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