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 | |
17 | $subref = sub BLOCK; |
18 | |
a0d0e21e |
19 | To import subroutines: |
20 | |
21 | use PACKAGE qw(NAME1 NAME2 NAME3); |
22 | |
23 | To call subroutines: |
24 | |
5f05dabc |
25 | NAME(LIST); # & is optional with parentheses. |
26 | NAME LIST; # Parentheses optional if pre-declared/imported. |
cb1a09d0 |
27 | &NAME; # Passes current @_ to subroutine. |
a0d0e21e |
28 | |
29 | =head1 DESCRIPTION |
30 | |
cb1a09d0 |
31 | Like many languages, Perl provides for user-defined subroutines. These |
32 | may be located anywhere in the main program, loaded in from other files |
33 | via the C<do>, C<require>, or C<use> keywords, or even generated on the |
34 | fly using C<eval> or anonymous subroutines (closures). You can even call |
c07a80fd |
35 | a function indirectly using a variable containing its name or a CODE reference |
36 | to it, as in C<$var = \&function>. |
cb1a09d0 |
37 | |
38 | The Perl model for function call and return values is simple: all |
39 | functions are passed as parameters one single flat list of scalars, and |
40 | all functions likewise return to their caller one single flat list of |
41 | scalars. Any arrays or hashes in these call and return lists will |
42 | collapse, losing their identities--but you may always use |
43 | pass-by-reference instead to avoid this. Both call and return lists may |
44 | contain as many or as few scalar elements as you'd like. (Often a |
45 | function without an explicit return statement is called a subroutine, but |
46 | there's really no difference from the language's perspective.) |
47 | |
48 | Any arguments passed to the routine come in as the array @_. Thus if you |
49 | called a function with two arguments, those would be stored in C<$_[0]> |
50 | and C<$_[1]>. The array @_ is a local array, but its values are implicit |
dfdcb7a0 |
51 | references (predating L<perlref>) to the actual scalar parameters. What |
52 | this means in practice is that when you explicitly modify C<$_[0]> et al., |
53 | you will be changing the actual arguments. As a result, all arguments |
54 | to functions are treated as lvalues. Any hash or array elements that are |
55 | passed to functions will get created if they do not exist (irrespective |
56 | of whether the function does modify the contents of C<@_>). This is |
57 | frequently a source of surprise. See L<perltrap> for an example. |
58 | |
59 | The return value of the subroutine is the value of the last expression |
cb1a09d0 |
60 | evaluated. Alternatively, a return statement may be used to specify the |
61 | returned value and exit the subroutine. If you return one or more arrays |
62 | and/or hashes, these will be flattened together into one large |
63 | indistinguishable list. |
64 | |
65 | Perl does not have named formal parameters, but in practice all you do is |
66 | assign to a my() list of these. Any variables you use in the function |
67 | that aren't declared private are global variables. For the gory details |
1fef88e7 |
68 | on creating private variables, see |
6d28dffb |
69 | L<"Private Variables via my()"> and L<"Temporary Values via local()">. |
70 | To create protected environments for a set of functions in a separate |
71 | package (and probably a separate file), see L<perlmod/"Packages">. |
a0d0e21e |
72 | |
73 | Example: |
74 | |
cb1a09d0 |
75 | sub max { |
76 | my $max = shift(@_); |
a0d0e21e |
77 | foreach $foo (@_) { |
78 | $max = $foo if $max < $foo; |
79 | } |
cb1a09d0 |
80 | return $max; |
a0d0e21e |
81 | } |
cb1a09d0 |
82 | $bestday = max($mon,$tue,$wed,$thu,$fri); |
a0d0e21e |
83 | |
84 | Example: |
85 | |
86 | # get a line, combining continuation lines |
87 | # that start with whitespace |
88 | |
89 | sub get_line { |
cb1a09d0 |
90 | $thisline = $lookahead; # GLOBAL VARIABLES!! |
a0d0e21e |
91 | LINE: while ($lookahead = <STDIN>) { |
92 | if ($lookahead =~ /^[ \t]/) { |
93 | $thisline .= $lookahead; |
94 | } |
95 | else { |
96 | last LINE; |
97 | } |
98 | } |
99 | $thisline; |
100 | } |
101 | |
102 | $lookahead = <STDIN>; # get first line |
103 | while ($_ = get_line()) { |
104 | ... |
105 | } |
106 | |
