5 $overload::hint_bits = 0x20000;
13 $ {$package . "::OVERLOAD"}{dummy}++; # Register with magic by touching.
14 *{$package . "::()"} = \&nil; # Make it findable via fetchmethod.
16 if ($_ eq 'fallback') {
20 if (not ref $sub and $sub !~ /::/) {
21 $ {$package . "::(" . $_} = $sub;
24 #print STDERR "Setting `$ {'package'}::\cO$_' to \\&`$sub'.\n";
25 *{$package . "::(" . $_} = \&{ $sub };
28 ${$package . "::()"} = $fb; # Make it findable too (fallback only).
32 $package = (caller())[0];
33 # *{$package . "::OVERLOAD"} = \&OVERLOAD;
35 $package->overload::OVERLOAD(@_);
39 $package = (caller())[0];
40 ${$package . "::OVERLOAD"}{dummy}++; # Upgrade the table
43 if ($_ eq 'fallback') {
44 undef $ {$package . "::()"};
46 delete $ {$package . "::"}{"(" . $_};
53 $package = ref $package if ref $package;
59 return undef unless $globref;
60 my $sub = \&{*$globref};
61 return $sub if $sub ne \&nil;
62 return shift->can($ {*$globref});
65 sub OverloadedStringify {
67 $package = ref $package if ref $package;
69 ov_method mycan($package, '(""'), $package
70 or ov_method mycan($package, '(0+'), $package
71 or ov_method mycan($package, '(bool'), $package
72 or ov_method mycan($package, '(nomethod'), $package;
77 $package = ref $package if ref $package;
78 #my $meth = $package->can('(' . shift);
79 ov_method mycan($package, '(' . shift), $package;
80 #return $meth if $meth ne \&nil;
85 my $package = ref $_[0];
86 return "$_[0]" unless $package;
87 bless $_[0], overload::Fake; # Non-overloaded package
89 bless $_[0], $package; # Back
90 $package . substr $str, index $str, '=';
94 (ref $_[0] && OverloadedStringify($_[0]) or ref($_[0]) eq 'Regexp') ?
99 sub mycan { # Real can would leave stubs.
100 my ($package, $meth) = @_;
101 return \*{$package . "::$meth"} if defined &{$package . "::$meth"};
103 foreach $p (@{$package . "::ISA"}) {
104 my $out = mycan($p, $meth);
118 %ops = ( with_assign => "+ - * / % ** << >> x .",
119 assign => "+= -= *= /= %= **= <<= >>= x= .=",
120 num_comparison => "< <= > >= == !=",
121 '3way_comparison'=> "<=> cmp",
122 str_comparison => "lt le gt ge eq ne",
126 func => "atan2 cos sin exp abs log sqrt int",
127 conversion => 'bool "" 0+',
129 dereferencing => '${} @{} %{} &{} *{}',
130 special => 'nomethod fallback =');
132 use warnings::register;
134 # Arguments: what, sub
137 warnings::warnif ("Odd number of arguments for overload::constant");
140 elsif (!exists $constants {$_ [0]}) {
141 warnings::warnif ("`$_[0]' is not an overloadable type");
143 elsif (!ref $_ [1] || "$_[1]" !~ /CODE\(0x[\da-f]+\)$/) {
144 # Can't use C<ref $_[1] eq "CODE"> above as code references can be
145 # blessed, and C<ref> would return the package the ref is blessed into.
146 if (warnings::enabled) {
147 $_ [1] = "undef" unless defined $_ [1];
148 warnings::warn ("`$_[1]' is not a code reference");
153 $^H |= $constants{$_[0]} | $overload::hint_bits;
159 sub remove_constant {
160 # Arguments: what, sub
163 $^H &= ~ $constants{$_[0]};
174 overload - Package for overloading perl operations
187 $a = new SomeThing 57;
190 if (overload::Overloaded $b) {...}
192 $strval = overload::StrVal $b;
196 =head2 Declaration of overloaded functions
198 The compilation directive
205 declares function Number::add() for addition, and method muas() in
206 the "class" C<Number> (or one of its base classes)
207 for the assignment form C<*=> of multiplication.
209 Arguments of this directive come in (key, value) pairs. Legal values
210 are values legal inside a C<&{ ... }> call, so the name of a
211 subroutine, a reference to a subroutine, or an anonymous subroutine
212 will all work. Note that values specified as strings are
213 interpreted as methods, not subroutines. Legal keys are listed below.
215 The subroutine C<add> will be called to execute C<$a+$b> if $a
216 is a reference to an object blessed into the package C<Number>, or if $a is
217 not an object from a package with defined mathemagic addition, but $b is a
218 reference to a C<Number>. It can also be called in other situations, like
219 C<$a+=7>, or C<$a++>. See L<MAGIC AUTOGENERATION>. (Mathemagical
220 methods refer to methods triggered by an overloaded mathematical
223 Since overloading respects inheritance via the @ISA hierarchy, the
224 above declaration would also trigger overloading of C<+> and C<*=> in
225 all the packages which inherit from C<Number>.
227 =head2 Calling Conventions for Binary Operations
229 The functions specified in the C<use overload ...> directive are called
230 with three (in one particular case with four, see L<Last Resort>)
231 arguments. If the corresponding operation is binary, then the first
232 two arguments are the two arguments of the operation. However, due to
233 general object calling conventions, the first argument should always be
234 an object in the package, so in the situation of C<7+$a>, the
235 order of the arguments is interchanged. It probably does not matter
236 when implementing the addition method, but whether the arguments
237 are reversed is vital to the subtraction method. The method can
238 query this information by examining the third argument, which can take
239 three different values:
245 the order of arguments is as in the current operation.
