Commit | Line | Data |
---|---|---|
a0d0e21e LW |
1 | =head1 NAME |
2 | ||
3 | perlsub - Perl subroutines | |
4 | ||
5 | =head1 SYNOPSIS | |
6 | ||
7 | To declare subroutines: | |
8 | ||
09bef843 SB |
9 | sub NAME; # A "forward" declaration. |
10 | sub NAME(PROTO); # ditto, but with prototypes | |
11 | sub NAME : ATTRS; # with attributes | |
12 | sub NAME(PROTO) : ATTRS; # with attributes and prototypes | |
cb1a09d0 | 13 | |
09bef843 SB |
14 | sub NAME BLOCK # A declaration and a definition. |
15 | sub NAME(PROTO) BLOCK # ditto, but with prototypes | |
16 | sub NAME : ATTRS BLOCK # with attributes | |
17 | sub NAME(PROTO) : ATTRS BLOCK # with prototypes and attributes | |
a0d0e21e | 18 | |
748a9306 LW |
19 | To define an anonymous subroutine at runtime: |
20 | ||
09bef843 SB |
21 | $subref = sub BLOCK; # no proto |
22 | $subref = sub (PROTO) BLOCK; # with proto | |
23 | $subref = sub : ATTRS BLOCK; # with attributes | |
24 | $subref = sub (PROTO) : ATTRS BLOCK; # with proto and attributes | |
748a9306 | 25 | |
a0d0e21e LW |
26 | To import subroutines: |
27 | ||
19799a22 | 28 | use MODULE qw(NAME1 NAME2 NAME3); |
a0d0e21e LW |
29 | |
30 | To call subroutines: | |
31 | ||
5f05dabc | 32 | NAME(LIST); # & is optional with parentheses. |
54310121 | 33 | NAME LIST; # Parentheses optional if predeclared/imported. |
19799a22 | 34 | &NAME(LIST); # Circumvent prototypes. |
5a964f20 | 35 | &NAME; # Makes current @_ visible to called subroutine. |
a0d0e21e LW |
36 | |
37 | =head1 DESCRIPTION | |
38 | ||
19799a22 GS |
39 | Like many languages, Perl provides for user-defined subroutines. |
40 | These may be located anywhere in the main program, loaded in from | |
41 | other files via the C<do>, C<require>, or C<use> keywords, or | |
be3174d2 | 42 | generated on the fly using C<eval> or anonymous subroutines. |
19799a22 GS |
43 | You can even call a function indirectly using a variable containing |
44 | its name or a CODE reference. | |
cb1a09d0 AD |
45 | |
46 | The Perl model for function call and return values is simple: all | |
47 | functions are passed as parameters one single flat list of scalars, and | |
48 | all functions likewise return to their caller one single flat list of | |
49 | scalars. Any arrays or hashes in these call and return lists will | |
50 | collapse, losing their identities--but you may always use | |
51 | pass-by-reference instead to avoid this. Both call and return lists may | |
52 | contain as many or as few scalar elements as you'd like. (Often a | |
53 | function without an explicit return statement is called a subroutine, but | |
19799a22 GS |
54 | there's really no difference from Perl's perspective.) |
55 | ||
56 | Any arguments passed in show up in the array C<@_>. Therefore, if | |
57 | you called a function with two arguments, those would be stored in | |
58 | C<$_[0]> and C<$_[1]>. The array C<@_> is a local array, but its | |
59 | elements are aliases for the actual scalar parameters. In particular, | |
60 | if an element C<$_[0]> is updated, the corresponding argument is | |
61 | updated (or an error occurs if it is not updatable). If an argument | |
62 | is an array or hash element which did not exist when the function | |
63 | was called, that element is created only when (and if) it is modified | |
64 | or a reference to it is taken. (Some earlier versions of Perl | |
65 | created the element whether or not the element was assigned to.) | |
66 | Assigning to the whole array C<@_> removes that aliasing, and does | |
67 | not update any arguments. | |
68 | ||
69 | The return value of a subroutine is the value of the last expression | |
70 | evaluated. More explicitly, a C<return> statement may be used to exit the | |
54310121 | 71 | subroutine, optionally specifying the returned value, which will be |
72 | evaluated in the appropriate context (list, scalar, or void) depending | |
73 | on the context of the subroutine call. If you specify no return value, | |
19799a22 GS |
74 | the subroutine returns an empty list in list context, the undefined |
75 | value in scalar context, or nothing in void context. If you return | |
76 | one or more aggregates (arrays and hashes), these will be flattened | |
77 | together into one large indistinguishable list. | |
78 | ||
79 | Perl does not have named formal parameters. In practice all you | |
80 | do is assign to a C<my()> list of these. Variables that aren't | |
81 | declared to be private are global variables. For gory details | |
82 | on creating private variables, see L<"Private Variables via my()"> | |
83 | and L<"Temporary Values via local()">. To create protected | |
84 | environments for a set of functions in a separate package (and | |
85 | probably a separate file), see L<perlmod/"Packages">. | |
a0d0e21e LW |
86 | |
87 | Example: | |
88 | ||
cb1a09d0 AD |
89 | sub max { |
90 | my $max = shift(@_); | |
a0d0e21e LW |
91 | foreach $foo (@_) { |
92 | $max = $foo if $max < $foo; | |
93 | } | |
cb1a09d0 | 94 | return $max; |
a0d0e21e | 95 | } |
cb1a09d0 | 96 | $bestday = max($mon,$tue,$wed,$thu,$fri); |
a0d0e21e LW |
97 | |
98 | Example: | |
99 | ||
100 | # get a line, combining continuation lines | |
101 | # that start with whitespace | |
102 | ||
103 | sub get_line { | |
19799a22 | 104 | $thisline = $lookahead; # global variables! |
54310121 | 105 | LINE: while (defined($lookahead = <STDIN>)) { |
a0d0e21e LW |
106 | if ($lookahead =~ /^[ \t]/) { |
107 | $thisline .= $lookahead; | |
108 | } | |
109 | else { | |
110 | last LINE; | |
111 | } | |
112 | } | |
19799a22 | 113 | return $thisline; |
a0d0e21e LW |
114 | } |
115 | ||
116 | $lookahead = <STDIN>; # get first line | |
19799a22 | 117 | while (defined($line = get_line())) { |
a0d0e21e LW |
118 | ... |
119 | } | |
120 | ||
09bef843 | 121 | Assigning to a list of private variables to name your arguments: |
a0d0e21e LW |
122 | |
123 | sub maybeset { | |
124 | my($key, $value) = @_; | |
cb1a09d0 | 125 | $Foo{$key} = $value unless $Foo{$key}; |
a0d0e21e LW |
126 | } |
127 | ||
19799a22 GS |
128 | Because the assignment copies the values, this also has the effect |
129 | of turning call-by-reference into call-by-value. Otherwise a | |
130 | function is free to do in-place modifications of C<@_> and change | |
131 | its caller's values. | |
cb1a09d0 AD |
132 | |
133 | upcase_in($v1, $v2); # this changes $v1 and $v2 | |
134 | sub upcase_in { | |
54310121 | 135 | for (@_) { tr/a-z/A-Z/ } |
136 | } | |
cb1a09d0 AD |
137 | |
138 | You aren't allowed to modify constants in this way, of course. If an | |
139 | argument were actually literal and you tried to change it, you'd take a | |
140 | (presumably fatal) exception. For example, this won't work: | |
141 | ||
142 | upcase_in("frederick"); | |
143 | ||
f86cebdf | 144 | It would be much safer if the C<upcase_in()> function |
cb1a09d0 AD |
145 | were written to return a copy of its parameters instead |
146 | of changing them in place: | |
147 | ||
19799a22 | 148 | ($v3, $v4) = upcase($v1, $v2); # this doesn't change $v1 and $v2 |
cb1a09d0 | 149 | sub upcase { |
54310121 | 150 | return unless defined wantarray; # void context, do nothing |
cb1a09d0 | 151 | my @parms = @_; |
54310121 | 152 | for (@parms) { tr/a-z/A-Z/ } |
c07a80fd | 153 | return wantarray ? @parms : $parms[0]; |
54310121 | 154 | } |
cb1a09d0 | 155 | |
19799a22 | 156 | Notice how this (unprototyped) function doesn't care whether it was |
a2293a43 | 157 | passed real scalars or arrays. Perl sees all arguments as one big, |
19799a22 GS |
158 | long, flat parameter list in C<@_>. This is one area where |
159 | Perl's simple argument-passing style shines. The C<upcase()> | |
160 | function would work perfectly well without changing the C<upcase()> | |
161 | definition even if we fed it things like this: | |
cb1a09d0 AD |
162 | |
163 | @newlist = upcase(@list1, @list2); | |
164 | @newlist = upcase( split /:/, $var ); | |
165 | ||
166 | Do not, however, be tempted to do this: | |
167 | ||
168 | (@a, @b) = upcase(@list1, @list2); | |
169 | ||
19799a22 GS |
170 | Like the flattened incoming parameter list, the return list is also |
171 | flattened on return. So all you have managed to do here is stored | |
17b63f68 | 172 | everything in C<@a> and made C<@b> empty. See |
13a2d996 | 173 | L<Pass by Reference> for alternatives. |
19799a22 GS |
174 | |
175 | A subroutine may be called using an explicit C<&> prefix. The | |