107 | Use array assignment to a local list to name your formal arguments: |
108 | |
109 | sub maybeset { |
110 | my($key, $value) = @_; |
cb1a09d0 |
111 | $Foo{$key} = $value unless $Foo{$key}; |
a0d0e21e |
112 | } |
113 | |
cb1a09d0 |
114 | This also has the effect of turning call-by-reference into call-by-value, |
5f05dabc |
115 | because the assignment copies the values. Otherwise a function is free to |
1fef88e7 |
116 | do in-place modifications of @_ and change its caller's values. |
cb1a09d0 |
117 | |
118 | upcase_in($v1, $v2); # this changes $v1 and $v2 |
119 | sub upcase_in { |
120 | for (@_) { tr/a-z/A-Z/ } |
121 | } |
122 | |
123 | You aren't allowed to modify constants in this way, of course. If an |
124 | argument were actually literal and you tried to change it, you'd take a |
125 | (presumably fatal) exception. For example, this won't work: |
126 | |
127 | upcase_in("frederick"); |
128 | |
129 | It would be much safer if the upcase_in() function |
130 | were written to return a copy of its parameters instead |
131 | of changing them in place: |
132 | |
133 | ($v3, $v4) = upcase($v1, $v2); # this doesn't |
134 | sub upcase { |
135 | my @parms = @_; |
136 | for (@parms) { tr/a-z/A-Z/ } |
c07a80fd |
137 | # wantarray checks if we were called in list context |
138 | return wantarray ? @parms : $parms[0]; |
cb1a09d0 |
139 | } |
140 | |
141 | Notice how this (unprototyped) function doesn't care whether it was passed |
142 | real scalars or arrays. Perl will see everything as one big long flat @_ |
143 | parameter list. This is one of the ways where Perl's simple |
144 | argument-passing style shines. The upcase() function would work perfectly |
145 | well without changing the upcase() definition even if we fed it things |
146 | like this: |
147 | |
148 | @newlist = upcase(@list1, @list2); |
149 | @newlist = upcase( split /:/, $var ); |
150 | |
151 | Do not, however, be tempted to do this: |
152 | |
153 | (@a, @b) = upcase(@list1, @list2); |
154 | |
155 | Because like its flat incoming parameter list, the return list is also |
156 | flat. So all you have managed to do here is stored everything in @a and |
157 | made @b an empty list. See L</"Pass by Reference"> for alternatives. |
158 | |
5f05dabc |
159 | A subroutine may be called using the "&" prefix. The "&" is optional |
160 | in modern Perls, and so are the parentheses if the subroutine has been |
161 | pre-declared. (Note, however, that the "&" is I<NOT> optional when |
162 | you're just naming the subroutine, such as when it's used as an |
163 | argument to defined() or undef(). Nor is it optional when you want to |
164 | do an indirect subroutine call with a subroutine name or reference |
165 | using the C<&$subref()> or C<&{$subref}()> constructs. See L<perlref> |
166 | for more on that.) |
a0d0e21e |
167 | |
168 | Subroutines may be called recursively. If a subroutine is called using |
cb1a09d0 |
169 | the "&" form, the argument list is optional, and if omitted, no @_ array is |
170 | set up for the subroutine: the @_ array at the time of the call is |
171 | visible to subroutine instead. This is an efficiency mechanism that |
172 | new users may wish to avoid. |
a0d0e21e |
173 | |
174 | &foo(1,2,3); # pass three arguments |
175 | foo(1,2,3); # the same |
176 | |
177 | foo(); # pass a null list |
178 | &foo(); # the same |
a0d0e21e |
179 | |
cb1a09d0 |
180 | &foo; # foo() get current args, like foo(@_) !! |
181 | foo; # like foo() IFF sub foo pre-declared, else "foo" |
182 | |
c07a80fd |
183 | Not only does the "&" form make the argument list optional, but it also |
184 | disables any prototype checking on the arguments you do provide. This |
185 | is partly for historical reasons, and partly for having a convenient way |
186 | to cheat if you know what you're doing. See the section on Prototypes below. |
187 | |
cb1a09d0 |
188 | =head2 Private Variables via my() |
189 | |
190 | Synopsis: |
191 | |
192 | my $foo; # declare $foo lexically local |
193 | my (@wid, %get); # declare list of variables local |
194 | my $foo = "flurp"; # declare $foo lexical, and init it |
195 | my @oof = @bar; # declare @oof lexical, and init it |
196 | |