249 the arguments are reversed.
253 the current operation is an assignment variant (as in
254 C<$a+=7>), but the usual function is called instead. This additional
255 information can be used to generate some optimizations. Compare
256 L<Calling Conventions for Mutators>.
260 =head2 Calling Conventions for Unary Operations
262 Unary operation are considered binary operations with the second
263 argument being C<undef>. Thus the functions that overloads C<{"++"}>
264 is called with arguments C<($a,undef,'')> when $a++ is executed.
266 =head2 Calling Conventions for Mutators
268 Two types of mutators have different calling conventions:
272 =item C<++> and C<-->
274 The routines which implement these operators are expected to actually
275 I<mutate> their arguments. So, assuming that $obj is a reference to a
278 sub incr { my $n = $ {$_[0]}; ++$n; $_[0] = bless \$n}
280 is an appropriate implementation of overloaded C<++>. Note that
282 sub incr { ++$ {$_[0]} ; shift }
284 is OK if used with preincrement and with postincrement. (In the case
285 of postincrement a copying will be performed, see L<Copy Constructor>.)
287 =item C<x=> and other assignment versions
289 There is nothing special about these methods. They may change the
290 value of their arguments, and may leave it as is. The result is going
291 to be assigned to the value in the left-hand-side if different from
294 This allows for the same method to be used as overloaded C<+=> and
295 C<+>. Note that this is I<allowed>, but not recommended, since by the
296 semantic of L<"Fallback"> Perl will call the method for C<+> anyway,
297 if C<+=> is not overloaded.
301 B<Warning.> Due to the presence of assignment versions of operations,
302 routines which may be called in assignment context may create
303 self-referential structures. Currently Perl will not free self-referential
304 structures until cycles are C<explicitly> broken. You may get problems
305 when traversing your structures too.
309 use overload '+' => sub { bless [ \$_[0], \$_[1] ] };
311 is asking for trouble, since for code C<$obj += $foo> the subroutine
312 is called as C<$obj = add($obj, $foo, undef)>, or C<$obj = [\$obj,
313 \$foo]>. If using such a subroutine is an important optimization, one
314 can overload C<+=> explicitly by a non-"optimized" version, or switch
315 to non-optimized version if C<not defined $_[2]> (see
316 L<Calling Conventions for Binary Operations>).
318 Even if no I<explicit> assignment-variants of operators are present in
319 the script, they may be generated by the optimizer. Say, C<",$obj,"> or
320 C<',' . $obj . ','> may be both optimized to
322 my $tmp = ',' . $obj; $tmp .= ',';
324 =head2 Overloadable Operations
326 The following symbols can be specified in C<use overload> directive:
330 =item * I<Arithmetic operations>
332 "+", "+=", "-", "-=", "*", "*=", "/", "/=", "%", "%=",
333 "**", "**=", "<<", "<<=", ">>", ">>=", "x", "x=", ".", ".=",
335 For these operations a substituted non-assignment variant can be called if
336 the assignment variant is not available. Methods for operations C<+>,
337 C<->, C<+=>, and C<-=> can be called to automatically generate
338 increment and decrement methods. The operation C<-> can be used to
339 autogenerate missing methods for unary minus or C<abs>.
341 See L<"MAGIC AUTOGENERATION">, L<"Calling Conventions for Mutators"> and
342 L<"Calling Conventions for Binary Operations">) for details of these
345 =item * I<Comparison operations>
347 "<", "<=", ">", ">=", "==", "!=", "<=>",
348 "lt", "le", "gt", "ge", "eq", "ne", "cmp",
350 If the corresponding "spaceship" variant is available, it can be
351 used to substitute for the missing operation. During C<sort>ing
352 arrays, C<cmp> is used to compare values subject to C<use overload>.
354 =item * I<Bit operations>
356 "&", "^", "|", "neg", "!", "~",
358 C<neg> stands for unary minus. If the method for C<neg> is not
359 specified, it can be autogenerated using the method for
360 subtraction. If the method for C<!> is not specified, it can be
361 autogenerated using the methods for C<bool>, or C<"">, or C<0+>.
363 =item * I<Increment and decrement>
367 If undefined, addition and subtraction methods can be
368 used instead. These operations are called both in prefix and
371 =item * I<Transcendental functions>
373 "atan2", "cos", "sin", "exp", "abs", "log", "sqrt", "int"
375 If C<abs> is unavailable, it can be autogenerated using methods
376 for "E<lt>" or "E<lt>=E<gt>" combined with either unary minus or subtraction.
378 Note that traditionally the Perl function L<int> rounds to 0, thus for
379 floating-point-like types one should follow the same semantic. If
380 C<int> is unavailable, it can be autogenerated using the overloading of
383 =item * I<Boolean, string and numeric conversion>
387 If one or two of these operations are not overloaded, the remaining ones can
388 be used instead. C<bool> is used in the flow control operators
389 (like C<while>) and for the ternary C<?:> operation. These functions can
390 return any arbitrary Perl value. If the corresponding operation for this value
391 is overloaded too, that operation will be called again with this value.
