'\fR symbol will return \f(CW\*(C`undef\*(C'\fR for end-of-file only once. If you call it again after this, it will assume you are processing another \&\f(CW@ARGV\fR list, and if you haven't set \f(CW@ARGV\fR, will read input from STDIN. .PP If what the angle brackets contain is a simple scalar variable (for example, \&\f(CW$foo\fR), then that variable contains the name of the filehandle to input from, or its typeglob, or a reference to the same. For example: .PP .Vb 2 \& $fh = \e*STDIN; \& $line = <$fh>; .Ve .PP If what's within the angle brackets is neither a filehandle nor a simple scalar variable containing a filehandle name, typeglob, or typeglob reference, it is interpreted as a filename pattern to be globbed, and either a list of filenames or the next filename in the list is returned, depending on context. This distinction is determined on syntactic grounds alone. That means \f(CW\*(C`<$x>\*(C'\fR is always a \f(CWreadline()\fR from an indirect handle, but \f(CW\*(C`<$hash{key}>\*(C'\fR is always a \f(CWglob()\fR. That's because \f(CW$x\fR is a simple scalar variable, but \f(CW$hash{key}\fR is not\-\-it's a hash element. Even \f(CW\*(C`<$x >\*(C'\fR (note the extra space) is treated as \f(CW\*(C`glob("$x ")\*(C'\fR, not \f(CWreadline($x)\fR. .PP One level of double-quote interpretation is done first, but you can't say \f(CW\*(C`<$foo>\*(C'\fR because that's an indirect filehandle as explained in the previous paragraph. (In older versions of Perl, programmers would insert curly brackets to force interpretation as a filename glob: \&\f(CW\*(C`<${foo}>\*(C'\fR. These days, it's considered cleaner to call the internal function directly as \f(CWglob($foo)\fR, which is probably the right way to have done it in the first place.) For example: .PP .Vb 3 \& while (<*.c>) { \& chmod 0644, $_; \& } .Ve .PP is roughly equivalent to: .PP .Vb 5 \& open(FOO, "echo *.c | tr \-s \*(Aq \et\er\ef\*(Aq \*(Aq\e\e012\e\e012\e\e012\e\e012\*(Aq|"); \& while () { \& chomp; \& chmod 0644, $_; \& } .Ve .PP except that the globbing is actually done internally using the standard \&\f(CW\*(C`File::Glob\*(C'\fR extension. Of course, the shortest way to do the above is: .PP .Vb 1 \& chmod 0644, <*.c>; .Ve .PP A (file)glob evaluates its (embedded) argument only when it is starting a new list. All values must be read before it will start over. In list context, this isn't important because you automatically get them all anyway. However, in scalar context the operator returns the next value each time it's called, or \f(CW\*(C`undef\*(C'\fR when the list has run out. As with filehandle reads, an automatic \f(CW\*(C`defined\*(C'\fR is generated when the glob occurs in the test part of a \f(CW\*(C`while\*(C'\fR, because legal glob returns (for example, a file called \fI0\fR) would otherwise terminate the loop. Again, \f(CW\*(C`undef\*(C'\fR is returned only once. So if you're expecting a single value from a glob, it is much better to say .PP .Vb 1 \& ($file) = ; .Ve .PP than .PP .Vb 1 \& $file = ; .Ve .PP because the latter will alternate between returning a filename and returning false. .PP If you're trying to do variable interpolation, it's definitely better to use the \f(CWglob()\fR function, because the older notation can cause people to become confused with the indirect filehandle notation. .PP .Vb 2 \& @files = glob("$dir/*.[ch]"); \& @files = glob($files[$i]); .Ve .PP If an angle-bracket-based globbing expression is used as the condition of a \f(CW\*(C`while\*(C'\fR or \f(CW\*(C`for\*(C'\fR loop, then it will be implicitly assigned to \f(CW$_\fR. If either a globbing expression or an explicit assignment of a globbing expression to a scalar is used as a \f(CW\*(C`while\*(C'\fR/\f(CW\*(C`for\*(C'\fR condition, then the condition actually tests for definedness of the expression's value, not for its regular truth value. .SS "Constant Folding" .IX Xref "constant folding folding" .IX Subsection "Constant Folding" Like C, Perl does a certain amount of expression evaluation at compile time whenever it determines that all arguments to an operator are static and have no side effects. In particular, string concatenation happens at compile time between literals that don't do variable substitution. Backslash interpolation also happens at compile time. You can say .PP .Vb 3 \& \*(AqNow is the time for all\*(Aq \& . "\en" \& . \*(Aqgood men to come to.