176 | C<&> is optional in modern Perl, as are parentheses if the | |
177 | subroutine has been predeclared. The C<&> is I<not> optional | |
178 | when just naming the subroutine, such as when it's used as | |
179 | an argument to defined() or undef(). Nor is it optional when you | |
180 | want to do an indirect subroutine call with a subroutine name or | |
181 | reference using the C<&$subref()> or C<&{$subref}()> constructs, | |
c47ff5f1 | 182 | although the C<< $subref->() >> notation solves that problem. |
19799a22 GS |
183 | See L<perlref> for more about all that. |
184 | ||
185 | Subroutines may be called recursively. If a subroutine is called | |
186 | using the C<&> form, the argument list is optional, and if omitted, | |
187 | no C<@_> array is set up for the subroutine: the C<@_> array at the | |
188 | time of the call is visible to subroutine instead. This is an | |
189 | efficiency mechanism that new users may wish to avoid. | |
a0d0e21e LW |
190 | |
191 | &foo(1,2,3); # pass three arguments | |
192 | foo(1,2,3); # the same | |
193 | ||
194 | foo(); # pass a null list | |
195 | &foo(); # the same | |
a0d0e21e | 196 | |
cb1a09d0 | 197 | &foo; # foo() get current args, like foo(@_) !! |
54310121 | 198 | foo; # like foo() IFF sub foo predeclared, else "foo" |
cb1a09d0 | 199 | |
19799a22 GS |
200 | Not only does the C<&> form make the argument list optional, it also |
201 | disables any prototype checking on arguments you do provide. This | |
c07a80fd | 202 | is partly for historical reasons, and partly for having a convenient way |
19799a22 | 203 | to cheat if you know what you're doing. See L<Prototypes> below. |
c07a80fd | 204 | |
09bef843 | 205 | Functions whose names are in all upper case are reserved to the Perl |
19799a22 GS |
206 | core, as are modules whose names are in all lower case. A |
207 | function in all capitals is a loosely-held convention meaning it | |
208 | will be called indirectly by the run-time system itself, usually | |
209 | due to a triggered event. Functions that do special, pre-defined | |
f2fc0a40 AB |
210 | things include C<BEGIN>, C<CHECK>, C<INIT>, C<END>, C<AUTOLOAD>, |
211 | C<CLONE> and C<DESTROY>--plus all functions mentioned in L<perltie>. | |
5a964f20 | 212 | |
b687b08b | 213 | =head2 Private Variables via my() |
cb1a09d0 AD |
214 | |
215 | Synopsis: | |
216 | ||
217 | my $foo; # declare $foo lexically local | |
218 | my (@wid, %get); # declare list of variables local | |
219 | my $foo = "flurp"; # declare $foo lexical, and init it | |
220 | my @oof = @bar; # declare @oof lexical, and init it | |
09bef843 SB |
221 | my $x : Foo = $y; # similar, with an attribute applied |
222 | ||
a0ae32d3 JH |
223 | B<WARNING>: The use of attribute lists on C<my> declarations is still |
224 | evolving. The current semantics and interface are subject to change. | |
225 | See L<attributes> and L<Attribute::Handlers>. | |
cb1a09d0 | 226 | |
19799a22 GS |
227 | The C<my> operator declares the listed variables to be lexically |
228 | confined to the enclosing block, conditional (C<if/unless/elsif/else>), | |
229 | loop (C<for/foreach/while/until/continue>), subroutine, C<eval>, | |
230 | or C<do/require/use>'d file. If more than one value is listed, the | |
231 | list must be placed in parentheses. All listed elements must be | |
232 | legal lvalues. Only alphanumeric identifiers may be lexically | |
09bef843 | 233 | scoped--magical built-ins like C<$/> must currently be C<local>ize |
19799a22 GS |
234 | with C<local> instead. |
235 | ||
236 | Unlike dynamic variables created by the C<local> operator, lexical | |
237 | variables declared with C<my> are totally hidden from the outside | |
238 | world, including any called subroutines. This is true if it's the | |
239 | same subroutine called from itself or elsewhere--every call gets | |
240 | its own copy. | |
241 | ||
242 | This doesn't mean that a C<my> variable declared in a statically | |
243 | enclosing lexical scope would be invisible. Only dynamic scopes | |
244 | are cut off. For example, the C<bumpx()> function below has access | |
245 | to the lexical $x variable because both the C<my> and the C<sub> | |
246 | occurred at the same scope, presumably file scope. | |
5a964f20 TC |
247 | |
248 | my $x = 10; | |
249 | sub bumpx { $x++ } | |
250 | ||
19799a22 GS |
251 | An C<eval()>, however, can see lexical variables of the scope it is |
252 | being evaluated in, so long as the names aren't hidden by declarations within | |
253 | the C<eval()> itself. See L<perlref>. | |
cb1a09d0 | 254 | |
19799a22 | 255 | The parameter list to my() may be assigned to if desired, which allows you |
cb1a09d0 AD |
256 | to initialize your variables. (If no initializer is given for a |
257 | particular variable, it is created with the undefined value.) Commonly | |
19799a22 | 258 | this is used to name input parameters to a subroutine. Examples: |
cb1a09d0 AD |
259 | |
260 | $arg = "fred"; # "global" variable | |
261 | $n = cube_root(27); | |
262 | print "$arg thinks the root is $n\n"; | |
263 | fred thinks the root is 3 | |
264 | ||
265 | sub cube_root { | |
266 | my $arg = shift; # name doesn't matter | |
267 | $arg **= 1/3; | |
268 | return $arg; | |
54310121 | 269 | } |
cb1a09d0 | 270 | |
19799a22 GS |
271 | The C<my> is simply a modifier on something you might assign to. So when |
272 | you do assign to variables in its argument list, C<my> doesn't | |
6cc33c6d | 273 | change whether those variables are viewed as a scalar or an array. So |
cb1a09d0 | 274 | |
5a964f20 | 275 | my ($foo) = <STDIN>; # WRONG? |
cb1a09d0 AD |
276 | my @FOO = <STDIN>; |
277 | ||
5f05dabc | 278 | both supply a list context to the right-hand side, while |
cb1a09d0 AD |
279 | |
280 | my $foo = <STDIN>; | |
281 | ||
5f05dabc | 282 | supplies a scalar context. But the following declares only one variable: |
748a9306 | 283 | |
5a964f20 | 284 | my $foo, $bar = 1; # WRONG |
748a9306 | 285 | |
cb1a09d0 | 286 | That has the same effect as |
748a9306 | 287 | |
cb1a09d0 AD |
288 | my $foo; |
289 | $bar = 1; | |
a0d0e21e | 290 | |
cb1a09d0 AD |
291 | The declared variable is not introduced (is not visible) until after |
292 | the current statement. Thus, | |
293 | ||
294 | my $x = $x; | |
295 | ||
19799a22 | 296 | can be used to initialize a new $x with the value of the old $x, and |
cb1a09d0 AD |
297 | the expression |
298 | ||
299 | my $x = 123 and $x == 123 | |
300 | ||
19799a22 | 301 | is false unless the old $x happened to have the value C<123>. |
cb1a09d0 | 302 | |
55497cff | 303 | Lexical scopes of control structures are not bounded precisely by the |
304 | braces that delimit their controlled blocks; control expressions are | |
19799a22 | 305 | part of that scope, too. Thus in the loop |
55497cff | 306 | |
19799a22 | 307 | while (my $line = <>) { |
55497cff | 308 | $line = lc $line; |
309 | } continue { | |
310 | print $line; | |
311 | } | |
312 | ||
19799a22 | 313 | the scope of $line extends from its declaration throughout the rest of |
55497cff | 314 | the loop construct (including the C<continue> clause), but not beyond |
315 | it. Similarly, in the conditional | |
316 | ||
317 | if ((my $answer = <STDIN>) =~ /^yes$/i) { | |
318 | user_agrees(); | |
319 | } elsif ($answer =~ /^no$/i) { | |
320 | user_disagrees(); | |
321 | } else { | |
322 | chomp $answer; | |
323 | die "'$answer' is neither 'yes' nor 'no'"; | |
324 | } | |
325 | ||
19799a22 GS |
326 | the scope of $answer extends from its declaration through the rest |
327 | of that conditional, including any C<elsif> and C<else> clauses, | |
55497cff | 328 | but not beyond it. |
329 | ||
bc92d8c0 NC |
330 | B<NOTE:> The behaviour of a C<my> statement modified with a statement |
331 | modifier conditional or loop construct (e.g. C<my $x if ...>) is | |
332 | B<undefined>. The value of the C<my> variable may be C<undef>, any | |
333 | previously assigned value, or possibly anything else. Don't rely on | |
334 | it. Future versions of perl might do something different from the | |
335 | version of perl you try it out on. Here be dragons. | |
55497cff | 336 | |
5f05dabc | 337 | The C<foreach> loop defaults to scoping its index variable dynamically |
19799a22 GS |
338 | in the manner of C<local>. However, if the index variable is |
339 | prefixed with the keyword C<my>, or if there is already a lexical | |
