Cari di Perl 
     Perl User Manual
Daftar Isi ☛
(Sebelumnya) Perl pragma to restrict unsafe ...Perl pragma to lexically contr ... (Berikutnya)
Pragmas

Package for overloading Perl operations

Daftar Isi

NAME

overload - Package for overloading Perl operations

SYNOPSIS

  1.  package SomeThing;
    2.  
  2.  
  3.  use overload
    4.  
  4. '+' => \&myadd,
    5.  
  5. '-' => \&mysub;
    6.  
  6. # etc
    7.  
  7.  ...
    8.  
  8.  
  9.  package main;
    10.  
  10.  $a = SomeThing->new( 57 );
    11.  
  11.  $b = 5 + $a;
    12.  
  12.  ...
    13.  
  13.  if (overload::Overloaded $b) {...}
    14.  
  14.  ...
    15.  
  15.  $strval = overload::StrVal $b;
    16.

DESCRIPTION

This pragma allows overloading of Perl's operators for a class.To overload built-in functions, see Overriding Built-in Functions in perlsub instead.

Fundamentals

Declaration

Arguments of the use overload directive are (key, value) pairs.For the full set of legal keys, see Overloadable Operations below.

Operator implementations (the values) can be subroutines,references to subroutines, or anonymous subroutines- in other words, anything legal inside a &{ ... } call.Values specified as strings are interpreted as method names.Thus

  1.  package Number;
    2.  
  2.  use overload
    3.  
  3.  "-" => "minus",
    4.  
  4.  "*=" => \&muas,
    5.  
  5.  '""' => sub { ...; };
    6.

declares that subtraction is to be implemented by method minus()in the class Number (or one of its base classes),and that the function Number::muas() is to be used for theassignment form of multiplication, *=.It also defines an anonymous subroutine to implement stringification:this is called whenever an object blessed into the package Numberis used in a string context (this subroutine might, for example,return the number as a Roman numeral).

Calling Conventions and Magic Autogeneration

The following sample implementation of minus() (which assumesthat Number objects are simply blessed references to scalars)illustrates the calling conventions:

  1.  package Number;
    2.  
  2.  sub minus {
    3.  
  3.  my ($self, $other, $swap) = @_;
    4.  
  4.  my $result = $$self - $other; # *
    5.  
  5.  $result = -$result if $swap;
    6.  
  6.  ref $result ? $result : bless \$result;
    7.  
  7.  }
    8.  
  8.  # * may recurse once - see table below
    9.

Three arguments are passed to all subroutines specified in theuse overload directive (with one exception - see nomethod).The first of these is the operand providing the overloadedoperator implementation -in this case, the object whose minus() method is being called.

The second argument is the other operand, or undef in thecase of a unary operator.

The third argument is set to TRUE if (and only if) the twooperands have been swapped. Perl may do this to ensure that thefirst argument ($self) is an object implementing the overloadedoperation, in line with general object calling conventions.For example, if $x and $y are Numbers:

  1.  operation   |   generates a call to
    2.  
  2.  ============|======================
    3.  
  3.  $x - $y |   minus($x, $y, '')
    4.  
  4.  $x - 7  |   minus($x, 7, '')
    5.  
  5.  7 - $x  |   minus($x, 7, 1)
    6.

Perl may also use minus() to implement other operators whichhave not been specified in the use overload directive,according to the rules for Magic Autogeneration described later.For example, the use overload above declared no subroutinefor any of the operators --, neg (the overload key forunary minus), or -=. Thus

  1.  operation   |   generates a call to
    2.  
  2.  ============|======================
    3.  
  3.  -$x |   minus($x, 0, 1)
    4.  
  4.  $x-- |   minus($x, 1, undef)
    5.  
  5.  $x -= 3 |   minus($x, 3, undef)
    6.

Note the undefs:where autogeneration results in the method for a standardoperator which does not change either of its operands, suchas -, being used to implement an operator which changesthe operand ("mutators": here, -- and -=),Perl passes undef as the third argument.This still evaluates as FALSE, consistent with the fact thatthe operands have not been swapped, but gives the subroutinea chance to alter its behaviour in these cases.

In all the above examples, minus() is requiredonly to return the result of the subtraction:Perl takes care of the assignment to $x.In fact, such methods should not modify their operands,even if undef is passed as the third argument(see Overloadable Operations).

The same is not true of implementations of ++ and --:these are expected to modify their operand.An appropriate implementation of -- might look like

  1.  use overload '--' => "decr",
    2.  
  2.  # ...
    3.  
  3.  sub decr { --${$_[0]}; }
    4.

Mathemagic, Mutators, and Copy Constructors

The term 'mathemagic' describes the overloaded implementationof mathematical operators.Mathemagical operations raise an issue.Consider the code:

  1.  $a = $b;
    2.  
  2.  --$a;
    3.

If $a and $b are scalars then after these statements

  1.  $a == $b - 1
    2.

An object, however, is a reference to blessed data, so if$a and $b are objects then the assignment $a = $bcopies only the reference, leaving $a and $b referringto the same object data.One might therefore expect the operation --$a to decrement$b as well as $a.However, this would not be consistent with how we expect themathematical operators to work.

Perl resolves this dilemma by transparently calling a copyconstructor before calling a method defined to implementa mutator (--, +=, and so on.).In the above example, when Perl reaches the decrementstatement, it makes a copy of the object data in $a andassigns to $a a reference to the copied data.Only then does it call decr(), which alters the copieddata, leaving $b unchanged.Thus the object metaphor is preserved as far as possible,while mathemagical operations still work according to thearithmetic metaphor.

Note: the preceding paragraph describes what happens whenPerl autogenerates the copy constructor for an object basedon a scalar.For other cases, see Copy Constructor.

Overloadable Operations

The complete list of keys that can be specified in the use overloaddirective are given, separated by spaces, in the values of thehash %overload::ops:

  1.  with_assign  => '+ - * / % ** << >> x .',
    2.  
  2.  assign  => '+= -= *= /= %= **= <<= >>= x= .=',
    3.  
  3.  num_comparison  => '< <= > >= == !=',
    4.  
  4.  '3way_comparison'=> '<=> cmp',
    5.  
  5.  str_comparison  => 'lt le gt ge eq ne',
    6.  
  6.  binary  => '& &= | |= ^ ^=',
    7.  
  7.  unary  => 'neg ! ~',
    8.  
  8.  mutators  => '++ --',
    9.  
  9.  func  => 'atan2 cos sin exp abs log sqrt int',
    10.  
  10.  conversion  => 'bool "" 0+ qr',
    11.  
  11.  iterators  => '<>',
    12.  
  12.  filetest => '-X',
    13.  
  13.  dereferencing  => '${} @{} %{} &{} *{}',
    14.  
  14.  matching  => '~~',
    15.  
  15.  special  => 'nomethod fallback ='
    16.