393 As a special case if the overload returns the object itself then it will
394 be used directly. An overloaded conversion returning the object is
395 probably a bug, because you're likely to get something that looks like
396 C<YourPackage=HASH(0x8172b34)>.
402 If not overloaded, the argument will be converted to a filehandle or
403 glob (which may require a stringification). The same overloading
404 happens both for the I<read-filehandle> syntax C<E<lt>$varE<gt>> and
405 I<globbing> syntax C<E<lt>${var}E<gt>>.
407 B<BUGS> Even in list context, the iterator is currently called only
408 once and with scalar context.
410 =item * I<Dereferencing>
412 '${}', '@{}', '%{}', '&{}', '*{}'.
414 If not overloaded, the argument will be dereferenced I<as is>, thus
415 should be of correct type. These functions should return a reference
416 of correct type, or another object with overloaded dereferencing.
418 As a special case if the overload returns the object itself then it
419 will be used directly (provided it is the correct type).
421 The dereference operators must be specified explicitly they will not be passed to
426 "nomethod", "fallback", "=",
428 see L<SPECIAL SYMBOLS FOR C<use overload>>.
432 See L<"Fallback"> for an explanation of when a missing method can be
435 A computer-readable form of the above table is available in the hash
436 %overload::ops, with values being space-separated lists of names:
438 with_assign => '+ - * / % ** << >> x .',
439 assign => '+= -= *= /= %= **= <<= >>= x= .=',
440 num_comparison => '< <= > >= == !=',
441 '3way_comparison'=> '<=> cmp',
442 str_comparison => 'lt le gt ge eq ne',
446 func => 'atan2 cos sin exp abs log sqrt',
447 conversion => 'bool "" 0+',
449 dereferencing => '${} @{} %{} &{} *{}',
450 special => 'nomethod fallback ='
452 =head2 Inheritance and overloading
454 Inheritance interacts with overloading in two ways.
458 =item Strings as values of C<use overload> directive
462 use overload key => value;
464 is a string, it is interpreted as a method name.
466 =item Overloading of an operation is inherited by derived classes
468 Any class derived from an overloaded class is also overloaded. The
469 set of overloaded methods is the union of overloaded methods of all
470 the ancestors. If some method is overloaded in several ancestor, then
471 which description will be used is decided by the usual inheritance
474 If C<A> inherits from C<B> and C<C> (in this order), C<B> overloads
475 C<+> with C<\&D::plus_sub>, and C<C> overloads C<+> by C<"plus_meth">,
476 then the subroutine C<D::plus_sub> will be called to implement
477 operation C<+> for an object in package C<A>.
481 Note that since the value of the C<fallback> key is not a subroutine,
482 its inheritance is not governed by the above rules. In the current
483 implementation, the value of C<fallback> in the first overloaded
484 ancestor is used, but this is accidental and subject to change.
486 =head1 SPECIAL SYMBOLS FOR C<use overload>
488 Three keys are recognized by Perl that are not covered by the above
493 C<"nomethod"> should be followed by a reference to a function of four
494 parameters. If defined, it is called when the overloading mechanism
495 cannot find a method for some operation. The first three arguments of
496 this function coincide with the arguments for the corresponding method if
497 it were found, the fourth argument is the symbol
498 corresponding to the missing method. If several methods are tried,
499 the last one is used. Say, C<1-$a> can be equivalent to
501 &nomethodMethod($a,1,1,"-")
503 if the pair C<"nomethod" =E<gt> "nomethodMethod"> was specified in the
504 C<use overload> directive.
506 The C<"nomethod"> mechanism is I<not> used for the dereference operators
507 ( ${} @{} %{} &{} *{} ).
510 If some operation cannot be resolved, and there is no function
511 assigned to C<"nomethod">, then an exception will be raised via die()--
512 unless C<"fallback"> was specified as a key in C<use overload> directive.
517 The key C<"fallback"> governs what to do if a method for a particular
518 operation is not found. Three different cases are possible depending on
519 the value of C<"fallback">:
526 substituted method (see L<MAGIC AUTOGENERATION>). If this fails, it
527 then tries to calls C<"nomethod"> value; if missing, an exception
532 The same as for the C<undef> value, but no exception is raised. Instead,
533 it silently reverts to what it would have done were there no C<use overload>
536 =item * defined, but FALSE
538 No autogeneration is tried. Perl tries to call
539 C<"nomethod"> value, and if this is missing, raises an exception.
543 B<Note.> C<"fallback"> inheritance via @ISA is not carved in stone
544 yet, see L<"Inheritance and overloading">.
546 =head2 Copy Constructor
548 The value for C<"="> is a reference to a function with three
549 arguments, i.e., it looks like the other values in C<use
550 overload>. However, it does not overload the Perl assignment
551 operator. This would go against Camel hair.
553 This operation is called in the situations when a mutator is applied
554 to a reference that shares its object with some other reference, such
560 To make this change $a and not change $b, a copy of C<$$a> is made,
561 and $a is assigned a reference to this new object. This operation is
562 done during execution of the C<++$a>, and not during the assignment,
563 (so before the increment C<$$a> coincides with C<$$b>). This is only
564 done if C<++> is expressed via a method for C<'++'> or C<'+='> (or
565 C<nomethod>). Note that if this operation is expressed via C<'+'>
566 a nonmutator, i.e., as in
571 then C<$a> does not reference a new copy of C<$$a>, since $$a does not
572 appear as lvalue when the above code is executed.