340 | by that name in scope, then a new lexical is created instead. Thus | |
341 | in the loop | |
55497cff | 342 | |
343 | for my $i (1, 2, 3) { | |
344 | some_function(); | |
345 | } | |
346 | ||
19799a22 GS |
347 | the scope of $i extends to the end of the loop, but not beyond it, |
348 | rendering the value of $i inaccessible within C<some_function()>. | |
55497cff | 349 | |
cb1a09d0 | 350 | Some users may wish to encourage the use of lexically scoped variables. |
19799a22 GS |
351 | As an aid to catching implicit uses to package variables, |
352 | which are always global, if you say | |
cb1a09d0 AD |
353 | |
354 | use strict 'vars'; | |
355 | ||
19799a22 GS |
356 | then any variable mentioned from there to the end of the enclosing |
357 | block must either refer to a lexical variable, be predeclared via | |
77ca0c92 | 358 | C<our> or C<use vars>, or else must be fully qualified with the package name. |
19799a22 GS |
359 | A compilation error results otherwise. An inner block may countermand |
360 | this with C<no strict 'vars'>. | |
361 | ||
362 | A C<my> has both a compile-time and a run-time effect. At compile | |
8593bda5 | 363 | time, the compiler takes notice of it. The principal usefulness |
19799a22 GS |
364 | of this is to quiet C<use strict 'vars'>, but it is also essential |
365 | for generation of closures as detailed in L<perlref>. Actual | |
366 | initialization is delayed until run time, though, so it gets executed | |
367 | at the appropriate time, such as each time through a loop, for | |
368 | example. | |
369 | ||
370 | Variables declared with C<my> are not part of any package and are therefore | |
cb1a09d0 AD |
371 | never fully qualified with the package name. In particular, you're not |
372 | allowed to try to make a package variable (or other global) lexical: | |
373 | ||
374 | my $pack::var; # ERROR! Illegal syntax | |
375 | my $_; # also illegal (currently) | |
376 | ||
377 | In fact, a dynamic variable (also known as package or global variables) | |
f86cebdf | 378 | are still accessible using the fully qualified C<::> notation even while a |
cb1a09d0 AD |
379 | lexical of the same name is also visible: |
380 | ||
381 | package main; | |
382 | local $x = 10; | |
383 | my $x = 20; | |
384 | print "$x and $::x\n"; | |
385 | ||
f86cebdf | 386 | That will print out C<20> and C<10>. |
cb1a09d0 | 387 | |
19799a22 GS |
388 | You may declare C<my> variables at the outermost scope of a file |
389 | to hide any such identifiers from the world outside that file. This | |
390 | is similar in spirit to C's static variables when they are used at | |
391 | the file level. To do this with a subroutine requires the use of | |
392 | a closure (an anonymous function that accesses enclosing lexicals). | |
393 | If you want to create a private subroutine that cannot be called | |
394 | from outside that block, it can declare a lexical variable containing | |
395 | an anonymous sub reference: | |
cb1a09d0 AD |
396 | |
397 | my $secret_version = '1.001-beta'; | |
398 | my $secret_sub = sub { print $secret_version }; | |
399 | &$secret_sub(); | |
400 | ||
401 | As long as the reference is never returned by any function within the | |
5f05dabc | 402 | module, no outside module can see the subroutine, because its name is not in |
cb1a09d0 | 403 | any package's symbol table. Remember that it's not I<REALLY> called |
19799a22 | 404 | C<$some_pack::secret_version> or anything; it's just $secret_version, |
cb1a09d0 AD |
405 | unqualified and unqualifiable. |
406 | ||
19799a22 GS |
407 | This does not work with object methods, however; all object methods |
408 | have to be in the symbol table of some package to be found. See | |
409 | L<perlref/"Function Templates"> for something of a work-around to | |
410 | this. | |
cb1a09d0 | 411 | |
c2611fb3 | 412 | =head2 Persistent Private Variables |
5a964f20 TC |
413 | |
414 | Just because a lexical variable is lexically (also called statically) | |
f86cebdf | 415 | scoped to its enclosing block, C<eval>, or C<do> FILE, this doesn't mean that |
5a964f20 TC |
416 | within a function it works like a C static. It normally works more |
417 | like a C auto, but with implicit garbage collection. | |
418 | ||
419 | Unlike local variables in C or C++, Perl's lexical variables don't | |
420 | necessarily get recycled just because their scope has exited. | |
421 | If something more permanent is still aware of the lexical, it will | |
422 | stick around. So long as something else references a lexical, that | |
423 | lexical won't be freed--which is as it should be. You wouldn't want | |
424 | memory being free until you were done using it, or kept around once you | |
425 | were done. Automatic garbage collection takes care of this for you. | |
426 | ||
427 | This means that you can pass back or save away references to lexical | |
428 | variables, whereas to return a pointer to a C auto is a grave error. | |
429 | It also gives us a way to simulate C's function statics. Here's a | |
430 | mechanism for giving a function private variables with both lexical | |
431 | scoping and a static lifetime. If you do want to create something like | |
432 | C's static variables, just enclose the whole function in an extra block, | |
433 | and put the static variable outside the function but in the block. | |
cb1a09d0 AD |
434 | |
435 | { | |
54310121 | 436 | my $secret_val = 0; |
cb1a09d0 AD |
437 | sub gimme_another { |
438 | return ++$secret_val; | |
54310121 | 439 | } |
440 | } | |
cb1a09d0 AD |
441 | # $secret_val now becomes unreachable by the outside |
442 | # world, but retains its value between calls to gimme_another | |
443 | ||
54310121 | 444 | If this function is being sourced in from a separate file |
cb1a09d0 | 445 | via C<require> or C<use>, then this is probably just fine. If it's |
19799a22 | 446 | all in the main program, you'll need to arrange for the C<my> |
cb1a09d0 | 447 | to be executed early, either by putting the whole block above |
f86cebdf | 448 | your main program, or more likely, placing merely a C<BEGIN> |
cb1a09d0 AD |
449 | sub around it to make sure it gets executed before your program |
450 | starts to run: | |
451 | ||
452 | sub BEGIN { | |
54310121 | 453 | my $secret_val = 0; |
cb1a09d0 AD |
454 | sub gimme_another { |
455 | return ++$secret_val; | |
54310121 | 456 | } |
457 | } | |
cb1a09d0 | 458 | |
19799a22 | 459 | See L<perlmod/"Package Constructors and Destructors"> about the |
7d30b5c4 | 460 | special triggered functions, C<BEGIN>, C<CHECK>, C<INIT> and C<END>. |
cb1a09d0 | 461 | |
19799a22 GS |
462 | If declared at the outermost scope (the file scope), then lexicals |
463 | work somewhat like C's file statics. They are available to all | |
464 | functions in that same file declared below them, but are inaccessible | |
465 | from outside that file. This strategy is sometimes used in modules | |
466 | to create private variables that the whole module can see. | |
5a964f20 | 467 | |
cb1a09d0 AD |
468 | =head2 Temporary Values via local() |
469 | ||
19799a22 | 470 | B<WARNING>: In general, you should be using C<my> instead of C<local>, because |
6d28dffb | 471 | it's faster and safer. Exceptions to this include the global punctuation |
cb1a09d0 | 472 | variables, filehandles and formats, and direct manipulation of the Perl |
19799a22 | 473 | symbol table itself. Format variables often use C<local> though, as do |
cb1a09d0 AD |
474 | other variables whose current value must be visible to called |
475 | subroutines. | |
476 | ||
477 | Synopsis: | |
478 | ||
479 | local $foo; # declare $foo dynamically local | |
480 | local (@wid, %get); # declare list of variables local | |
481 | local $foo = "flurp"; # declare $foo dynamic, and init it | |
482 | local @oof = @bar; # declare @oof dynamic, and init it | |
483 | ||
484 | local *FH; # localize $FH, @FH, %FH, &FH ... | |
485 | local *merlyn = *randal; # now $merlyn is really $randal, plus | |
486 | # @merlyn is really @randal, etc | |
487 | local *merlyn = 'randal'; # SAME THING: promote 'randal' to *randal | |
54310121 | 488 | local *merlyn = \$randal; # just alias $merlyn, not @merlyn etc |
cb1a09d0 | 489 | |
19799a22 GS |
490 | A C<local> modifies its listed variables to be "local" to the |
491 | enclosing block, C<eval>, or C<do FILE>--and to I<any subroutine | |