Most of the overloadable operators map one-to-one to these keys.Exceptions, including additional overloadable operations notapparent from this hash, are included in the notes which follow.

A warning is issued if an attempt is made to register an operator not foundabove.

  • not

    The operator not is not a valid key for use overload.However, if the operator ! is overloaded then the sameimplementation will be used for not(since the two operators differ only in precedence).

  • neg

    The key neg is used for unary minus to disambiguate it frombinary -.

  • ++, --

    Assuming they are to behave analogously to Perl's ++ and --,overloaded implementations of these operators are required tomutate their operands.

    No distinction is made between prefix and postfix forms of theincrement and decrement operators: these differ only in thepoint at which Perl calls the associated subroutine whenevaluating an expression.

  • Assignments
    1.  +=  -=  *=  /=  %=  **=  <<=  >>=  x=  .=
      2.  
    2.  &=  |=  ^=
      3.

    Simple assignment is not overloadable (the '=' key is usedfor the Copy Constructor).Perl does have a way to make assignments to an object do whateveryou want, but this involves using tie(), not overload -see tie and the COOKBOOK examples below.

    The subroutine for the assignment variant of an operator isrequired only to return the result of the operation.It is permitted to change the value of its operand(this is safe because Perl calls the copy constructor first),but this is optional since Perl assigns the returned value tothe left-hand operand anyway.

    An object that overloads an assignment operator does so only inrespect of assignments to that object.In other words, Perl never calls the corresponding methods withthe third argument (the "swap" argument) set to TRUE.For example, the operation

    1.  $a *= $b
      2.

    cannot lead to $b's implementation of *= being called,even if $a is a scalar.(It can, however, generate a call to $b's method for *).

  • Non-mutators with a mutator variant
    1.  +  -  *  /  %  **  <<  >>  x  .
      2.  
    2.  &  |  ^
      3.

    As described above,Perl may call methods for operators like + and & in the courseof implementing missing operations like ++, +=, and &=.While these methods may detect this usage by testing the definednessof the third argument, they should in all cases avoid changing theiroperands.This is because Perl does not call the copy constructor beforeinvoking these methods.

  • int

    Traditionally, the Perl function int rounds to 0(see int), and so for floating-point-like types oneshould follow the same semantic.

  • String, numeric, boolean, and regexp conversions
    1.  ""  0+  bool
      2.

    These conversions are invoked according to context as necessary.For example, the subroutine for '""' (stringify) may be usedwhere the overloaded object is passed as an argument to print,and that for 'bool' where it is tested in the condition of a flowcontrol statement (like while) or the ternary ?: operation.

    Of course, in contexts like, for example, $obj + 1, Perl willinvoke $obj's implementation of + rather than (in thisexample) converting $obj to a number using the numify method'0+' (an exception to this is when no method has been providedfor '+' and fallback is set to TRUE).

    The subroutines for '""', '0+', and 'bool' can returnany arbitrary Perl value.If the corresponding operation for this value is overloaded too,the operation will be called again with this value.

    As a special case if the overload returns the object itself then it willbe used directly. An overloaded conversion returning the object isprobably a bug, because you're likely to get something that looks likeYourPackage=HASH(0x8172b34).

    1.  qr
      2.

    The subroutine for 'qr' is used wherever the object isinterpolated into or used as a regexp, including when itappears on the RHS of a =~ or !~ operator.

    qr must return a compiled regexp, or a ref to a compiled regexp(such as qr// returns), and any further overloading on the returnvalue will be ignored.

  • Iteration

    If <> is overloaded then the same implementation is usedfor both the read-filehandle syntax <$var> andglobbing syntax <${var}>.

    BUGS Even in list context, the iterator is currently called onlyonce and with scalar context.

  • File tests

    The key '-X' is used to specify a subroutine to handle all thefiletest operators (-f, -x, and so on: see -X forthe full list);it is not possible to overload any filetest operator individually.To distinguish them, the letter following the '-' is passed as thesecond argument (that is, in the slot that for binary operatorsis used to pass the second operand).

    Calling an overloaded filetest operator does not affect the stat valueassociated with the special filehandle _. It still refers to theresult of the last stat, lstat or unoverloaded filetest.

    This overload was introduced in Perl 5.12.

  • Matching

    The key "~~" allows you to override the smart matching logic used bythe ~~ operator and the switch construct (given/when). SeeSwitch Statements in perlsyn and feature.

    Unusually, the overloaded implementation of the smart match operatordoes not get full control of the smart match behaviour.In particular, in the following code:

    1.  package Foo;
      2.  
    2.  use overload '~~' => 'match';
      3.  
    3.  
    4.  my $obj =  Foo->new();
      5.  
    5.  $obj ~~ [ 1,2,3 ];
      6.

    the smart match does not invoke the method call like this:

    1.  $obj->match([1,2,3],0);
      2.

    rather, the smart match distributive rule takes precedence, so $obj issmart matched against each array element in turn until a match is found,so you may see between one and three of these calls instead:

    1.  $obj->match(1,0);
      2.  
    2.  $obj->match(2,0);
      3.  
    3.  $obj->match(3,0);
      4.

    Consult the match table in Smartmatch Operator in perlop fordetails of when overloading is invoked.

  • Dereferencing
    1.  ${}  @{}  %{}  &{}  *{}
      2.

    If these operators are not explicitly overloaded then theywork in the normal way, yielding the underlying scalar,array, or whatever stores the object data (or the appropriateerror message if the dereference operator doesn't match it).Defining a catch-all 'nomethod' (see below)makes no difference to this as the catch-all function willnot be called to implement a missing dereference operator.

    If a dereference operator is overloaded then it must return areference of the appropriate type (for example, thesubroutine for key '${}' should return a reference to ascalar, not a scalar), or another object which overloads theoperator: that is, the subroutine only determines what isdereferenced and the actual dereferencing is left to Perl.As a special case, if the subroutine returns the object itselfthen it will not be called again - avoiding infinite recursion.