574 If the copy constructor is required during the execution of some mutator,
575 but a method for C<'='> was not specified, it can be autogenerated as a
576 string copy if the object is a plain scalar.
582 The actually executed code for
585 Something else which does not modify $a or $b....
591 Something else which does not modify $a or $b....
592 $a = $a->clone(undef,"");
595 if $b was mathemagical, and C<'++'> was overloaded with C<\&incr>,
596 C<'='> was overloaded with C<\&clone>.
600 Same behaviour is triggered by C<$b = $a++>, which is consider a synonym for
603 =head1 MAGIC AUTOGENERATION
605 If a method for an operation is not found, and the value for C<"fallback"> is
606 TRUE or undefined, Perl tries to autogenerate a substitute method for
607 the missing operation based on the defined operations. Autogenerated method
608 substitutions are possible for the following operations:
612 =item I<Assignment forms of arithmetic operations>
614 C<$a+=$b> can use the method for C<"+"> if the method for C<"+=">
617 =item I<Conversion operations>
619 String, numeric, and boolean conversion are calculated in terms of one
620 another if not all of them are defined.
622 =item I<Increment and decrement>
624 The C<++$a> operation can be expressed in terms of C<$a+=1> or C<$a+1>,
625 and C<$a--> in terms of C<$a-=1> and C<$a-1>.
629 can be expressed in terms of C<$aE<lt>0> and C<-$a> (or C<0-$a>).
633 can be expressed in terms of subtraction.
637 C<!> and C<not> can be expressed in terms of boolean conversion, or
638 string or numerical conversion.
640 =item I<Concatenation>
642 can be expressed in terms of string conversion.
644 =item I<Comparison operations>
646 can be expressed in terms of its "spaceship" counterpart: either
647 C<E<lt>=E<gt>> or C<cmp>:
649 <, >, <=, >=, ==, != in terms of <=>
650 lt, gt, le, ge, eq, ne in terms of cmp
654 <> in terms of builtin operations
656 =item I<Dereferencing>
658 ${} @{} %{} &{} *{} in terms of builtin operations
660 =item I<Copy operator>
662 can be expressed in terms of an assignment to the dereferenced value, if this
663 value is a scalar and not a reference.
667 =head1 Losing overloading
669 The restriction for the comparison operation is that even if, for example,
670 `C<cmp>' should return a blessed reference, the autogenerated `C<lt>'
671 function will produce only a standard logical value based on the
672 numerical value of the result of `C<cmp>'. In particular, a working
673 numeric conversion is needed in this case (possibly expressed in terms of
676 Similarly, C<.=> and C<x=> operators lose their mathemagical properties
677 if the string conversion substitution is applied.
679 When you chop() a mathemagical object it is promoted to a string and its
680 mathemagical properties are lost. The same can happen with other
683 =head1 Run-time Overloading
685 Since all C<use> directives are executed at compile-time, the only way to
686 change overloading during run-time is to
688 eval 'use overload "+" => \&addmethod';
692 eval 'no overload "+", "--", "<="';
694 though the use of these constructs during run-time is questionable.
696 =head1 Public functions
698 Package C<overload.pm> provides the following public functions:
702 =item overload::StrVal(arg)
704 Gives string value of C<arg> as in absence of stringify overloading.
706 =item overload::Overloaded(arg)
708 Returns true if C<arg> is subject to overloading of some operations.
710 =item overload::Method(obj,op)
712 Returns C<undef> or a reference to the method that implements C<op>.
716 =head1 Overloading constants
718 For some application Perl parser mangles constants too much. It is possible
719 to hook into this process via overload::constant() and overload::remove_constant()
722 These functions take a hash as an argument. The recognized keys of this hash
729 to overload integer constants,
733 to overload floating point constants,
737 to overload octal and hexadecimal constants,
741 to overload C<q>-quoted strings, constant pieces of C<qq>- and C<qx>-quoted
742 strings and here-documents,
746 to overload constant pieces of regular expressions.
750 The corresponding values are references to functions which take three arguments:
751 the first one is the I<initial> string form of the constant, the second one
752 is how Perl interprets this constant, the third one is how the constant is used.
753 Note that the initial string form does not
754 contain string delimiters, and has backslashes in backslash-delimiter
755 combinations stripped (thus the value of delimiter is not relevant for
756 processing of this string). The return value of this function is how this
757 constant is going to be interpreted by Perl. The third argument is undefined
758 unless for overloaded C<q>- and C<qr>- constants, it is C<q> in single-quote
759 context (comes from strings, regular expressions, and single-quote HERE
760 documents), it is C<tr> for arguments of C<tr>/C<y> operators,
761 it is C<s> for right-hand side of C<s>-operator, and it is C<qq> otherwise.
763 Since an expression C<"ab$cd,,"> is just a shortcut for C<'ab' . $cd . ',,'>,
764 it is expected that overloaded constant strings are equipped with reasonable
765 overloaded catenation operator, otherwise absurd results will result.
766 Similarly, negative numbers are considered as negations of positive constants.
768 Note that it is probably meaningless to call the functions overload::constant()
769 and overload::remove_constant() from anywhere but import() and unimport() methods.