492 | called from within that block>. A C<local> just gives temporary | |
493 | values to global (meaning package) variables. It does I<not> create | |
494 | a local variable. This is known as dynamic scoping. Lexical scoping | |
495 | is done with C<my>, which works more like C's auto declarations. | |
cb1a09d0 | 496 | |
19799a22 | 497 | If more than one variable is given to C<local>, they must be placed in |
5f05dabc | 498 | parentheses. All listed elements must be legal lvalues. This operator works |
cb1a09d0 | 499 | by saving the current values of those variables in its argument list on a |
5f05dabc | 500 | hidden stack and restoring them upon exiting the block, subroutine, or |
cb1a09d0 AD |
501 | eval. This means that called subroutines can also reference the local |
502 | variable, but not the global one. The argument list may be assigned to if | |
503 | desired, which allows you to initialize your local variables. (If no | |
504 | initializer is given for a particular variable, it is created with an | |
505 | undefined value.) Commonly this is used to name the parameters to a | |
506 | subroutine. Examples: | |
507 | ||
508 | for $i ( 0 .. 9 ) { | |
509 | $digits{$i} = $i; | |
54310121 | 510 | } |
cb1a09d0 | 511 | # assume this function uses global %digits hash |
54310121 | 512 | parse_num(); |
cb1a09d0 AD |
513 | |
514 | # now temporarily add to %digits hash | |
515 | if ($base12) { | |
516 | # (NOTE: not claiming this is efficient!) | |
517 | local %digits = (%digits, 't' => 10, 'e' => 11); | |
518 | parse_num(); # parse_num gets this new %digits! | |
519 | } | |
520 | # old %digits restored here | |
521 | ||
19799a22 | 522 | Because C<local> is a run-time operator, it gets executed each time |
cb1a09d0 AD |
523 | through a loop. In releases of Perl previous to 5.0, this used more stack |
524 | storage each time until the loop was exited. Perl now reclaims the space | |
525 | each time through, but it's still more efficient to declare your variables | |
526 | outside the loop. | |
527 | ||
f86cebdf GS |
528 | A C<local> is simply a modifier on an lvalue expression. When you assign to |
529 | a C<local>ized variable, the C<local> doesn't change whether its list is viewed | |
cb1a09d0 AD |
530 | as a scalar or an array. So |
531 | ||
532 | local($foo) = <STDIN>; | |
533 | local @FOO = <STDIN>; | |
534 | ||
5f05dabc | 535 | both supply a list context to the right-hand side, while |
cb1a09d0 AD |
536 | |
537 | local $foo = <STDIN>; | |
538 | ||
539 | supplies a scalar context. | |
540 | ||
3e3baf6d TB |
541 | A note about C<local()> and composite types is in order. Something |
542 | like C<local(%foo)> works by temporarily placing a brand new hash in | |
543 | the symbol table. The old hash is left alone, but is hidden "behind" | |
544 | the new one. | |
545 | ||
546 | This means the old variable is completely invisible via the symbol | |
547 | table (i.e. the hash entry in the C<*foo> typeglob) for the duration | |
548 | of the dynamic scope within which the C<local()> was seen. This | |
549 | has the effect of allowing one to temporarily occlude any magic on | |
550 | composite types. For instance, this will briefly alter a tied | |
551 | hash to some other implementation: | |
552 | ||
553 | tie %ahash, 'APackage'; | |
554 | [...] | |
555 | { | |
556 | local %ahash; | |
557 | tie %ahash, 'BPackage'; | |
558 | [..called code will see %ahash tied to 'BPackage'..] | |
559 | { | |
560 | local %ahash; | |
561 | [..%ahash is a normal (untied) hash here..] | |
562 | } | |
563 | } | |
564 | [..%ahash back to its initial tied self again..] | |
565 | ||
fd5a896a DM |
566 | B<WARNING> The code example above does not currently work as described. |
567 | This will be fixed in a future release of Perl; in the meantime, avoid | |
568 | code that relies on any particular behaviour of localising tied arrays | |
569 | or hashes (localising individual elements is still okay). | |
570 | See L<perldelta/"Localising Tied Arrays and Hashes Is Broken"> for more | |
571 | details. | |
572 | ||
3e3baf6d TB |
573 | As another example, a custom implementation of C<%ENV> might look |
574 | like this: | |
575 | ||
576 | { | |
577 | local %ENV; | |
578 | tie %ENV, 'MyOwnEnv'; | |
579 | [..do your own fancy %ENV manipulation here..] | |
580 | } | |
581 | [..normal %ENV behavior here..] | |
582 | ||
6ee623d5 | 583 | It's also worth taking a moment to explain what happens when you |
f86cebdf GS |
584 | C<local>ize a member of a composite type (i.e. an array or hash element). |
585 | In this case, the element is C<local>ized I<by name>. This means that | |
6ee623d5 GS |
586 | when the scope of the C<local()> ends, the saved value will be |
587 | restored to the hash element whose key was named in the C<local()>, or | |
588 | the array element whose index was named in the C<local()>. If that | |
589 | element was deleted while the C<local()> was in effect (e.g. by a | |
590 | C<delete()> from a hash or a C<shift()> of an array), it will spring | |
591 | back into existence, possibly extending an array and filling in the | |
592 | skipped elements with C<undef>. For instance, if you say | |
593 | ||
594 | %hash = ( 'This' => 'is', 'a' => 'test' ); | |
595 | @ary = ( 0..5 ); | |
596 | { | |
597 | local($ary[5]) = 6; | |
598 | local($hash{'a'}) = 'drill'; | |
599 | while (my $e = pop(@ary)) { | |
600 | print "$e . . .\n"; | |
601 | last unless $e > 3; | |
602 | } | |
603 | if (@ary) { | |
604 | $hash{'only a'} = 'test'; | |
605 | delete $hash{'a'}; | |
606 | } | |
607 | } | |
608 | print join(' ', map { "$_ $hash{$_}" } sort keys %hash),".\n"; | |
609 | print "The array has ",scalar(@ary)," elements: ", | |
610 | join(', ', map { defined $_ ? $_ : 'undef' } @ary),"\n"; | |
611 | ||
612 | Perl will print | |
613 | ||
614 | 6 . . . | |
615 | 4 . . . | |
616 | 3 . . . | |
617 | This is a test only a test. | |
618 | The array has 6 elements: 0, 1, 2, undef, undef, 5 | |
619 | ||
19799a22 | 620 | The behavior of local() on non-existent members of composite |
7185e5cc GS |
621 | types is subject to change in future. |
622 | ||
cd06dffe GS |
623 | =head2 Lvalue subroutines |
624 | ||
e6a32221 JC |
625 | B<WARNING>: Lvalue subroutines are still experimental and the |
626 | implementation may change in future versions of Perl. | |
cd06dffe GS |
627 | |
628 | It is possible to return a modifiable value from a subroutine. | |
629 | To do this, you have to declare the subroutine to return an lvalue. | |
630 | ||
631 | my $val; | |
632 | sub canmod : lvalue { | |
e6a32221 | 633 | # return $val; this doesn't work, don't say "return" |
cd06dffe GS |
634 | $val; |
635 | } | |
636 | sub nomod { | |
637 | $val; | |
638 | } | |
639 | ||
640 | canmod() = 5; # assigns to $val | |
641 | nomod() = 5; # ERROR | |
642 | ||
643 | The scalar/list context for the subroutine and for the right-hand | |
644 | side of assignment is determined as if the subroutine call is replaced | |
645 | by a scalar. For example, consider: | |
646 | ||
647 | data(2,3) = get_data(3,4); | |
648 | ||
649 | Both subroutines here are called in a scalar context, while in: | |
650 | ||
651 | (data(2,3)) = get_data(3,4); | |
652 | ||
653 | and in: | |
654 | ||
655 | (data(2),data(3)) = get_data(3,4); | |
656 | ||
657 | all the subroutines are called in a list context. | |
658 | ||
e6a32221 JC |
659 | =over 4 |
660 | ||
661 | =item Lvalue subroutines are EXPERIMENTAL | |
662 | ||
663 | They appear to be convenient, but there are several reasons to be | |
664 | circumspect. | |
665 | ||
666 | You can't use the return keyword, you must pass out the value before | |
667 | falling out of subroutine scope. (see comment in example above). This | |
668 | is usually not a problem, but it disallows an explicit return out of a | |
669 | deeply nested loop, which is sometimes a nice way out. | |
670 | ||
671 | They violate encapsulation. A normal mutator can check the supplied | |
672 | argument before setting the attribute it is protecting, an lvalue | |
673 | subroutine never gets that chance. Consider; | |
674 | ||
675 | my $some_array_ref = []; # protected by mutators ?? | |
676 | ||
677 | sub set_arr { # normal mutator | |
678 | my $val = shift; | |
679 | die("expected array, you supplied ", ref $val) | |
680 | unless ref $val eq 'ARRAY'; | |
681 | $some_array_ref = $val; | |
682 | } | |