  • Special
    1.  nomethod  fallback  =
      2.

    See Special Keys for use overload.

Magic Autogeneration

If a method for an operation is not found then Perl tries toautogenerate a substitute implementation from the operationsthat have been defined.

Note: the behaviour described in this section can be disabledby setting fallback to FALSE (see fallback).

In the following tables, numbers indicate priority.For example, the table below states that,if no implementation for '!' has been defined then Perl willimplement it using 'bool' (that is, by inverting the valuereturned by the method for 'bool');if boolean conversion is also unimplemented then Perl willuse '0+' or, failing that, '""'.

  1.  operator | can be autogenerated from
    2.  
  2.  |
    3.  
  3.  | 0+   ""   bool   .   x
    4.  
  4.  =========|==========================
    5.  
  5.    0+ |   1 2
    6.  
  6.    "" |  1  2
    7.  
  7.    bool  |  1 2
    8.  
  8.    int   |  1 2 3
    9.  
  9.    ! |  2 3 1
    10.  
  10.    qr |  2 1 3
    11.  
  11.       . |  2 1 3
    12.  
  12.    x |  2 1 3
    13.  
  13.    .= |  3 2 4 1
    14.  
  14.    x= |  3 2 4 1
    15.  
  15.    <> |  2 1 3
    16.  
  16.    -X |  2 1 3
    17.

Note: The iterator ('<>') and file test ('-X')operators work as normal: if the operand is not a blessed glob orIO reference then it is converted to a string (using the methodfor '""', '0+', or 'bool') to be interpreted as a globor filename.

  1.  operator | can be autogenerated from
    2.  
  2.  |
    3.  
  3.  |  <   <=>   neg   -= -
    4.  
  4.  =========|==========================
    5.  
  5.    neg   | 1
    6.  
  6.    -= | 1
    7.  
  7.    -- |   1 2
    8.  
  8.    abs   | a1 a2 b1 b2 [*]
    9.  
  9.    < | 1
    10.  
  10.    <= | 1
    11.  
  11.    > | 1
    12.  
  12.    >= | 1
    13.  
  13.    == | 1
    14.  
  14.    != | 1
    15.  
  15.  
  16.  * one from [a1, a2] and one from [b1, b2]
    17.

Just as numeric comparisons can be autogenerated from the methodfor '<=>', string comparisons can be autogenerated fromthat for 'cmp':

  1.  operators  |  can be autogenerated from
    2.  
  2.  ====================|===========================
    3.  
  3.  lt gt le ge eq ne  |  cmp
    4.

Similarly, autogeneration for keys '+=' and '++' is analogousto '-=' and '--' above:

  1.  operator | can be autogenerated from
    2.  
  2.  |
    3.  
  3.  |  += +
    4.  
  4.  =========|==========================
    5.  
  5.  +=   | 1
    6.  
  6.  ++   |   1 2
    7.

And other assignment variations are analogous to'+=' and '-=' (and similar to '.=' and 'x=' above):

  1.   operator ||  *= /= %= **= <<= >>= &= ^= |=
    2.  
  2.  -------------------||--------------------------------
    3.  
  3.  autogenerated from ||  *  /  %  **  <<  >>  &  ^  |
    4.

Note also that the copy constructor (key '=') may beautogenerated, but only for objects based on scalars.See Copy Constructor.

Minimal Set of Overloaded Operations

Since some operations can be automatically generated from others, there isa minimal set of operations that need to be overloaded in order to havethe complete set of overloaded operations at one's disposal.Of course, the autogenerated operations may not do exactly what the userexpects. The minimal set is:

  1.  + - * / % ** << >> x
    2.  
  2.  <=> cmp
    3.  
  3.  & | ^ ~
    4.  
  4.  atan2 cos sin exp log sqrt int
    5.  
  5.  "" 0+ bool
    6.  
  6.  ~~
    7.

Of the conversions, only one of string, boolean or numeric isneeded because each can be generated from either of the other two.

Special Keys for use overload

nomethod

The 'nomethod' key is used to specify a catch-all function tobe called for any operator that is not individually overloaded.The specified function will be passed four parameters.The first three arguments coincide with those that would have beenpassed to the corresponding method if it had been defined.The fourth argument is the use overload key for that missingmethod.

For example, if $a is an object blessed into a package declaring

  1.  use overload 'nomethod' => 'catch_all', # ...
    2.

then the operation

  1.  3 + $a
    2.

could (unless a method is specifically declared for the key'+') result in a call

  1.  catch_all($a, 3, 1, '+')
    2.

See How Perl Chooses an Operator Implementation.

fallback

The value assigned to the key 'fallback' tells Perl how hardit should try to find an alternative way to implement a missingoperator.

  • defined, but FALSE
    1.  use overload "fallback" => 0, # ... ;
      2.

    This disables Magic Autogeneration.

  • undef

    In the default case where no value is explicitly assigned tofallback, magic autogeneration is enabled.

  • TRUE

    The same as for undef, but if a missing operator cannot beautogenerated then, instead of issuing an error message, Perlis allowed to revert to what it would have done for thatoperator if there had been no use overload directive.

    Note: in most cases, particularly the Copy Constructor,this is unlikely to be appropriate behaviour.

See How Perl Chooses an Operator Implementation.

Copy Constructor

As mentioned above,this operation is called when a mutator is applied to a referencethat shares its object with some other reference.For example, if $b is mathemagical, and '++' is overloadedwith 'incr', and '=' is overloaded with 'clone', then thecode

  1.  $a = $b;
    2.  
  2.  # ... (other code which does not modify $a or $b) ...
    3.  
  3.  ++$b;
    4.

would be executed in a manner equivalent to

  1.  $a = $b;
    2.  
  2.  # ...
    3.  
  3.  $b = $b->clone(undef, "");
    4.  
  4.  $b->incr(undef, "");
    5.

Note:

  • The subroutine for '=' does not overload the Perl assignmentoperator: it is used only to allow mutators to work as describedhere. (See Assignments above.)

  • As for other operations, the subroutine implementing '=' is passedthree arguments, though the last two are always undef and ''.