770 From these methods they may be called as
775 die "unknown import: @_" unless @_ == 1 and $_[0] eq ':constant';
776 overload::constant integer => sub {Math::BigInt->new(shift)};
779 B<BUGS> Currently overloaded-ness of constants does not propagate
782 =head1 IMPLEMENTATION
784 What follows is subject to change RSN.
786 The table of methods for all operations is cached in magic for the
787 symbol table hash for the package. The cache is invalidated during
788 processing of C<use overload>, C<no overload>, new function
789 definitions, and changes in @ISA. However, this invalidation remains
790 unprocessed until the next C<bless>ing into the package. Hence if you
791 want to change overloading structure dynamically, you'll need an
792 additional (fake) C<bless>ing to update the table.
794 (Every SVish thing has a magic queue, and magic is an entry in that
795 queue. This is how a single variable may participate in multiple
796 forms of magic simultaneously. For instance, environment variables
797 regularly have two forms at once: their %ENV magic and their taint
798 magic. However, the magic which implements overloading is applied to
799 the stashes, which are rarely used directly, thus should not slow down
802 If an object belongs to a package using overload, it carries a special
803 flag. Thus the only speed penalty during arithmetic operations without
804 overloading is the checking of this flag.
806 In fact, if C<use overload> is not present, there is almost no overhead
807 for overloadable operations, so most programs should not suffer
808 measurable performance penalties. A considerable effort was made to
809 minimize the overhead when overload is used in some package, but the
810 arguments in question do not belong to packages using overload. When
811 in doubt, test your speed with C<use overload> and without it. So far
812 there have been no reports of substantial speed degradation if Perl is
813 compiled with optimization turned on.
815 There is no size penalty for data if overload is not used. The only
816 size penalty if overload is used in some package is that I<all> the
817 packages acquire a magic during the next C<bless>ing into the
818 package. This magic is three-words-long for packages without
819 overloading, and carries the cache table if the package is overloaded.
821 Copying (C<$a=$b>) is shallow; however, a one-level-deep copying is
822 carried out before any operation that can imply an assignment to the
823 object $a (or $b) refers to, like C<$a++>. You can override this
824 behavior by defining your own copy constructor (see L<"Copy Constructor">).
826 It is expected that arguments to methods that are not explicitly supposed
827 to be changed are constant (but this is not enforced).
829 =head1 Metaphor clash
831 One may wonder why the semantic of overloaded C<=> is so counter intuitive.
832 If it I<looks> counter intuitive to you, you are subject to a metaphor
835 Here is a Perl object metaphor:
837 I< object is a reference to blessed data>
839 and an arithmetic metaphor:
841 I< object is a thing by itself>.
843 The I<main> problem of overloading C<=> is the fact that these metaphors
844 imply different actions on the assignment C<$a = $b> if $a and $b are
845 objects. Perl-think implies that $a becomes a reference to whatever
846 $b was referencing. Arithmetic-think implies that the value of "object"
847 $a is changed to become the value of the object $b, preserving the fact
848 that $a and $b are separate entities.
850 The difference is not relevant in the absence of mutators. After
851 a Perl-way assignment an operation which mutates the data referenced by $a
852 would change the data referenced by $b too. Effectively, after
853 C<$a = $b> values of $a and $b become I<indistinguishable>.
855 On the other hand, anyone who has used algebraic notation knows the
856 expressive power of the arithmetic metaphor. Overloading works hard
857 to enable this metaphor while preserving the Perlian way as far as
858 possible. Since it is not possible to freely mix two contradicting
859 metaphors, overloading allows the arithmetic way to write things I<as
860 far as all the mutators are called via overloaded access only>. The
861 way it is done is described in L<Copy Constructor>.
863 If some mutator methods are directly applied to the overloaded values,
864 one may need to I<explicitly unlink> other values which references the
869 $b = $a; # $b is "linked" to $a
871 $a = $a->clone; # Unlink $b from $a
874 Note that overloaded access makes this transparent:
877 $b = $a; # $b is "linked" to $a
878 $a += 4; # would unlink $b automagically
880 However, it would not make
883 $a = 4; # Now $a is a plain 4, not 'Data'
885 preserve "objectness" of $a. But Perl I<has> a way to make assignments
886 to an object do whatever you want. It is just not the overload, but
887 tie()ing interface (see L<perlfunc/tie>). Adding a FETCH() method
888 which returns the object itself, and STORE() method which changes the
889 value of the object, one can reproduce the arithmetic metaphor in its
890 completeness, at least for variables which were tie()d from the start.
892 (Note that a workaround for a bug may be needed, see L<"BUGS">.)
896 Please add examples to what follows!
898 =head2 Two-face scalars
900 Put this in F<two_face.pm> in your Perl library directory:
902 package two_face; # Scalars with separate string and
904 sub new { my $p = shift; bless [@_], $p }
905 use overload '""' => \&str, '0+' => \&num, fallback => 1;
912 my $seven = new two_face ("vii", 7);
913 printf "seven=$seven, seven=%d, eight=%d\n", $seven, $seven+1;
914 print "seven contains `i'\n" if $seven =~ /i/;
916 (The second line creates a scalar which has both a string value, and a
917 numeric value.) This prints:
919 seven=vii, seven=7, eight=8
922 =head2 Two-face references
924 Suppose you want to create an object which is accessible as both an
925 array reference and a hash reference.