683 | sub set_arr_lv : lvalue { # lvalue mutator | |
684 | $some_array_ref; | |
685 | } | |
686 | ||
687 | # set_arr_lv cannot stop this ! | |
688 | set_arr_lv() = { a => 1 }; | |
818c4caa | 689 | |
e6a32221 JC |
690 | =back |
691 | ||
cb1a09d0 AD |
692 | =head2 Passing Symbol Table Entries (typeglobs) |
693 | ||
19799a22 GS |
694 | B<WARNING>: The mechanism described in this section was originally |
695 | the only way to simulate pass-by-reference in older versions of | |
696 | Perl. While it still works fine in modern versions, the new reference | |
697 | mechanism is generally easier to work with. See below. | |
a0d0e21e LW |
698 | |
699 | Sometimes you don't want to pass the value of an array to a subroutine | |
700 | but rather the name of it, so that the subroutine can modify the global | |
701 | copy of it rather than working with a local copy. In perl you can | |
cb1a09d0 | 702 | refer to all objects of a particular name by prefixing the name |
5f05dabc | 703 | with a star: C<*foo>. This is often known as a "typeglob", because the |
a0d0e21e LW |
704 | star on the front can be thought of as a wildcard match for all the |
705 | funny prefix characters on variables and subroutines and such. | |
706 | ||
55497cff | 707 | When evaluated, the typeglob produces a scalar value that represents |
5f05dabc | 708 | all the objects of that name, including any filehandle, format, or |
a0d0e21e | 709 | subroutine. When assigned to, it causes the name mentioned to refer to |
19799a22 | 710 | whatever C<*> value was assigned to it. Example: |
a0d0e21e LW |
711 | |
712 | sub doubleary { | |
713 | local(*someary) = @_; | |
714 | foreach $elem (@someary) { | |
715 | $elem *= 2; | |
716 | } | |
717 | } | |
718 | doubleary(*foo); | |
719 | doubleary(*bar); | |
720 | ||
19799a22 | 721 | Scalars are already passed by reference, so you can modify |
a0d0e21e | 722 | scalar arguments without using this mechanism by referring explicitly |
1fef88e7 | 723 | to C<$_[0]> etc. You can modify all the elements of an array by passing |
f86cebdf GS |
724 | all the elements as scalars, but you have to use the C<*> mechanism (or |
725 | the equivalent reference mechanism) to C<push>, C<pop>, or change the size of | |
a0d0e21e LW |
726 | an array. It will certainly be faster to pass the typeglob (or reference). |
727 | ||
728 | Even if you don't want to modify an array, this mechanism is useful for | |
5f05dabc | 729 | passing multiple arrays in a single LIST, because normally the LIST |
a0d0e21e | 730 | mechanism will merge all the array values so that you can't extract out |
55497cff | 731 | the individual arrays. For more on typeglobs, see |
2ae324a7 | 732 | L<perldata/"Typeglobs and Filehandles">. |
cb1a09d0 | 733 | |
5a964f20 TC |
734 | =head2 When to Still Use local() |
735 | ||
19799a22 GS |
736 | Despite the existence of C<my>, there are still three places where the |
737 | C<local> operator still shines. In fact, in these three places, you | |
5a964f20 TC |
738 | I<must> use C<local> instead of C<my>. |
739 | ||
13a2d996 | 740 | =over 4 |
5a964f20 | 741 | |
551e1d92 RB |
742 | =item 1. |
743 | ||
744 | You need to give a global variable a temporary value, especially $_. | |
5a964f20 | 745 | |
f86cebdf GS |
746 | The global variables, like C<@ARGV> or the punctuation variables, must be |
747 | C<local>ized with C<local()>. This block reads in F</etc/motd>, and splits | |
5a964f20 | 748 | it up into chunks separated by lines of equal signs, which are placed |
f86cebdf | 749 | in C<@Fields>. |
5a964f20 TC |
750 | |
751 | { | |
752 | local @ARGV = ("/etc/motd"); | |
753 | local $/ = undef; | |
754 | local $_ = <>; | |
755 | @Fields = split /^\s*=+\s*$/; | |
756 | } | |
757 | ||
19799a22 | 758 | It particular, it's important to C<local>ize $_ in any routine that assigns |
5a964f20 TC |
759 | to it. Look out for implicit assignments in C<while> conditionals. |
760 | ||
551e1d92 RB |
761 | =item 2. |
762 | ||
763 | You need to create a local file or directory handle or a local function. | |
5a964f20 | 764 | |
09bef843 SB |
765 | A function that needs a filehandle of its own must use |
766 | C<local()> on a complete typeglob. This can be used to create new symbol | |
5a964f20 TC |
767 | table entries: |
768 | ||
769 | sub ioqueue { | |
770 | local (*READER, *WRITER); # not my! | |
17b63f68 | 771 | pipe (READER, WRITER) or die "pipe: $!"; |
5a964f20 TC |
772 | return (*READER, *WRITER); |
773 | } | |
774 | ($head, $tail) = ioqueue(); | |
775 | ||
776 | See the Symbol module for a way to create anonymous symbol table | |
777 | entries. | |
778 | ||
779 | Because assignment of a reference to a typeglob creates an alias, this | |
780 | can be used to create what is effectively a local function, or at least, | |
781 | a local alias. | |
782 | ||
783 | { | |
f86cebdf GS |
784 | local *grow = \&shrink; # only until this block exists |
785 | grow(); # really calls shrink() | |
786 | move(); # if move() grow()s, it shrink()s too | |
5a964f20 | 787 | } |
f86cebdf | 788 | grow(); # get the real grow() again |
5a964f20 TC |
789 | |
790 | See L<perlref/"Function Templates"> for more about manipulating | |
791 | functions by name in this way. | |
792 | ||
551e1d92 RB |
793 | =item 3. |
794 | ||
795 | You want to temporarily change just one element of an array or hash. | |
5a964f20 | 796 | |
f86cebdf | 797 | You can C<local>ize just one element of an aggregate. Usually this |
5a964f20 TC |
798 | is done on dynamics: |
799 | ||
800 | { | |
801 | local $SIG{INT} = 'IGNORE'; | |
802 | funct(); # uninterruptible | |
803 | } | |
804 | # interruptibility automatically restored here | |
805 | ||
806 | But it also works on lexically declared aggregates. Prior to 5.005, | |
807 | this operation could on occasion misbehave. | |
808 | ||
809 | =back | |
810 | ||
cb1a09d0 AD |
811 | =head2 Pass by Reference |
812 | ||
55497cff | 813 | If you want to pass more than one array or hash into a function--or |
814 | return them from it--and have them maintain their integrity, then | |
815 | you're going to have to use an explicit pass-by-reference. Before you | |
816 | do that, you need to understand references as detailed in L<perlref>. | |
c07a80fd | 817 | This section may not make much sense to you otherwise. |
cb1a09d0 | 818 | |
19799a22 GS |
819 | Here are a few simple examples. First, let's pass in several arrays |
820 | to a function and have it C<pop> all of then, returning a new list | |
821 | of all their former last elements: | |
cb1a09d0 AD |
822 | |
823 | @tailings = popmany ( \@a, \@b, \@c, \@d ); | |
824 | ||
825 | sub popmany { | |
826 | my $aref; | |
827 | my @retlist = (); | |
828 | foreach $aref ( @_ ) { | |
829 | push @retlist, pop @$aref; | |
54310121 | 830 | } |
cb1a09d0 | 831 | return @retlist; |
54310121 | 832 | } |
cb1a09d0 | 833 | |
54310121 | 834 | Here's how you might write a function that returns a |
cb1a09d0 AD |
835 | list of keys occurring in all the hashes passed to it: |
836 | ||
54310121 | 837 | @common = inter( \%foo, \%bar, \%joe ); |
cb1a09d0 AD |
838 | sub inter { |
839 | my ($k, $href, %seen); # locals | |
840 | foreach $href (@_) { | |
841 | while ( $k = each %$href ) { | |
842 | $seen{$k}++; | |
54310121 | 843 | } |
844 | } | |
cb1a09d0 | 845 | return grep { $seen{$_} == @_ } keys %seen; |
54310121 | 846 | } |
cb1a09d0 | 847 | |
5f05dabc | 848 | So far, we're using just the normal list return mechanism. |
54310121 | 849 | What happens if you want to pass or return a hash? Well, |
850 | if you're using only one of them, or you don't mind them | |
cb1a09d0 | 851 | concatenating, then the normal calling convention is ok, although |
54310121 | 852 | a little expensive. |
cb1a09d0 AD |
853 | |
854 | Where people get into trouble is here: | |
855 | ||
856 | (@a, @b) = func(@c, @d); | |
857 | or | |
858 | (%a, %b) = func(%c, %d); | |
859 | ||
19799a22 GS |
860 | That syntax simply won't work. It sets just C<@a> or C<%a> and |
861 | clears the C<@b> or C<%b>. Plus the function didn't get passed | |
862 | into two separate arrays or hashes: it got one long list in C<@_>, | |
863 | as always. | |
cb1a09d0 AD |
864 | |
865 | If you can arrange for everyone to deal with this through references, it's | |
866 | cleaner code, although not so nice to look at. Here's a function that | |
867 | takes two array references as arguments, returning the two array elements | |