  • The copy constructor is called only before a call to a functiondeclared to implement a mutator, for example, if ++$b; in thecode above is effected via a method declared for key '++'(or 'nomethod', passed '++' as the fourth argument) or, byautogeneration, '+='.It is not called if the increment operation is effected by a callto the method for '+' since, in the equivalent code,

    1.  $a = $b;
      2.  
    2.  $b = $b + 1;
      3.

    the data referred to by $a is unchanged by the assignment to$b of a reference to new object data.

  • The copy constructor is not called if Perl determines that it isunnecessary because there is no other reference to the data beingmodified.

  • If 'fallback' is undefined or TRUE then a copy constructorcan be autogenerated, but only for objects based on scalars.In other cases it needs to be defined explicitly.Where an object's data is stored as, for example, an array ofscalars, the following might be appropriate:

    1.  use overload '=' => sub { bless [ @{$_[0]} ] },  # ...
      2.

  • If 'fallback' is TRUE and no copy constructor is defined then,for objects not based on scalars, Perl may silently fall back onsimple assignment - that is, assignment of the object reference.In effect, this disables the copy constructor mechanism sinceno new copy of the object data is created.This is almost certainly not what you want.(It is, however, consistent: for example, Perl's fallback for the++ operator is to increment the reference itself.)

How Perl Chooses an Operator Implementation

Which is checked first, nomethod or fallback?If the two operands of an operator are of different types andboth overload the operator, which implementation is used?The following are the precedence rules:

1.

If the first operand has declared a subroutine to overload theoperator then use that implementation.

2.

Otherwise, if fallback is TRUE or undefined for thefirst operand then see if therules for autogenerationallows another of its operators to be used instead.

3.

Unless the operator is an assignment (+=, -=, etc.),repeat step (1) in respect of the second operand.

4.

Repeat Step (2) in respect of the second operand.

5.

If the first operand has a "nomethod" method then use that.

6.

If the second operand has a "nomethod" method then use that.

7.

If fallback is TRUE for both operandsthen perform the usual operation for the operator,treating the operands as numbers, strings, or booleansas appropriate for the operator (see note).

8.

Nothing worked - die.

Where there is only one operand (or only one operand withoverloading) the checks in respect of the other operand above areskipped.

There are exceptions to the above rules for dereference operations(which, if Step 1 fails, always fall back to the normal, built-inimplementations - see Dereferencing), and for ~~ (which has itsown set of rules - see Matching under Overloadable Operationsabove).

Note on Step 7: some operators have a different semantic dependingon the type of their operands.As there is no way to instruct Perl to treat the operands as, e.g.,numbers instead of strings, the result here may not be what youexpect.See BUGS AND PITFALLS.

Losing Overloading

The restriction for the comparison operation is that even if, for example,cmp should return a blessed reference, the autogenerated ltfunction will produce only a standard logical value based on thenumerical value of the result of cmp. In particular, a workingnumeric conversion is needed in this case (possibly expressed in terms ofother conversions).

Similarly, .= and x= operators lose their mathemagical propertiesif the string conversion substitution is applied.

When you chop() a mathemagical object it is promoted to a string and itsmathemagical properties are lost. The same can happen with otheroperations as well.

Inheritance and Overloading

Overloading respects inheritance via the @ISA hierarchy.Inheritance interacts with overloading in two ways.

  • Method names in the use overload directive

    If value in

    1.   use overload key => value;
      2.

    is a string, it is interpreted as a method name - which may(in the usual way) be inherited from another class.

  • Overloading of an operation is inherited by derived classes

    Any class derived from an overloaded class is also overloadedand inherits its operator implementations.If the same operator is overloaded in more than one ancestorthen the implementation is determined by the usual inheritancerules.

    For example, if A inherits from B and C (in that order),B overloads + with \&D::plus_sub, and C overloads+ by "plus_meth", then the subroutine D::plus_sub willbe called to implement operation + for an object in package A.

Note that since the value of the fallback key is not a subroutine,its inheritance is not governed by the above rules. In the currentimplementation, the value of fallback in the first overloadedancestor is used, but this is accidental and subject to change.

Run-time Overloading

Since all use directives are executed at compile-time, the only way tochange overloading during run-time is to

  1.  eval 'use overload "+" => \&addmethod';
    2.

You can also use

  1.  eval 'no overload "+", "--", "<="';
    2.

though the use of these constructs during run-time is questionable.

Public Functions

Package overload.pm provides the following public functions:

  • overload::StrVal(arg)

    Gives the string value of arg as in theabsence of stringify overloading. If youare using this to get the address of a reference (useful for checking if tworeferences point to the same thing) then you may be better off usingScalar::Util::refaddr(), which is faster.

  • overload::Overloaded(arg)

    Returns true if arg is subject to overloading of some operations.

  • overload::Method(obj,op)

    Returns undef or a reference to the method that implements op.

Overloading Constants

For some applications, the Perl parser mangles constants too much.It is possible to hook into this process via overload::constant()and overload::remove_constant() functions.

These functions take a hash as an argument. The recognized keys of this hashare:

  • integer

    to overload integer constants,

  • float

    to overload floating point constants,

  • binary

    to overload octal and hexadecimal constants,

  • q

    to overload q-quoted strings, constant pieces of qq- and qx-quotedstrings and here-documents,

  • qr

    to overload constant pieces of regular expressions.

The corresponding values are references to functions which take three arguments:the first one is the initial string form of the constant, the second oneis how Perl interprets this constant, the third one is how the constant is used.Note that the initial string form does notcontain string delimiters, and has backslashes in backslash-delimitercombinations stripped (thus the value of delimiter is not relevant forprocessing of this string). The return value of this function is how thisconstant is going to be interpreted by Perl. The third argument is undefinedunless for overloaded q- and qr- constants, it is q in single-quotecontext (comes from strings, regular expressions, and single-quote HEREdocuments), it is tr for arguments of tr/y operators,it is s for right-hand side of s-operator, and it is qq otherwise.

Since an expression "ab$cd,," is just a shortcut for 'ab' . $cd . ',,',it is expected that overloaded constant strings are equipped with reasonableoverloaded catenation operator, otherwise absurd results will result.Similarly, negative numbers are considered as negations of positive constants.

Note that it is probably meaningless to call the functions overload::constant()and overload::remove_constant() from anywhere but import() and unimport() methods.From these methods they may be called as

  1.  sub import {
    2.  
  2.    shift;
    3.  
  3.    return unless @_;
    4.  
  4.    die "unknown import: @_" unless @_ == 1 and $_[0] eq ':constant';
    5.  
  5.    overload::constant integer => sub {Math::BigInt->new(shift)};
    6.  
  6.  }
    7.