928 use overload '%{}' => \&gethash, '@{}' => sub { $ {shift()} };
936 tie %h, ref $self, $self;
940 sub TIEHASH { my $p = shift; bless \ shift, $p }
943 $fields{$_} = $i++ foreach qw{zero one two three};
945 my $self = ${shift()};
946 my $key = $fields{shift()};
947 defined $key or die "Out of band access";
948 $$self->[$key] = shift;
951 my $self = ${shift()};
952 my $key = $fields{shift()};
953 defined $key or die "Out of band access";
957 Now one can access an object using both the array and hash syntax:
959 my $bar = new two_refs 3,4,5,6;
961 $bar->{two} == 11 or die 'bad hash fetch';
963 Note several important features of this example. First of all, the
964 I<actual> type of $bar is a scalar reference, and we do not overload
965 the scalar dereference. Thus we can get the I<actual> non-overloaded
966 contents of $bar by just using C<$$bar> (what we do in functions which
967 overload dereference). Similarly, the object returned by the
968 TIEHASH() method is a scalar reference.
970 Second, we create a new tied hash each time the hash syntax is used.
971 This allows us not to worry about a possibility of a reference loop,
972 which would lead to a memory leak.
974 Both these problems can be cured. Say, if we want to overload hash
975 dereference on a reference to an object which is I<implemented> as a
976 hash itself, the only problem one has to circumvent is how to access
977 this I<actual> hash (as opposed to the I<virtual> hash exhibited by the
978 overloaded dereference operator). Here is one possible fetching routine:
981 my ($self, $key) = (shift, shift);
982 my $class = ref $self;
983 bless $self, 'overload::dummy'; # Disable overloading of %{}
984 my $out = $self->{$key};
985 bless $self, $class; # Restore overloading
989 To remove creation of the tied hash on each access, one may an extra
990 level of indirection which allows a non-circular structure of references:
993 use overload '%{}' => sub { ${shift()}->[1] },
994 '@{}' => sub { ${shift()}->[0] };
1000 bless \ [$a, \%h], $p;
1005 tie %h, ref $self, $self;
1009 sub TIEHASH { my $p = shift; bless \ shift, $p }
1012 $fields{$_} = $i++ foreach qw{zero one two three};
1015 my $key = $fields{shift()};
1016 defined $key or die "Out of band access";
1021 my $key = $fields{shift()};
1022 defined $key or die "Out of band access";
1026 Now if $baz is overloaded like this, then C<$baz> is a reference to a
1027 reference to the intermediate array, which keeps a reference to an
1028 actual array, and the access hash. The tie()ing object for the access
1029 hash is a reference to a reference to the actual array, so
1035 There are no loops of references.
1039 Both "objects" which are blessed into the class C<two_refs1> are
1040 references to a reference to an array, thus references to a I<scalar>.
1041 Thus the accessor expression C<$$foo-E<gt>[$ind]> involves no
1042 overloaded operations.
1046 =head2 Symbolic calculator
1048 Put this in F<symbolic.pm> in your Perl library directory:
1050 package symbolic; # Primitive symbolic calculator
1051 use overload nomethod => \&wrap;
1053 sub new { shift; bless ['n', @_] }
1055 my ($obj, $other, $inv, $meth) = @_;
1056 ($obj, $other) = ($other, $obj) if $inv;
1057 bless [$meth, $obj, $other];
1060 This module is very unusual as overloaded modules go: it does not
1061 provide any usual overloaded operators, instead it provides the L<Last
1062 Resort> operator C<nomethod>. In this example the corresponding
1063 subroutine returns an object which encapsulates operations done over
1064 the objects: C<new symbolic 3> contains C<['n', 3]>, C<2 + new
1065 symbolic 3> contains C<['+', 2, ['n', 3]]>.
1067 Here is an example of the script which "calculates" the side of
1068 circumscribed octagon using the above package:
1071 my $iter = 1; # 2**($iter+2) = 8
1072 my $side = new symbolic 1;
1076 $side = (sqrt(1 + $side**2) - 1)/$side;
1080 The value of $side is
1082 ['/', ['-', ['sqrt', ['+', 1, ['**', ['n', 1], 2]],
1083 undef], 1], ['n', 1]]
1085 Note that while we obtained this value using a nice little script,
1086 there is no simple way to I<use> this value. In fact this value may
1087 be inspected in debugger (see L<perldebug>), but ony if
1088 C<bareStringify> B<O>ption is set, and not via C<p> command.
1090 If one attempts to print this value, then the overloaded operator
1091 C<""> will be called, which will call C<nomethod> operator. The
1092 result of this operator will be stringified again, but this result is
1093 again of type C<symbolic>, which will lead to an infinite loop.
1095 Add a pretty-printer method to the module F<symbolic.pm>:
1098 my ($meth, $a, $b) = @{+shift};
1099 $a = 'u' unless defined $a;
1100 $b = 'u' unless defined $b;
1101 $a = $a->pretty if ref $a;
1102 $b = $b->pretty if ref $b;
1106 Now one can finish the script by
1108 print "side = ", $side->pretty, "\n";
1110 The method C<pretty> is doing object-to-string conversion, so it
1111 is natural to overload the operator C<""> using this method. However,
1112 inside such a method it is not necessary to pretty-print the
1113 I<components> $a and $b of an object. In the above subroutine
1114 C<"[$meth $a $b]"> is a catenation of some strings and components $a
1115 and $b. If these components use overloading, the catenation operator
1116 will look for an overloaded operator C<.>; if not present, it will
1117 look for an overloaded operator C<"">. Thus it is enough to use
1119 use overload nomethod => \&wrap, '""' => \&str;
1121 my ($meth, $a, $b) = @{+shift};
1122 $a = 'u' unless defined $a;
1123 $b = 'u' unless defined $b;
1127 Now one can change the last line of the script to
1129 print "side = $side\n";
1133 side = [/ [- [sqrt [+ 1 [** [n 1 u] 2]] u] 1] [n 1 u]]
1135 and one can inspect the value in debugger using all the possible
1138 Something is still amiss: consider the loop variable $cnt of the
1139 script. It was a number, not an object. We cannot make this value of
1140 type C<symbolic>, since then the loop will not terminate.