868 | in order of how many elements they have in them: | |
869 | ||
870 | ($aref, $bref) = func(\@c, \@d); | |
871 | print "@$aref has more than @$bref\n"; | |
872 | sub func { | |
873 | my ($cref, $dref) = @_; | |
874 | if (@$cref > @$dref) { | |
875 | return ($cref, $dref); | |
876 | } else { | |
c07a80fd | 877 | return ($dref, $cref); |
54310121 | 878 | } |
879 | } | |
cb1a09d0 AD |
880 | |
881 | It turns out that you can actually do this also: | |
882 | ||
883 | (*a, *b) = func(\@c, \@d); | |
884 | print "@a has more than @b\n"; | |
885 | sub func { | |
886 | local (*c, *d) = @_; | |
887 | if (@c > @d) { | |
888 | return (\@c, \@d); | |
889 | } else { | |
890 | return (\@d, \@c); | |
54310121 | 891 | } |
892 | } | |
cb1a09d0 AD |
893 | |
894 | Here we're using the typeglobs to do symbol table aliasing. It's | |
19799a22 | 895 | a tad subtle, though, and also won't work if you're using C<my> |
09bef843 | 896 | variables, because only globals (even in disguise as C<local>s) |
19799a22 | 897 | are in the symbol table. |
5f05dabc | 898 | |
899 | If you're passing around filehandles, you could usually just use the bare | |
19799a22 GS |
900 | typeglob, like C<*STDOUT>, but typeglobs references work, too. |
901 | For example: | |
5f05dabc | 902 | |
903 | splutter(\*STDOUT); | |
904 | sub splutter { | |
905 | my $fh = shift; | |
906 | print $fh "her um well a hmmm\n"; | |
907 | } | |
908 | ||
909 | $rec = get_rec(\*STDIN); | |
910 | sub get_rec { | |
911 | my $fh = shift; | |
912 | return scalar <$fh>; | |
913 | } | |
914 | ||
19799a22 GS |
915 | If you're planning on generating new filehandles, you could do this. |
916 | Notice to pass back just the bare *FH, not its reference. | |
5f05dabc | 917 | |
918 | sub openit { | |
19799a22 | 919 | my $path = shift; |
5f05dabc | 920 | local *FH; |
e05a3a1e | 921 | return open (FH, $path) ? *FH : undef; |
54310121 | 922 | } |
5f05dabc | 923 | |
cb1a09d0 AD |
924 | =head2 Prototypes |
925 | ||
19799a22 GS |
926 | Perl supports a very limited kind of compile-time argument checking |
927 | using function prototyping. If you declare | |
cb1a09d0 AD |
928 | |
929 | sub mypush (\@@) | |
930 | ||
19799a22 GS |
931 | then C<mypush()> takes arguments exactly like C<push()> does. The |
932 | function declaration must be visible at compile time. The prototype | |
933 | affects only interpretation of new-style calls to the function, | |
934 | where new-style is defined as not using the C<&> character. In | |
935 | other words, if you call it like a built-in function, then it behaves | |
936 | like a built-in function. If you call it like an old-fashioned | |
937 | subroutine, then it behaves like an old-fashioned subroutine. It | |
938 | naturally falls out from this rule that prototypes have no influence | |
939 | on subroutine references like C<\&foo> or on indirect subroutine | |
c47ff5f1 | 940 | calls like C<&{$subref}> or C<< $subref->() >>. |
c07a80fd | 941 | |
942 | Method calls are not influenced by prototypes either, because the | |
19799a22 GS |
943 | function to be called is indeterminate at compile time, since |
944 | the exact code called depends on inheritance. | |
cb1a09d0 | 945 | |
19799a22 GS |
946 | Because the intent of this feature is primarily to let you define |
947 | subroutines that work like built-in functions, here are prototypes | |
948 | for some other functions that parse almost exactly like the | |
949 | corresponding built-in. | |
cb1a09d0 AD |
950 | |
951 | Declared as Called as | |
952 | ||
f86cebdf GS |
953 | sub mylink ($$) mylink $old, $new |
954 | sub myvec ($$$) myvec $var, $offset, 1 | |
955 | sub myindex ($$;$) myindex &getstring, "substr" | |
956 | sub mysyswrite ($$$;$) mysyswrite $buf, 0, length($buf) - $off, $off | |
957 | sub myreverse (@) myreverse $a, $b, $c | |
958 | sub myjoin ($@) myjoin ":", $a, $b, $c | |
959 | sub mypop (\@) mypop @array | |
960 | sub mysplice (\@$$@) mysplice @array, @array, 0, @pushme | |
961 | sub mykeys (\%) mykeys %{$hashref} | |
962 | sub myopen (*;$) myopen HANDLE, $name | |
963 | sub mypipe (**) mypipe READHANDLE, WRITEHANDLE | |
964 | sub mygrep (&@) mygrep { /foo/ } $a, $b, $c | |
965 | sub myrand ($) myrand 42 | |
966 | sub mytime () mytime | |
cb1a09d0 | 967 | |
c07a80fd | 968 | Any backslashed prototype character represents an actual argument |
6e47f808 | 969 | that absolutely must start with that character. The value passed |
19799a22 GS |
970 | as part of C<@_> will be a reference to the actual argument given |
971 | in the subroutine call, obtained by applying C<\> to that argument. | |
c07a80fd | 972 | |
5b794e05 JH |
973 | You can also backslash several argument types simultaneously by using |
974 | the C<\[]> notation: | |
975 | ||
976 | sub myref (\[$@%&*]) | |
977 | ||
978 | will allow calling myref() as | |
979 | ||
980 | myref $var | |
981 | myref @array | |
982 | myref %hash | |
983 | myref &sub | |
984 | myref *glob | |
985 | ||
986 | and the first argument of myref() will be a reference to | |
987 | a scalar, an array, a hash, a code, or a glob. | |
988 | ||
c07a80fd | 989 | Unbackslashed prototype characters have special meanings. Any |
19799a22 | 990 | unbackslashed C<@> or C<%> eats all remaining arguments, and forces |
f86cebdf GS |
991 | list context. An argument represented by C<$> forces scalar context. An |
992 | C<&> requires an anonymous subroutine, which, if passed as the first | |
0df79f0c GS |
993 | argument, does not require the C<sub> keyword or a subsequent comma. |
994 | ||
995 | A C<*> allows the subroutine to accept a bareword, constant, scalar expression, | |
648ca4f7 GS |
996 | typeglob, or a reference to a typeglob in that slot. The value will be |
997 | available to the subroutine either as a simple scalar, or (in the latter | |
0df79f0c GS |
998 | two cases) as a reference to the typeglob. If you wish to always convert |
999 | such arguments to a typeglob reference, use Symbol::qualify_to_ref() as | |
1000 | follows: | |
1001 | ||
1002 | use Symbol 'qualify_to_ref'; | |
1003 | ||
1004 | sub foo (*) { | |
1005 | my $fh = qualify_to_ref(shift, caller); | |
1006 | ... | |
1007 | } | |
c07a80fd | 1008 | |
1009 | A semicolon separates mandatory arguments from optional arguments. | |
19799a22 | 1010 | It is redundant before C<@> or C<%>, which gobble up everything else. |
cb1a09d0 | 1011 | |
19799a22 GS |
1012 | Note how the last three examples in the table above are treated |
1013 | specially by the parser. C<mygrep()> is parsed as a true list | |
1014 | operator, C<myrand()> is parsed as a true unary operator with unary | |
1015 | precedence the same as C<rand()>, and C<mytime()> is truly without | |
1016 | arguments, just like C<time()>. That is, if you say | |
cb1a09d0 AD |
1017 | |
1018 | mytime +2; | |
1019 | ||
f86cebdf | 1020 | you'll get C<mytime() + 2>, not C<mytime(2)>, which is how it would be parsed |
19799a22 | 1021 | without a prototype. |
cb1a09d0 | 1022 | |
19799a22 GS |
1023 | The interesting thing about C<&> is that you can generate new syntax with it, |
1024 | provided it's in the initial position: | |
cb1a09d0 | 1025 | |
6d28dffb | 1026 | sub try (&@) { |
cb1a09d0 AD |
1027 | my($try,$catch) = @_; |
1028 | eval { &$try }; | |
1029 | if ($@) { | |
1030 | local $_ = $@; | |
1031 | &$catch; | |
1032 | } | |
1033 | } | |
55497cff | 1034 | sub catch (&) { $_[0] } |
cb1a09d0 AD |
1035 | |
1036 | try { | |
1037 | die "phooey"; | |
1038 | } catch { | |
1039 | /phooey/ and print "unphooey\n"; | |
1040 | }; | |
1041 | ||
f86cebdf | 1042 | That prints C<"unphooey">. (Yes, there are still unresolved |
19799a22 | 1043 | issues having to do with visibility of C<@_>. I'm ignoring that |
f86cebdf | 1044 | question for the moment. (But note that if we make C<@_> lexically |
cb1a09d0 | 1045 | scoped, those anonymous subroutines can act like closures... (Gee, |
5f05dabc | 1046 | is this sounding a little Lispish? (Never mind.)))) |
cb1a09d0 | 1047 | |
19799a22 | 1048 | And here's a reimplementation of the Perl C<grep> operator: |
cb1a09d0 AD |
1049 | |
1050 | sub mygrep (&@) { | |
1051 | my $code = shift; | |
1052 | my @result; | |
1053 | foreach $_ (@_) { | |
6e47f808 | 1054 | push(@result, $_) if &$code; |
cb1a09d0 AD |
1055 | } |
1056 | @result; | |
1057 | } | |
a0d0e21e | 1058 | |
cb1a09d0 AD |
1059 | Some folks would prefer full alphanumeric prototypes. Alphanumerics have |