IMPLEMENTATION

What follows is subject to change RSN.

The table of methods for all operations is cached in magic for thesymbol table hash for the package. The cache is invalidated duringprocessing of use overload, no overload, new functiondefinitions, and changes in @ISA. However, this invalidation remainsunprocessed until the next blessing into the package. Hence if youwant to change overloading structure dynamically, you'll need anadditional (fake) blessing to update the table.

(Every SVish thing has a magic queue, and magic is an entry in thatqueue. This is how a single variable may participate in multipleforms of magic simultaneously. For instance, environment variablesregularly have two forms at once: their %ENV magic and their taintmagic. However, the magic which implements overloading is applied tothe stashes, which are rarely used directly, thus should not slow downPerl.)

If an object belongs to a package using overload, it carries a specialflag. Thus the only speed penalty during arithmetic operations withoutoverloading is the checking of this flag.

In fact, if use overload is not present, there is almost no overheadfor overloadable operations, so most programs should not suffermeasurable performance penalties. A considerable effort was made tominimize the overhead when overload is used in some package, but thearguments in question do not belong to packages using overload. Whenin doubt, test your speed with use overload and without it. So farthere have been no reports of substantial speed degradation if Perl iscompiled with optimization turned on.

There is no size penalty for data if overload is not used. The onlysize penalty if overload is used in some package is that all thepackages acquire a magic during the next blessing into thepackage. This magic is three-words-long for packages withoutoverloading, and carries the cache table if the package is overloaded.

It is expected that arguments to methods that are not explicitly supposedto be changed are constant (but this is not enforced).

COOKBOOK

Please add examples to what follows!

Two-face Scalars

Put this in two_face.pm in your Perl library directory:

  1.   package two_face;# Scalars with separate string and
    2.  
  2.  # numeric values.
    3.  
  3.   sub new { my $p = shift; bless [@_], $p }
    4.  
  4.   use overload '""' => \&str, '0+' => \&num, fallback => 1;
    5.  
  5.   sub num {shift->[1]}
    6.  
  6.   sub str {shift->[0]}
    7.

Use it as follows:

  1.   require two_face;
    2.  
  2.   my $seven = two_face->new("vii", 7);
    3.  
  3.   printf "seven=$seven, seven=%d, eight=%d\n", $seven, $seven+1;
    4.  
  4.   print "seven contains 'i'\n" if $seven =~ /i/;
    5.

(The second line creates a scalar which has both a string value, and anumeric value.) This prints:

  1.   seven=vii, seven=7, eight=8
    2.  
  2.   seven contains 'i'
    3.

Two-face References

Suppose you want to create an object which is accessible as both anarray reference and a hash reference.

  1.   package two_refs;
    2.  
  2.   use overload '%{}' => \&gethash, '@{}' => sub { $ {shift()} };
    3.  
  3.   sub new {
    4.  
  4.  my $p = shift;
    5.  
  5.  bless \ [@_], $p;
    6.  
  6.   }
    7.  
  7.   sub gethash {
    8.  
  8.  my %h;
    9.  
  9.  my $self = shift;
    10.  
  10.  tie %h, ref $self, $self;
    11.  
  11.  \%h;
    12.  
  12.   }
    13.  
  13.  
  14.   sub TIEHASH { my $p = shift; bless \ shift, $p }
    15.  
  15.   my %fields;
    16.  
  16.   my $i = 0;
    17.  
  17.   $fields{$_} = $i++ foreach qw{zero one two three};
    18.  
  18.   sub STORE {
    19.  
  19.  my $self = ${shift()};
    20.  
  20.  my $key = $fields{shift()};
    21.  
  21.  defined $key or die "Out of band access";
    22.  
  22.  $$self->[$key] = shift;
    23.  
  23.   }
    24.  
  24.   sub FETCH {
    25.  
  25.  my $self = ${shift()};
    26.  
  26.  my $key = $fields{shift()};
    27.  
  27.  defined $key or die "Out of band access";
    28.  
  28.  $$self->[$key];
    29.  
  29.   }
    30.

Now one can access an object using both the array and hash syntax:

  1.   my $bar = two_refs->new(3,4,5,6);
    2.  
  2.   $bar->[2] = 11;
    3.  
  3.   $bar->{two} == 11 or die 'bad hash fetch';
    4.

Note several important features of this example. First of all, theactual type of $bar is a scalar reference, and we do not overloadthe scalar dereference. Thus we can get the actual non-overloadedcontents of $bar by just using $$bar (what we do in functions whichoverload dereference). Similarly, the object returned by theTIEHASH() method is a scalar reference.

Second, we create a new tied hash each time the hash syntax is used.This allows us not to worry about a possibility of a reference loop,which would lead to a memory leak.

Both these problems can be cured. Say, if we want to overload hashdereference on a reference to an object which is implemented as ahash itself, the only problem one has to circumvent is how to accessthis actual hash (as opposed to the virtual hash exhibited by theoverloaded dereference operator). Here is one possible fetching routine:

  1.   sub access_hash {
    2.  
  2.  my ($self, $key) = (shift, shift);
    3.  
  3.  my $class = ref $self;
    4.  
  4.  bless $self, 'overload::dummy'; # Disable overloading of %{}
    5.  
  5.  my $out = $self->{$key};
    6.  
  6.  bless $self, $class;# Restore overloading
    7.  
  7.  $out;
    8.  
  8.   }
    9.

To remove creation of the tied hash on each access, one may an extralevel of indirection which allows a non-circular structure of references:

  1.   package two_refs1;
    2.  
  2.   use overload '%{}' => sub { ${shift()}->[1] },
    3.  
  3.    '@{}' => sub { ${shift()}->[0] };
    4.  
  4.   sub new {
    5.  
  5.  my $p = shift;
    6.  
  6.  my $a = [@_];
    7.  
  7.  my %h;
    8.  
  8.  tie %h, $p, $a;
    9.  
  9.  bless \ [$a, \%h], $p;
    10.  
  10.   }
    11.  
  11.   sub gethash {
    12.  
  12.  my %h;
    13.  
  13.  my $self = shift;
    14.  
  14.  tie %h, ref $self, $self;
    15.  
  15.  \%h;
    16.  
  16.   }
    17.  
  17.  
  18.   sub TIEHASH { my $p = shift; bless \ shift, $p }
    19.  
  19.   my %fields;
    20.  
  20.   my $i = 0;
    21.  
  21.   $fields{$_} = $i++ foreach qw{zero one two three};
    22.  
  22.   sub STORE {
    23.  
  23.  my $a = ${shift()};
    24.  
  24.  my $key = $fields{shift()};
    25.  
  25.  defined $key or die "Out of band access";
    26.  
  26.  $a->[$key] = shift;
    27.  
  27.   }
    28.  
  28.   sub FETCH {
    29.  
  29.  my $a = ${shift()};
    30.  
  30.  my $key = $fields{shift()};
    31.  
  31.  defined $key or die "Out of band access";
    32.  
  32.  $a->[$key];
    33.  
  33.   }
    34.