197 | A "my" declares the listed variables to be confined (lexically) to the |
55497cff |
198 | enclosing block, conditional (C<if/unless/elsif/else>), loop |
199 | (C<for/foreach/while/until/continue>), subroutine, C<eval>, or |
200 | C<do/require/use>'d file. If more than one value is listed, the list |
5f05dabc |
201 | must be placed in parentheses. All listed elements must be legal lvalues. |
55497cff |
202 | Only alphanumeric identifiers may be lexically scoped--magical |
203 | builtins like $/ must currently be localized with "local" instead. |
cb1a09d0 |
204 | |
205 | Unlike dynamic variables created by the "local" statement, lexical |
206 | variables declared with "my" are totally hidden from the outside world, |
207 | including any called subroutines (even if it's the same subroutine called |
208 | from itself or elsewhere--every call gets its own copy). |
209 | |
210 | (An eval(), however, can see the lexical variables of the scope it is |
211 | being evaluated in so long as the names aren't hidden by declarations within |
212 | the eval() itself. See L<perlref>.) |
213 | |
214 | The parameter list to my() may be assigned to if desired, which allows you |
215 | to initialize your variables. (If no initializer is given for a |
216 | particular variable, it is created with the undefined value.) Commonly |
217 | this is used to name the parameters to a subroutine. Examples: |
218 | |
219 | $arg = "fred"; # "global" variable |
220 | $n = cube_root(27); |
221 | print "$arg thinks the root is $n\n"; |
222 | fred thinks the root is 3 |
223 | |
224 | sub cube_root { |
225 | my $arg = shift; # name doesn't matter |
226 | $arg **= 1/3; |
227 | return $arg; |
228 | } |
229 | |
230 | The "my" is simply a modifier on something you might assign to. So when |
231 | you do assign to the variables in its argument list, the "my" doesn't |
232 | change whether those variables is viewed as a scalar or an array. So |
233 | |
234 | my ($foo) = <STDIN>; |
235 | my @FOO = <STDIN>; |
236 | |
5f05dabc |
237 | both supply a list context to the right-hand side, while |
cb1a09d0 |
238 | |
239 | my $foo = <STDIN>; |
240 | |
5f05dabc |
241 | supplies a scalar context. But the following declares only one variable: |
748a9306 |
242 | |
cb1a09d0 |
243 | my $foo, $bar = 1; |
748a9306 |
244 | |
cb1a09d0 |
245 | That has the same effect as |
748a9306 |
246 | |
cb1a09d0 |
247 | my $foo; |
248 | $bar = 1; |
a0d0e21e |
249 | |
cb1a09d0 |
250 | The declared variable is not introduced (is not visible) until after |
251 | the current statement. Thus, |
252 | |
253 | my $x = $x; |
254 | |
255 | can be used to initialize the new $x with the value of the old $x, and |
256 | the expression |
257 | |
258 | my $x = 123 and $x == 123 |
259 | |
260 | is false unless the old $x happened to have the value 123. |
261 | |
55497cff |
262 | Lexical scopes of control structures are not bounded precisely by the |
263 | braces that delimit their controlled blocks; control expressions are |
264 | part of the scope, too. Thus in the loop |
265 | |
266 | while (my $line = <>) { |
267 | $line = lc $line; |
268 | } continue { |
269 | print $line; |
270 | } |
271 | |
272 | the scope of $line extends from its declaration throughout the rest of |
273 | the loop construct (including the C<continue> clause), but not beyond |
274 | it. Similarly, in the conditional |
275 | |
276 | if ((my $answer = <STDIN>) =~ /^yes$/i) { |
277 | user_agrees(); |
278 | } elsif ($answer =~ /^no$/i) { |
279 | user_disagrees(); |
280 | } else { |
281 | chomp $answer; |
282 | die "'$answer' is neither 'yes' nor 'no'"; |
283 | } |
284 | |
285 | the scope of $answer extends from its declaration throughout the rest |
286 | of the conditional (including C<elsif> and C<else> clauses, if any), |
287 | but not beyond it. |
288 | |
289 | (None of the foregoing applies to C<if/unless> or C<while/until> |
290 | modifiers appended to simple statements. Such modifiers are not |
291 | control structures and have no effect on scoping.) |
292 | |
5f05dabc |
293 | The C<foreach> loop defaults to scoping its index variable dynamically |
55497cff |
294 | (in the manner of C<local>; see below). However, if the index |