1060 | been intentionally left out of prototypes for the express purpose of | |
1061 | someday in the future adding named, formal parameters. The current | |
1062 | mechanism's main goal is to let module writers provide better diagnostics | |
1063 | for module users. Larry feels the notation quite understandable to Perl | |
1064 | programmers, and that it will not intrude greatly upon the meat of the | |
1065 | module, nor make it harder to read. The line noise is visually | |
1066 | encapsulated into a small pill that's easy to swallow. | |
1067 | ||
420cdfc1 ST |
1068 | If you try to use an alphanumeric sequence in a prototype you will |
1069 | generate an optional warning - "Illegal character in prototype...". | |
1070 | Unfortunately earlier versions of Perl allowed the prototype to be | |
1071 | used as long as its prefix was a valid prototype. The warning may be | |
1072 | upgraded to a fatal error in a future version of Perl once the | |
1073 | majority of offending code is fixed. | |
1074 | ||
cb1a09d0 AD |
1075 | It's probably best to prototype new functions, not retrofit prototyping |
1076 | into older ones. That's because you must be especially careful about | |
1077 | silent impositions of differing list versus scalar contexts. For example, | |
1078 | if you decide that a function should take just one parameter, like this: | |
1079 | ||
1080 | sub func ($) { | |
1081 | my $n = shift; | |
1082 | print "you gave me $n\n"; | |
54310121 | 1083 | } |
cb1a09d0 AD |
1084 | |
1085 | and someone has been calling it with an array or expression | |
1086 | returning a list: | |
1087 | ||
1088 | func(@foo); | |
1089 | func( split /:/ ); | |
1090 | ||
19799a22 | 1091 | Then you've just supplied an automatic C<scalar> in front of their |
f86cebdf | 1092 | argument, which can be more than a bit surprising. The old C<@foo> |
cb1a09d0 | 1093 | which used to hold one thing doesn't get passed in. Instead, |
19799a22 GS |
1094 | C<func()> now gets passed in a C<1>; that is, the number of elements |
1095 | in C<@foo>. And the C<split> gets called in scalar context so it | |
1096 | starts scribbling on your C<@_> parameter list. Ouch! | |
cb1a09d0 | 1097 | |
5f05dabc | 1098 | This is all very powerful, of course, and should be used only in moderation |
54310121 | 1099 | to make the world a better place. |
44a8e56a | 1100 | |
1101 | =head2 Constant Functions | |
1102 | ||
1103 | Functions with a prototype of C<()> are potential candidates for | |
19799a22 GS |
1104 | inlining. If the result after optimization and constant folding |
1105 | is either a constant or a lexically-scoped scalar which has no other | |
54310121 | 1106 | references, then it will be used in place of function calls made |
19799a22 GS |
1107 | without C<&>. Calls made using C<&> are never inlined. (See |
1108 | F<constant.pm> for an easy way to declare most constants.) | |
44a8e56a | 1109 | |
5a964f20 | 1110 | The following functions would all be inlined: |
44a8e56a | 1111 | |
699e6cd4 TP |
1112 | sub pi () { 3.14159 } # Not exact, but close. |
1113 | sub PI () { 4 * atan2 1, 1 } # As good as it gets, | |
1114 | # and it's inlined, too! | |
44a8e56a | 1115 | sub ST_DEV () { 0 } |
1116 | sub ST_INO () { 1 } | |
1117 | ||
1118 | sub FLAG_FOO () { 1 << 8 } | |
1119 | sub FLAG_BAR () { 1 << 9 } | |
1120 | sub FLAG_MASK () { FLAG_FOO | FLAG_BAR } | |
54310121 | 1121 | |
1122 | sub OPT_BAZ () { not (0x1B58 & FLAG_MASK) } | |
44a8e56a | 1123 | sub BAZ_VAL () { |
1124 | if (OPT_BAZ) { | |
1125 | return 23; | |
1126 | } | |
1127 | else { | |
1128 | return 42; | |
1129 | } | |
1130 | } | |
cb1a09d0 | 1131 | |
54310121 | 1132 | sub N () { int(BAZ_VAL) / 3 } |
1133 | BEGIN { | |
1134 | my $prod = 1; | |
1135 | for (1..N) { $prod *= $_ } | |
1136 | sub N_FACTORIAL () { $prod } | |
1137 | } | |
1138 | ||
5a964f20 | 1139 | If you redefine a subroutine that was eligible for inlining, you'll get |
4cee8e80 CS |
1140 | a mandatory warning. (You can use this warning to tell whether or not a |
1141 | particular subroutine is considered constant.) The warning is | |
1142 | considered severe enough not to be optional because previously compiled | |
1143 | invocations of the function will still be using the old value of the | |
19799a22 | 1144 | function. If you need to be able to redefine the subroutine, you need to |
4cee8e80 | 1145 | ensure that it isn't inlined, either by dropping the C<()> prototype |
19799a22 | 1146 | (which changes calling semantics, so beware) or by thwarting the |
4cee8e80 CS |
1147 | inlining mechanism in some other way, such as |
1148 | ||
4cee8e80 | 1149 | sub not_inlined () { |
54310121 | 1150 | 23 if $]; |
4cee8e80 CS |
1151 | } |
1152 | ||
19799a22 | 1153 | =head2 Overriding Built-in Functions |
a0d0e21e | 1154 | |
19799a22 | 1155 | Many built-in functions may be overridden, though this should be tried |
5f05dabc | 1156 | only occasionally and for good reason. Typically this might be |
19799a22 | 1157 | done by a package attempting to emulate missing built-in functionality |
a0d0e21e LW |
1158 | on a non-Unix system. |
1159 | ||
5f05dabc | 1160 | Overriding may be done only by importing the name from a |
a0d0e21e | 1161 | module--ordinary predeclaration isn't good enough. However, the |
19799a22 GS |
1162 | C<use subs> pragma lets you, in effect, predeclare subs |
1163 | via the import syntax, and these names may then override built-in ones: | |
a0d0e21e LW |
1164 | |
1165 | use subs 'chdir', 'chroot', 'chmod', 'chown'; | |
1166 | chdir $somewhere; | |
1167 | sub chdir { ... } | |
1168 | ||
19799a22 GS |
1169 | To unambiguously refer to the built-in form, precede the |
1170 | built-in name with the special package qualifier C<CORE::>. For example, | |
1171 | saying C<CORE::open()> always refers to the built-in C<open()>, even | |
fb73857a | 1172 | if the current package has imported some other subroutine called |
19799a22 | 1173 | C<&open()> from elsewhere. Even though it looks like a regular |
09bef843 | 1174 | function call, it isn't: you can't take a reference to it, such as |
19799a22 | 1175 | the incorrect C<\&CORE::open> might appear to produce. |
fb73857a | 1176 | |
19799a22 GS |
1177 | Library modules should not in general export built-in names like C<open> |
1178 | or C<chdir> as part of their default C<@EXPORT> list, because these may | |
a0d0e21e | 1179 | sneak into someone else's namespace and change the semantics unexpectedly. |
19799a22 | 1180 | Instead, if the module adds that name to C<@EXPORT_OK>, then it's |
a0d0e21e LW |
1181 | possible for a user to import the name explicitly, but not implicitly. |
1182 | That is, they could say | |
1183 | ||
1184 | use Module 'open'; | |
1185 | ||
19799a22 | 1186 | and it would import the C<open> override. But if they said |
a0d0e21e LW |
1187 | |
1188 | use Module; | |
1189 | ||
19799a22 | 1190 | they would get the default imports without overrides. |
a0d0e21e | 1191 | |
19799a22 | 1192 | The foregoing mechanism for overriding built-in is restricted, quite |
95d94a4f | 1193 | deliberately, to the package that requests the import. There is a second |
19799a22 | 1194 | method that is sometimes applicable when you wish to override a built-in |
95d94a4f GS |
1195 | everywhere, without regard to namespace boundaries. This is achieved by |
1196 | importing a sub into the special namespace C<CORE::GLOBAL::>. Here is an | |
1197 | example that quite brazenly replaces the C<glob> operator with something | |
1198 | that understands regular expressions. | |
1199 | ||
1200 | package REGlob; | |
1201 | require Exporter; | |
1202 | @ISA = 'Exporter'; | |
1203 | @EXPORT_OK = 'glob'; | |
1204 | ||
1205 | sub import { | |
1206 | my $pkg = shift; | |
1207 | return unless @_; | |
1208 | my $sym = shift; | |
1209 | my $where = ($sym =~ s/^GLOBAL_// ? 'CORE::GLOBAL' : caller(0)); | |
1210 | $pkg->export($where, $sym, @_); | |
1211 | } | |
1212 | ||
1213 | sub glob { | |
1214 | my $pat = shift; | |
1215 | my @got; | |
19799a22 GS |
1216 | local *D; |
1217 | if (opendir D, '.') { | |
1218 | @got = grep /$pat/, readdir D; | |
1219 | closedir D; | |
1220 | } | |
1221 | return @got; | |
95d94a4f GS |
1222 | } |
1223 | 1; | |
1224 | ||
1225 | And here's how it could be (ab)used: | |
1226 | ||
1227 | #use REGlob 'GLOBAL_glob'; # override glob() in ALL namespaces | |
1228 | package Foo; | |