Now if $baz is overloaded like this, then $baz is a reference to areference to the intermediate array, which keeps a reference to anactual array, and the access hash. The tie()ing object for the accesshash is a reference to a reference to the actual array, so

  • There are no loops of references.

  • Both "objects" which are blessed into the class two_refs1 arereferences to a reference to an array, thus references to a scalar.Thus the accessor expression $$foo->[$ind] involves nooverloaded operations.

Symbolic Calculator

Put this in symbolic.pm in your Perl library directory:

  1.   package symbolic;# Primitive symbolic calculator
    2.  
  2.   use overload nomethod => \&wrap;
    3.  
  3.  
  4.   sub new { shift; bless ['n', @_] }
    5.  
  5.   sub wrap {
    6.  
  6.  my ($obj, $other, $inv, $meth) = @_;
    7.  
  7.  ($obj, $other) = ($other, $obj) if $inv;
    8.  
  8.  bless [$meth, $obj, $other];
    9.  
  9.   }
    10.

This module is very unusual as overloaded modules go: it does notprovide any usual overloaded operators, instead it provides animplementation for nomethod. In this example the nomethodsubroutine returns an object which encapsulates operations done overthe objects: symbolic->new(3) contains ['n', 3], 2 +symbolic->new(3) contains ['+', 2, ['n', 3]].

Here is an example of the script which "calculates" the side ofcircumscribed octagon using the above package:

  1.   require symbolic;
    2.  
  2.   my $iter = 1;# 2**($iter+2) = 8
    3.  
  3.   my $side = symbolic->new(1);
    4.  
  4.   my $cnt = $iter;
    5.  
  5.  
  6.   while ($cnt--) {
    7.  
  7.  $side = (sqrt(1 + $side**2) - 1)/$side;
    8.  
  8.   }
    9.  
  9.   print "OK\n";
    10.

The value of $side is

  1.   ['/', ['-', ['sqrt', ['+', 1, ['**', ['n', 1], 2]],
    2.  
  2.    undef], 1], ['n', 1]]
    3.

Note that while we obtained this value using a nice little script,there is no simple way to use this value. In fact this value maybe inspected in debugger (see perldebug), but only ifbareStringify Option is set, and not via p command.

If one attempts to print this value, then the overloaded operator"" will be called, which will call nomethod operator. Theresult of this operator will be stringified again, but this result isagain of type symbolic, which will lead to an infinite loop.

Add a pretty-printer method to the module symbolic.pm:

  1.   sub pretty {
    2.  
  2.  my ($meth, $a, $b) = @{+shift};
    3.  
  3.  $a = 'u' unless defined $a;
    4.  
  4.  $b = 'u' unless defined $b;
    5.  
  5.  $a = $a->pretty if ref $a;
    6.  
  6.  $b = $b->pretty if ref $b;
    7.  
  7.  "[$meth $a $b]";
    8.  
  8.   }
    9.

Now one can finish the script by

  1.   print "side = ", $side->pretty, "\n";
    2.

The method pretty is doing object-to-string conversion, so itis natural to overload the operator "" using this method. However,inside such a method it is not necessary to pretty-print thecomponents $a and $b of an object. In the above subroutine"[$meth $a $b]" is a catenation of some strings and components $aand $b. If these components use overloading, the catenation operatorwill look for an overloaded operator .; if not present, it willlook for an overloaded operator "". Thus it is enough to use

  1.   use overload nomethod => \&wrap, '""' => \&str;
    2.  
  2.   sub str {
    3.  
  3.  my ($meth, $a, $b) = @{+shift};
    4.  
  4.  $a = 'u' unless defined $a;
    5.  
  5.  $b = 'u' unless defined $b;
    6.  
  6.  "[$meth $a $b]";
    7.  
  7.   }
    8.

Now one can change the last line of the script to

  1.   print "side = $side\n";
    2.

which outputs

  1.   side = [/ [- [sqrt [+ 1 [** [n 1 u] 2]] u] 1] [n 1 u]]
    2.

and one can inspect the value in debugger using all the possiblemethods.

Something is still amiss: consider the loop variable $cnt of thescript. It was a number, not an object. We cannot make this value oftype symbolic, since then the loop will not terminate.

Indeed, to terminate the cycle, the $cnt should become false.However, the operator bool for checking falsity is overloaded (thistime via overloaded ""), and returns a long string, thus any objectof type symbolic is true. To overcome this, we need a way tocompare an object to 0. In fact, it is easier to write a numericconversion routine.

Here is the text of symbolic.pm with such a routine added (andslightly modified str()):

  1.   package symbolic;# Primitive symbolic calculator
    2.  
  2.   use overload
    3.  
  3.  nomethod => \&wrap, '""' => \&str, '0+' => \&num;
    4.  
  4.  
  5.   sub new { shift; bless ['n', @_] }
    6.  
  6.   sub wrap {
    7.  
  7.  my ($obj, $other, $inv, $meth) = @_;
    8.  
  8.  ($obj, $other) = ($other, $obj) if $inv;
    9.  
  9.  bless [$meth, $obj, $other];
    10.  
  10.   }
    11.  
  11.   sub str {
    12.  
  12.  my ($meth, $a, $b) = @{+shift};
    13.  
  13.  $a = 'u' unless defined $a;
    14.  
  14.  if (defined $b) {
    15.  
  15.   "[$meth $a $b]";
    16.  
  16.  } else {
    17.  
  17.   "[$meth $a]";
    18.  
  18.  }
    19.  
  19.   }
    20.  
  20.   my %subr = ( n => sub {$_[0]},
    21.  
  21.    sqrt => sub {sqrt $_[0]},
    22.  
  22.    '-' => sub {shift() - shift()},
    23.  
  23.    '+' => sub {shift() + shift()},
    24.  
  24.    '/' => sub {shift() / shift()},
    25.  
  25.    '*' => sub {shift() * shift()},
    26.  
  26.    '**' => sub {shift() ** shift()},
    27.  
  27.  );
    28.  
  28.   sub num {
    29.  
  29.  my ($meth, $a, $b) = @{+shift};
    30.  
  30.  my $subr = $subr{$meth}
    31.  
  31.   or die "Do not know how to ($meth) in symbolic";
    32.  
  32.  $a = $a->num if ref $a eq __PACKAGE__;
    33.  
  33.  $b = $b->num if ref $b eq __PACKAGE__;
    34.  
  34.  $subr->($a,$b);
    35.  
  35.   }
    36.