\*(Aq .Ve .PP and this all reduces to one string internally. Likewise, if you say .PP .Vb 3 \& foreach $file (@filenames) { \& if (\-s $file > 5 + 100 * 2**16) { } \& } .Ve .PP the compiler precomputes the number which that expression represents so that the interpreter won't have to. .SS No-ops .IX Xref "no-op nop" .IX Subsection "No-ops" Perl doesn't officially have a no-op operator, but the bare constants \&\f(CW0\fR and \f(CW1\fR are special-cased not to produce a warning in void context, so you can for example safely do .PP .Vb 1 \& 1 while foo(); .Ve .SS "Bitwise String Operators" .IX Xref "operator, bitwise, string &. |. ^. ~." .IX Subsection "Bitwise String Operators" Bitstrings of any size may be manipulated by the bitwise operators (\f(CW\*(C`~ | & ^\*(C'\fR). .PP If the operands to a binary bitwise op are strings of different sizes, \fB|\fR and \fB^\fR ops act as though the shorter operand had additional zero bits on the right, while the \fB&\fR op acts as though the longer operand were truncated to the length of the shorter. The granularity for such extension or truncation is one or more bytes. .PP .Vb 5 \& # ASCII\-based examples \& print "j p \en" ^ " a h"; # prints "JAPH\en" \& print "JA" | " ph\en"; # prints "japh\en" \& print "japh\enJunk" & \*(Aq_\|_\|_\|_\|_\*(Aq; # prints "JAPH\en"; \& print \*(Aqp N$\*(Aq ^ " E>\*(C'\fR) always produce integral results. (But see also "Bitwise String Operators".) However, \f(CW\*(C`use\ integer\*(C'\fR still has meaning for them. By default, their results are interpreted as unsigned integers, but if \f(CW\*(C`use\ integer\*(C'\fR is in effect, their results are interpreted as signed integers. For example, \f(CW\*(C`~0\*(C'\fR usually evaluates to a large integral value. However, \f(CW\*(C`use\ integer;\ ~0\*(C'\fR is \f(CW\-1\fR on two's-complement machines. .SS "Floating-point Arithmetic" .IX Subsection "Floating-point Arithmetic" .IX Xref "floating-point floating point float real" .PP While \f(CW\*(C`use\ integer\*(C'\fR provides integer-only arithmetic, there is no analogous mechanism to provide automatic rounding or truncation to a certain number of decimal places. For rounding to a certain number of digits, \f(CWsprintf()\fR or \f(CWprintf()\fR is usually the easiest route. See perlfaq4. .PP Floating-point numbers are only approximations to what a mathematician would call real numbers. There are infinitely more reals than floats, so some corners must be cut. For example: .PP .Vb 2 \& printf "%.20g\en", 123456789123456789; \& # produces 123456789123456784 .Ve .PP Testing for exact floating-point equality or inequality is not a good idea. Here's a (relatively expensive) work-around to compare whether two floating-point numbers are equal to a particular number of decimal places. See Knuth, volume II, for a more robust treatment of this topic. .PP .Vb 7 \& sub fp_equal { \& my ($X, $Y, $POINTS) = @_; \& my ($tX, $tY); \& $tX = sprintf("%.${POINTS}g", $X); \& $tY = sprintf("%.${POINTS}g", $Y); \& return $tX eq $tY; \& } .Ve .PP The POSIX module (part of the standard perl distribution) implements \&\f(CWceil()\fR, \f(CWfloor()\fR, and other mathematical and trigonometric functions. The \f(CW\*(C`Math::Complex\*(C'\fR module (part of the standard perl distribution) defines mathematical functions that work on both the reals and the imaginary numbers. \f(CW\*(C`Math::Complex\*(C'\fR is not as efficient as POSIX, but POSIX can't work with complex numbers. .PP Rounding in financial applications can have serious implications, and the rounding method used should be specified precisely. In these cases, it probably pays not to trust whichever system rounding is being used by Perl, but to instead implement the rounding function you need yourself. .SS "Bigger Numbers" .IX Xref "number, arbitrary precision" .IX Subsection "Bigger Numbers" The standard \f(CW\*(C`Math::BigInt\*(C'\fR, \f(CW\*(C`Math::BigRat\*(C'\fR, and \&\f(CW\*(C`Math::BigFloat\*(C'\fR modules, along with the \f(CW\*(C`bignum\*(C'\fR, \f(CW\*(C`bigint\*(C'\fR, and \f(CW\*(C`bigrat\*(C'\fR pragmas, provide variable-precision arithmetic and overloaded operators, although they're currently pretty slow. At the cost of some space and considerable speed, they avoid the normal pitfalls associated with limited-precision representations. .PP .Vb 5 \& use 5.010; \& use bigint; # easy interface to Math::BigInt \& $x = 123456789123456789; \& say $x * $x; \& +15241578780673678515622620750190521 .Ve .PP Or with rationals: .PP .Vb 8 \& use 5.010; \& use bigrat; \& $x = 3/22; \& $y = 4/6; \& say "x/y is ", $x/$y; \& say "x*y is ", $x*$y; \& x/y is 9/44 \& x*y is 1/11 .Ve .PP Several modules let you calculate with unlimited or fixed precision (bound only by memory and CPU time). There are also some non-standard modules that provide faster implementations via external C libraries. .PP Here is a short, but incomplete summary: .PP .Vb 10 \& Math::String treat string sequences like numbers \& Math::FixedPrecision calculate with a fixed precision \& Math::Currency for currency calculations \& Bit::Vector manipulate bit vectors fast (uses C) \& Math::BigIntFast Bit::Vector wrapper for big numbers \& Math::Pari provides access to the Pari C library \& Math::Cephes uses the external Cephes C library (no \& big numbers) \& Math::Cephes::Fraction fractions via the Cephes library \& Math::GMP another one using an external C library \& Math::GMPz an alternative interface to libgmp\*(Aqs big ints \& Math::GMPq an interface to libgmp\*(Aqs fraction numbers \& Math::GMPf an interface to libgmp\*(Aqs floating point numbers .Ve .PP Choose wisely.