1229 | use REGlob 'glob'; # override glob() in Foo:: only | |
1230 | print for <^[a-z_]+\.pm\$>; # show all pragmatic modules | |
1231 | ||
19799a22 | 1232 | The initial comment shows a contrived, even dangerous example. |
95d94a4f | 1233 | By overriding C<glob> globally, you would be forcing the new (and |
19799a22 | 1234 | subversive) behavior for the C<glob> operator for I<every> namespace, |
95d94a4f GS |
1235 | without the complete cognizance or cooperation of the modules that own |
1236 | those namespaces. Naturally, this should be done with extreme caution--if | |
1237 | it must be done at all. | |
1238 | ||
1239 | The C<REGlob> example above does not implement all the support needed to | |
19799a22 | 1240 | cleanly override perl's C<glob> operator. The built-in C<glob> has |
95d94a4f | 1241 | different behaviors depending on whether it appears in a scalar or list |
19799a22 | 1242 | context, but our C<REGlob> doesn't. Indeed, many perl built-in have such |
95d94a4f GS |
1243 | context sensitive behaviors, and these must be adequately supported by |
1244 | a properly written override. For a fully functional example of overriding | |
1245 | C<glob>, study the implementation of C<File::DosGlob> in the standard | |
1246 | library. | |
1247 | ||
77bc9082 RGS |
1248 | When you override a built-in, your replacement should be consistent (if |
1249 | possible) with the built-in native syntax. You can achieve this by using | |
1250 | a suitable prototype. To get the prototype of an overridable built-in, | |
1251 | use the C<prototype> function with an argument of C<"CORE::builtin_name"> | |
1252 | (see L<perlfunc/prototype>). | |
1253 | ||
1254 | Note however that some built-ins can't have their syntax expressed by a | |
1255 | prototype (such as C<system> or C<chomp>). If you override them you won't | |
1256 | be able to fully mimic their original syntax. | |
1257 | ||
fe854a6f | 1258 | The built-ins C<do>, C<require> and C<glob> can also be overridden, but due |
77bc9082 RGS |
1259 | to special magic, their original syntax is preserved, and you don't have |
1260 | to define a prototype for their replacements. (You can't override the | |
1261 | C<do BLOCK> syntax, though). | |
1262 | ||
1263 | C<require> has special additional dark magic: if you invoke your | |
1264 | C<require> replacement as C<require Foo::Bar>, it will actually receive | |
1265 | the argument C<"Foo/Bar.pm"> in @_. See L<perlfunc/require>. | |
1266 | ||
1267 | And, as you'll have noticed from the previous example, if you override | |
fe854a6f | 1268 | C<glob>, the C<E<lt>*E<gt>> glob operator is overridden as well. |
77bc9082 | 1269 | |
9b3023bc RGS |
1270 | In a similar fashion, overriding the C<readline> function also overrides |
1271 | the equivalent I/O operator C<< <FILEHANDLE> >>. | |
1272 | ||
fe854a6f | 1273 | Finally, some built-ins (e.g. C<exists> or C<grep>) can't be overridden. |
77bc9082 | 1274 | |
a0d0e21e LW |
1275 | =head2 Autoloading |
1276 | ||
19799a22 GS |
1277 | If you call a subroutine that is undefined, you would ordinarily |
1278 | get an immediate, fatal error complaining that the subroutine doesn't | |
1279 | exist. (Likewise for subroutines being used as methods, when the | |
1280 | method doesn't exist in any base class of the class's package.) | |
1281 | However, if an C<AUTOLOAD> subroutine is defined in the package or | |
1282 | packages used to locate the original subroutine, then that | |
1283 | C<AUTOLOAD> subroutine is called with the arguments that would have | |
1284 | been passed to the original subroutine. The fully qualified name | |
1285 | of the original subroutine magically appears in the global $AUTOLOAD | |
1286 | variable of the same package as the C<AUTOLOAD> routine. The name | |
1287 | is not passed as an ordinary argument because, er, well, just | |
1288 | because, that's why... | |
1289 | ||
1290 | Many C<AUTOLOAD> routines load in a definition for the requested | |
1291 | subroutine using eval(), then execute that subroutine using a special | |
1292 | form of goto() that erases the stack frame of the C<AUTOLOAD> routine | |
1293 | without a trace. (See the source to the standard module documented | |
1294 | in L<AutoLoader>, for example.) But an C<AUTOLOAD> routine can | |
1295 | also just emulate the routine and never define it. For example, | |
1296 | let's pretend that a function that wasn't defined should just invoke | |
1297 | C<system> with those arguments. All you'd do is: | |
cb1a09d0 AD |
1298 | |
1299 | sub AUTOLOAD { | |
1300 | my $program = $AUTOLOAD; | |
1301 | $program =~ s/.*:://; | |
1302 | system($program, @_); | |
54310121 | 1303 | } |
cb1a09d0 | 1304 | date(); |
6d28dffb | 1305 | who('am', 'i'); |
cb1a09d0 AD |
1306 | ls('-l'); |
1307 | ||
19799a22 GS |
1308 | In fact, if you predeclare functions you want to call that way, you don't |
1309 | even need parentheses: | |
cb1a09d0 AD |
1310 | |
1311 | use subs qw(date who ls); | |
1312 | date; | |
1313 | who "am", "i"; | |
1314 | ls -l; | |
1315 | ||
1316 | A more complete example of this is the standard Shell module, which | |
19799a22 | 1317 | can treat undefined subroutine calls as calls to external programs. |
a0d0e21e | 1318 | |
19799a22 GS |
1319 | Mechanisms are available to help modules writers split their modules |
1320 | into autoloadable files. See the standard AutoLoader module | |
6d28dffb | 1321 | described in L<AutoLoader> and in L<AutoSplit>, the standard |
1322 | SelfLoader modules in L<SelfLoader>, and the document on adding C | |
19799a22 | 1323 | functions to Perl code in L<perlxs>. |
cb1a09d0 | 1324 | |
09bef843 SB |
1325 | =head2 Subroutine Attributes |
1326 | ||
1327 | A subroutine declaration or definition may have a list of attributes | |
1328 | associated with it. If such an attribute list is present, it is | |
0120eecf | 1329 | broken up at space or colon boundaries and treated as though a |
09bef843 SB |
1330 | C<use attributes> had been seen. See L<attributes> for details |
1331 | about what attributes are currently supported. | |
1332 | Unlike the limitation with the obsolescent C<use attrs>, the | |
1333 | C<sub : ATTRLIST> syntax works to associate the attributes with | |
1334 | a pre-declaration, and not just with a subroutine definition. | |
1335 | ||
1336 | The attributes must be valid as simple identifier names (without any | |
1337 | punctuation other than the '_' character). They may have a parameter | |
1338 | list appended, which is only checked for whether its parentheses ('(',')') | |
1339 | nest properly. | |
1340 | ||
1341 | Examples of valid syntax (even though the attributes are unknown): | |
1342 | ||
0120eecf GS |
1343 | sub fnord (&\%) : switch(10,foo(7,3)) : expensive ; |
1344 | sub plugh () : Ugly('\(") :Bad ; | |
09bef843 SB |
1345 | sub xyzzy : _5x5 { ... } |
1346 | ||
1347 | Examples of invalid syntax: | |
1348 | ||
1349 | sub fnord : switch(10,foo() ; # ()-string not balanced | |
1350 | sub snoid : Ugly('(') ; # ()-string not balanced | |
1351 | sub xyzzy : 5x5 ; # "5x5" not a valid identifier | |
1352 | sub plugh : Y2::north ; # "Y2::north" not a simple identifier | |
0120eecf | 1353 | sub snurt : foo + bar ; # "+" not a colon or space |
09bef843 SB |
1354 | |
1355 | The attribute list is passed as a list of constant strings to the code | |
1356 | which associates them with the subroutine. In particular, the second example | |
1357 | of valid syntax above currently looks like this in terms of how it's | |
1358 | parsed and invoked: | |
1359 | ||
1360 | use attributes __PACKAGE__, \&plugh, q[Ugly('\(")], 'Bad'; | |
1361 | ||
1362 | For further details on attribute lists and their manipulation, | |
a0ae32d3 | 1363 | see L<attributes> and L<Attribute::Handlers>. |
09bef843 | 1364 | |
cb1a09d0 | 1365 | =head1 SEE ALSO |
a0d0e21e | 1366 | |
19799a22 GS |
1367 | See L<perlref/"Function Templates"> for more about references and closures. |
1368 | See L<perlxs> if you'd like to learn about calling C subroutines from Perl. | |
a2293a43 | 1369 | See L<perlembed> if you'd like to learn about calling Perl subroutines from C. |
19799a22 GS |
1370 | See L<perlmod> to learn about bundling up your functions in separate files. |
1371 | See L<perlmodlib> to learn what library modules come standard on your system. | |
1372 | See L<perltoot> to learn how to make object method calls. |