All the work of numeric conversion is done in %subr and num(). Ofcourse, %subr is not complete, it contains only operators used in theexample below. Here is the extra-credit question: why do we need anexplicit recursion in num()? (Answer is at the end of this section.)

Use this module like this:

  1.   require symbolic;
    2.  
  2.   my $iter = symbolic->new(2);# 16-gon
    3.  
  3.   my $side = symbolic->new(1);
    4.  
  4.   my $cnt = $iter;
    5.  
  5.  
  6.   while ($cnt) {
    7.  
  7.  $cnt = $cnt - 1;# Mutator '--' not implemented
    8.  
  8.  $side = (sqrt(1 + $side**2) - 1)/$side;
    9.  
  9.   }
    10.  
  10.   printf "%s=%f\n", $side, $side;
    11.  
  11.   printf "pi=%f\n", $side*(2**($iter+2));
    12.

It prints (without so many line breaks)

  1.   [/ [- [sqrt [+ 1 [** [/ [- [sqrt [+ 1 [** [n 1] 2]]] 1]
    2.  
  2.   [n 1]] 2]]] 1]
    3.  
  3.  [/ [- [sqrt [+ 1 [** [n 1] 2]]] 1] [n 1]]]=0.198912
    4.  
  4.   pi=3.182598
    5.

The above module is very primitive. It does not implementmutator methods (++, -= and so on), does not do deep copying(not required without mutators!), and implements only those arithmeticoperations which are used in the example.

To implement most arithmetic operations is easy; one should just usethe tables of operations, and change the code which fills %subr to

  1.   my %subr = ( 'n' => sub {$_[0]} );
    2.  
  2.   foreach my $op (split " ", $overload::ops{with_assign}) {
    3.  
  3.  $subr{$op} = $subr{"$op="} = eval "sub {shift() $op shift()}";
    4.  
  4.   }
    5.  
  5.   my @bins = qw(binary 3way_comparison num_comparison str_comparison);
    6.  
  6.   foreach my $op (split " ", "@overload::ops{ @bins }") {
    7.  
  7.  $subr{$op} = eval "sub {shift() $op shift()}";
    8.  
  8.   }
    9.  
  9.   foreach my $op (split " ", "@overload::ops{qw(unary func)}") {
    10.  
  10.  print "defining '$op'\n";
    11.  
  11.  $subr{$op} = eval "sub {$op shift()}";
    12.  
  12.   }
    13.

Since subroutines implementing assignment operators are not requiredto modify their operands (see Overloadable Operations above),we do not need anything special to make += and friends work,besides adding these operators to %subr and defining a copyconstructor (needed since Perl has no way to know that theimplementation of '+=' does not mutate the argument -see Copy Constructor).

To implement a copy constructor, add '=' => \&cpy to use overloadline, and code (this code assumes that mutators change things one leveldeep only, so recursive copying is not needed):

  1.   sub cpy {
    2.  
  2.  my $self = shift;
    3.  
  3.  bless [@$self], ref $self;
    4.  
  4.   }
    5.

To make ++ and -- work, we need to implement actual mutators,either directly, or in nomethod. We continue to do things insidenomethod, thus add

  1.  if ($meth eq '++' or $meth eq '--') {
    2.  
  2.   @$obj = ($meth, (bless [@$obj]), 1); # Avoid circular reference
    3.  
  3.   return $obj;
    4.  
  4.  }
    5.

after the first line of wrap(). This is not a most effectiveimplementation, one may consider

  1.   sub inc { $_[0] = bless ['++', shift, 1]; }
    2.

instead.

As a final remark, note that one can fill %subr by

  1.   my %subr = ( 'n' => sub {$_[0]} );
    2.  
  2.   foreach my $op (split " ", $overload::ops{with_assign}) {
    3.  
  3.  $subr{$op} = $subr{"$op="} = eval "sub {shift() $op shift()}";
    4.  
  4.   }
    5.  
  5.   my @bins = qw(binary 3way_comparison num_comparison str_comparison);
    6.  
  6.   foreach my $op (split " ", "@overload::ops{ @bins }") {
    7.  
  7.  $subr{$op} = eval "sub {shift() $op shift()}";
    8.  
  8.   }
    9.  
  9.   foreach my $op (split " ", "@overload::ops{qw(unary func)}") {
    10.  
  10.  $subr{$op} = eval "sub {$op shift()}";
    11.  
  11.   }
    12.  
  12.   $subr{'++'} = $subr{'+'};
    13.  
  13.   $subr{'--'} = $subr{'-'};
    14.

This finishes implementation of a primitive symbolic calculator in50 lines of Perl code. Since the numeric values of subexpressionsare not cached, the calculator is very slow.

Here is the answer for the exercise: In the case of str(), we need noexplicit recursion since the overloaded .-operator will fall backto an existing overloaded operator "". Overloaded arithmeticoperators do not fall back to numeric conversion if fallback isnot explicitly requested. Thus without an explicit recursion num()would convert ['+', $a, $b] to $a + $b, which would just rebuildthe argument of num().

If you wonder why defaults for conversion are different for str() andnum(), note how easy it was to write the symbolic calculator. Thissimplicity is due to an appropriate choice of defaults. One extranote: due to the explicit recursion num() is more fragile than sym():we need to explicitly check for the type of $a and $b. If components$a and $b happen to be of some related type, this may lead to problems.

Really Symbolic Calculator

One may wonder why we call the above calculator symbolic. The reasonis that the actual calculation of the value of expression is postponeduntil the value is used.