1142 Indeed, to terminate the cycle, the $cnt should become false.
1143 However, the operator C<bool> for checking falsity is overloaded (this
1144 time via overloaded C<"">), and returns a long string, thus any object
1145 of type C<symbolic> is true. To overcome this, we need a way to
1146 compare an object to 0. In fact, it is easier to write a numeric
1149 Here is the text of F<symbolic.pm> with such a routine added (and
1150 slightly modified str()):
1152 package symbolic; # Primitive symbolic calculator
1154 nomethod => \&wrap, '""' => \&str, '0+' => \#
1156 sub new { shift; bless ['n', @_] }
1158 my ($obj, $other, $inv, $meth) = @_;
1159 ($obj, $other) = ($other, $obj) if $inv;
1160 bless [$meth, $obj, $other];
1163 my ($meth, $a, $b) = @{+shift};
1164 $a = 'u' unless defined $a;
1171 my %subr = ( n => sub {$_[0]},
1172 sqrt => sub {sqrt $_[0]},
1173 '-' => sub {shift() - shift()},
1174 '+' => sub {shift() + shift()},
1175 '/' => sub {shift() / shift()},
1176 '*' => sub {shift() * shift()},
1177 '**' => sub {shift() ** shift()},
1180 my ($meth, $a, $b) = @{+shift};
1181 my $subr = $subr{$meth}
1182 or die "Do not know how to ($meth) in symbolic";
1183 $a = $a->num if ref $a eq __PACKAGE__;
1184 $b = $b->num if ref $b eq __PACKAGE__;
1188 All the work of numeric conversion is done in %subr and num(). Of
1189 course, %subr is not complete, it contains only operators used in the
1190 example below. Here is the extra-credit question: why do we need an
1191 explicit recursion in num()? (Answer is at the end of this section.)
1193 Use this module like this:
1196 my $iter = new symbolic 2; # 16-gon
1197 my $side = new symbolic 1;
1201 $cnt = $cnt - 1; # Mutator `--' not implemented
1202 $side = (sqrt(1 + $side**2) - 1)/$side;
1204 printf "%s=%f\n", $side, $side;
1205 printf "pi=%f\n", $side*(2**($iter+2));
1207 It prints (without so many line breaks)
1209 [/ [- [sqrt [+ 1 [** [/ [- [sqrt [+ 1 [** [n 1] 2]]] 1]
1211 [/ [- [sqrt [+ 1 [** [n 1] 2]]] 1] [n 1]]]=0.198912
1214 The above module is very primitive. It does not implement
1215 mutator methods (C<++>, C<-=> and so on), does not do deep copying
1216 (not required without mutators!), and implements only those arithmetic
1217 operations which are used in the example.
1219 To implement most arithmetic operations is easy; one should just use
1220 the tables of operations, and change the code which fills %subr to
1222 my %subr = ( 'n' => sub {$_[0]} );
1223 foreach my $op (split " ", $overload::ops{with_assign}) {
1224 $subr{$op} = $subr{"$op="} = eval "sub {shift() $op shift()}";
1226 my @bins = qw(binary 3way_comparison num_comparison str_comparison);
1227 foreach my $op (split " ", "@overload::ops{ @bins }") {
1228 $subr{$op} = eval "sub {shift() $op shift()}";
1230 foreach my $op (split " ", "@overload::ops{qw(unary func)}") {
1231 print "defining `$op'\n";
1232 $subr{$op} = eval "sub {$op shift()}";
1235 Due to L<Calling Conventions for Mutators>, we do not need anything
1236 special to make C<+=> and friends work, except filling C<+=> entry of
1237 %subr, and defining a copy constructor (needed since Perl has no
1238 way to know that the implementation of C<'+='> does not mutate
1239 the argument, compare L<Copy Constructor>).
1241 To implement a copy constructor, add C<< '=' => \&cpy >> to C<use overload>
1242 line, and code (this code assumes that mutators change things one level
1243 deep only, so recursive copying is not needed):
1247 bless [@$self], ref $self;
1250 To make C<++> and C<--> work, we need to implement actual mutators,
1251 either directly, or in C<nomethod>. We continue to do things inside
1252 C<nomethod>, thus add
1254 if ($meth eq '++' or $meth eq '--') {
1255 @$obj = ($meth, (bless [@$obj]), 1); # Avoid circular reference
1259 after the first line of wrap(). This is not a most effective
1260 implementation, one may consider
1262 sub inc { $_[0] = bless ['++', shift, 1]; }
1266 As a final remark, note that one can fill %subr by
1268 my %subr = ( 'n' => sub {$_[0]} );
1269 foreach my $op (split " ", $overload::ops{with_assign}) {
1270 $subr{$op} = $subr{"$op="} = eval "sub {shift() $op shift()}";
1272 my @bins = qw(binary 3way_comparison num_comparison str_comparison);
1273 foreach my $op (split " ", "@overload::ops{ @bins }") {
1274 $subr{$op} = eval "sub {shift() $op shift()}";
1276 foreach my $op (split " ", "@overload::ops{qw(unary func)}") {
1277 $subr{$op} = eval "sub {$op shift()}";
1279 $subr{'++'} = $subr{'+'};
1280 $subr{'--'} = $subr{'-'};
1282 This finishes implementation of a primitive symbolic calculator in
1283 50 lines of Perl code. Since the numeric values of subexpressions
1284 are not cached, the calculator is very slow.