To see it in action, add a method

  1.   sub STORE {
    2.  
  2.  my $obj = shift;
    3.  
  3.  $#$obj = 1;
    4.  
  4.  @$obj->[0,1] = ('=', shift);
    5.  
  5.   }
    6.

to the package symbolic. After this change one can do

  1.   my $a = symbolic->new(3);
    2.  
  2.   my $b = symbolic->new(4);
    3.  
  3.   my $c = sqrt($a**2 + $b**2);
    4.

and the numeric value of $c becomes 5. However, after calling

  1.   $a->STORE(12);  $b->STORE(5);
    2.

the numeric value of $c becomes 13. There is no doubt now that the modulesymbolic provides a symbolic calculator indeed.

To hide the rough edges under the hood, provide a tie()d interface to thepackage symbolic. Add methods

  1.   sub TIESCALAR { my $pack = shift; $pack->new(@_) }
    2.  
  2.   sub FETCH { shift }
    3.  
  3.   sub nop {  }# Around a bug
    4.

(the bug, fixed in Perl 5.14, is described in BUGS). One can use thisnew interface as

  1.   tie $a, 'symbolic', 3;
    2.  
  2.   tie $b, 'symbolic', 4;
    3.  
  3.   $a->nop;  $b->nop;# Around a bug
    4.  
  4.  
  5.   my $c = sqrt($a**2 + $b**2);
    6.

Now numeric value of $c is 5. After $a = 12; $b = 5 the numeric valueof $c becomes 13. To insulate the user of the module add a method

  1.   sub vars { my $p = shift; tie($_, $p), $_->nop foreach @_; }
    2.

Now

  1.   my ($a, $b);
    2.  
  2.   symbolic->vars($a, $b);
    3.  
  3.   my $c = sqrt($a**2 + $b**2);
    4.  
  4.  
  5.   $a = 3; $b = 4;
    6.  
  6.   printf "c5  %s=%f\n", $c, $c;
    7.  
  7.  
  8.   $a = 12; $b = 5;
    9.  
  9.   printf "c13  %s=%f\n", $c, $c;
    10.

shows that the numeric value of $c follows changes to the values of $aand $b.

AUTHOR

Ilya Zakharevich <[email protected]>.

SEE ALSO

The overloading pragma can be used to enable or disable overloadedoperations within a lexical scope - see overloading.

DIAGNOSTICS

When Perl is run with the -Do switch or its equivalent, overloadinginduces diagnostic messages.

Using the m command of Perl debugger (see perldebug) one candeduce which operations are overloaded (and which ancestor triggersthis overloading). Say, if eq is overloaded, then the method (eqis shown by debugger. The method () corresponds to the fallbackkey (in fact a presence of this method shows that this package hasoverloading enabled, and it is what is used by the Overloadedfunction of module overload).

The module might issue the following warnings:

  • Odd number of arguments for overload::constant

    (W) The call to overload::constant contained an odd number of arguments.The arguments should come in pairs.

  • '%s' is not an overloadable type

    (W) You tried to overload a constant type the overload package is unaware of.

  • '%s' is not a code reference

    (W) The second (fourth, sixth, ...) argument of overload::constant needsto be a code reference. Either an anonymous subroutine, or a referenceto a subroutine.

  • overload arg '%s' is invalid

    (W) use overload was passed an argument it did notrecognize. Did you mistype an operator?

BUGS AND PITFALLS

  • No warning is issued for invalid use overload keys.Such errors are not always obvious:

    1.  use overload "+0" => sub { ...; },   # should be "0+"
      2.  
    2.  "not" => sub { ...; };   # should be "!"
      3.

    (Bug #74098)

  • A pitfall when fallback is TRUE and Perl resorts to a built-inimplementation of an operator is that some operators have morethan one semantic, for example |:

    1.  use overload '0+' => sub { $_[0]->{n}; },
      2.  
    2.  fallback => 1;
      3.  
    3.  my $x = bless { n => 4 }, "main";
      4.  
    4.  my $y = bless { n => 8 }, "main";
      5.  
    5.  print $x | $y, "\n";
      6.

    You might expect this to output "12".In fact, it prints "<": the ASCII result of treating "|"as a bitwise string operator - that is, the result of treatingthe operands as the strings "4" and "8" rather than numbers.The fact that numify (0+) is implemented but stringify("") isn't makes no difference since the latter is simplyautogenerated from the former.

    The only way to change this is to provide your own subroutinefor '|'.

  • Magic autogeneration increases the potential for inadvertentlycreating self-referential structures.Currently Perl will not free self-referentialstructures until cycles are explicitly broken.For example,

    1.  use overload '+' => 'add';
      2.  
    2.  sub add { bless [ \$_[0], \$_[1] ] };
      3.

    is asking for trouble, since

    1.  $obj += $y;
      2.

    will effectively become

    1.  $obj = add($obj, $y, undef);
      2.

    with the same result as

    1.  $obj = [\$obj, \$foo];
      2.

    Even if no explicit assignment-variants of operators are present inthe script, they may be generated by the optimizer.For example,

    1.  "obj = $obj\n"
      2.

    may be optimized to

    1.  my $tmp = 'obj = ' . $obj;  $tmp .= "\n";
      2.

  • Because it is used for overloading, the per-package hash%OVERLOAD now has a special meaning in Perl.The symbol table is filled with names looking like line-noise.

  • For the purpose of inheritance every overloaded package behaves as iffallback is present (possibly undefined). This may createinteresting effects if some package is not overloaded, but inheritsfrom two overloaded packages.

  • Before Perl 5.14, the relation between overloading and tie()ing was broken.Overloading was triggered or not based on the previous class of thetie()d variable.

    This happened because the presence of overloading was checkedtoo early, before any tie()d access was attempted. If theclass of the value FETCH()ed from the tied variable does notchange, a simple workaround for code that is to run on older Perlversions is to access the value (via () = $foo or some such)immediately after tie()ing, so that after this call the previous classcoincides with the current one.

  • Barewords are not covered by overloaded string constants.

Page index
 
Source : perldoc.perl.org - Official documentation for the Perl programming language
Site maintained by Jon Allen (JJ)     See the project page for more details
Documentation maintained by the Perl 5 Porters
(Sebelumnya) Perl pragma to restrict unsafe ...Perl pragma to lexically contr ... (Berikutnya)