1286 Here is the answer for the exercise: In the case of str(), we need no
1287 explicit recursion since the overloaded C<.>-operator will fall back
1288 to an existing overloaded operator C<"">. Overloaded arithmetic
1289 operators I<do not> fall back to numeric conversion if C<fallback> is
1290 not explicitly requested. Thus without an explicit recursion num()
1291 would convert C<['+', $a, $b]> to C<$a + $b>, which would just rebuild
1292 the argument of num().
1294 If you wonder why defaults for conversion are different for str() and
1295 num(), note how easy it was to write the symbolic calculator. This
1296 simplicity is due to an appropriate choice of defaults. One extra
1297 note: due to the explicit recursion num() is more fragile than sym():
1298 we need to explicitly check for the type of $a and $b. If components
1299 $a and $b happen to be of some related type, this may lead to problems.
1301 =head2 I<Really> symbolic calculator
1303 One may wonder why we call the above calculator symbolic. The reason
1304 is that the actual calculation of the value of expression is postponed
1305 until the value is I<used>.
1307 To see it in action, add a method
1312 @$obj->[0,1] = ('=', shift);
1315 to the package C<symbolic>. After this change one can do
1317 my $a = new symbolic 3;
1318 my $b = new symbolic 4;
1319 my $c = sqrt($a**2 + $b**2);
1321 and the numeric value of $c becomes 5. However, after calling
1323 $a->STORE(12); $b->STORE(5);
1325 the numeric value of $c becomes 13. There is no doubt now that the module
1326 symbolic provides a I<symbolic> calculator indeed.
1328 To hide the rough edges under the hood, provide a tie()d interface to the
1329 package C<symbolic> (compare with L<Metaphor clash>). Add methods
1331 sub TIESCALAR { my $pack = shift; $pack->new(@_) }
1333 sub nop { } # Around a bug
1335 (the bug is described in L<"BUGS">). One can use this new interface as
1337 tie $a, 'symbolic', 3;
1338 tie $b, 'symbolic', 4;
1339 $a->nop; $b->nop; # Around a bug
1341 my $c = sqrt($a**2 + $b**2);
1343 Now numeric value of $c is 5. After C<$a = 12; $b = 5> the numeric value
1344 of $c becomes 13. To insulate the user of the module add a method
1346 sub vars { my $p = shift; tie($_, $p), $_->nop foreach @_; }
1351 symbolic->vars($a, $b);
1352 my $c = sqrt($a**2 + $b**2);
1355 printf "c5 %s=%f\n", $c, $c;
1358 printf "c13 %s=%f\n", $c, $c;
1360 shows that the numeric value of $c follows changes to the values of $a
1365 Ilya Zakharevich E<lt>F<ilya@math.mps.ohio-state.edu>E<gt>.
1369 When Perl is run with the B<-Do> switch or its equivalent, overloading
1370 induces diagnostic messages.
1372 Using the C<m> command of Perl debugger (see L<perldebug>) one can
1373 deduce which operations are overloaded (and which ancestor triggers
1374 this overloading). Say, if C<eq> is overloaded, then the method C<(eq>
1375 is shown by debugger. The method C<()> corresponds to the C<fallback>
1376 key (in fact a presence of this method shows that this package has
1377 overloading enabled, and it is what is used by the C<Overloaded>
1378 function of module C<overload>).
1380 The module might issue the following warnings:
1384 =item Odd number of arguments for overload::constant
1386 (W) The call to overload::constant contained an odd number of arguments.
1387 The arguments should come in pairs.
1389 =item `%s' is not an overloadable type
1391 (W) You tried to overload a constant type the overload package is unaware of.
1393 =item `%s' is not a code reference
1395 (W) The second (fourth, sixth, ...) argument of overload::constant needs
1396 to be a code reference. Either an anonymous subroutine, or a reference
1403 Because it is used for overloading, the per-package hash %OVERLOAD now
1404 has a special meaning in Perl. The symbol table is filled with names
1405 looking like line-noise.
1407 For the purpose of inheritance every overloaded package behaves as if
1408 C<fallback> is present (possibly undefined). This may create
1409 interesting effects if some package is not overloaded, but inherits
1410 from two overloaded packages.
1412 Relation between overloading and tie()ing is broken. Overloading is
1413 triggered or not basing on the I<previous> class of tie()d value.
1415 This happens because the presence of overloading is checked too early,
1416 before any tie()d access is attempted. If the FETCH()ed class of the
1417 tie()d value does not change, a simple workaround is to access the value
1418 immediately after tie()ing, so that after this call the I<previous> class
1419 coincides with the current one.
1421 B<Needed:> a way to fix this without a speed penalty.
1423 Barewords are not covered by overloaded string constants.
1425 This document is confusing. There are grammos and misleading language
1426 used in places. It would seem a total rewrite is needed.