295 | variable is prefixed with the keyword "my", then it is lexically |
296 | scoped instead. Thus in the loop |
297 | |
298 | for my $i (1, 2, 3) { |
299 | some_function(); |
300 | } |
301 | |
302 | the scope of $i extends to the end of the loop, but not beyond it, and |
303 | so the value of $i is unavailable in some_function(). |
304 | |
cb1a09d0 |
305 | Some users may wish to encourage the use of lexically scoped variables. |
306 | As an aid to catching implicit references to package variables, |
307 | if you say |
308 | |
309 | use strict 'vars'; |
310 | |
311 | then any variable reference from there to the end of the enclosing |
312 | block must either refer to a lexical variable, or must be fully |
313 | qualified with the package name. A compilation error results |
314 | otherwise. An inner block may countermand this with S<"no strict 'vars'">. |
315 | |
316 | A my() has both a compile-time and a run-time effect. At compile time, |
317 | the compiler takes notice of it; the principle usefulness of this is to |
318 | quiet C<use strict 'vars'>. The actual initialization doesn't happen |
319 | until run time, so gets executed every time through a loop. |
320 | |
321 | Variables declared with "my" are not part of any package and are therefore |
322 | never fully qualified with the package name. In particular, you're not |
323 | allowed to try to make a package variable (or other global) lexical: |
324 | |
325 | my $pack::var; # ERROR! Illegal syntax |
326 | my $_; # also illegal (currently) |
327 | |
328 | In fact, a dynamic variable (also known as package or global variables) |
329 | are still accessible using the fully qualified :: notation even while a |
330 | lexical of the same name is also visible: |
331 | |
332 | package main; |
333 | local $x = 10; |
334 | my $x = 20; |
335 | print "$x and $::x\n"; |
336 | |
337 | That will print out 20 and 10. |
338 | |
5f05dabc |
339 | You may declare "my" variables at the outermost scope of a file to |
340 | hide any such identifiers totally from the outside world. This is similar |
6d28dffb |
341 | to C's static variables at the file level. To do this with a subroutine |
cb1a09d0 |
342 | requires the use of a closure (anonymous function). If a block (such as |
343 | an eval(), function, or C<package>) wants to create a private subroutine |
344 | that cannot be called from outside that block, it can declare a lexical |
345 | variable containing an anonymous sub reference: |
346 | |
347 | my $secret_version = '1.001-beta'; |
348 | my $secret_sub = sub { print $secret_version }; |
349 | &$secret_sub(); |
350 | |
351 | As long as the reference is never returned by any function within the |
5f05dabc |
352 | module, no outside module can see the subroutine, because its name is not in |
cb1a09d0 |
353 | any package's symbol table. Remember that it's not I<REALLY> called |
354 | $some_pack::secret_version or anything; it's just $secret_version, |
355 | unqualified and unqualifiable. |
356 | |
357 | This does not work with object methods, however; all object methods have |
358 | to be in the symbol table of some package to be found. |
359 | |
360 | Just because the lexical variable is lexically (also called statically) |
361 | scoped doesn't mean that within a function it works like a C static. It |
362 | normally works more like a C auto. But here's a mechanism for giving a |
363 | function private variables with both lexical scoping and a static |
364 | lifetime. If you do want to create something like C's static variables, |
365 | just enclose the whole function in an extra block, and put the |
366 | static variable outside the function but in the block. |
367 | |
368 | { |
369 | my $secret_val = 0; |
370 | sub gimme_another { |
371 | return ++$secret_val; |
372 | } |
373 | } |
374 | # $secret_val now becomes unreachable by the outside |
375 | # world, but retains its value between calls to gimme_another |
376 | |
377 | If this function is being sourced in from a separate file |
378 | via C<require> or C<use>, then this is probably just fine. If it's |
379 | all in the main program, you'll need to arrange for the my() |
380 | to be executed early, either by putting the whole block above |
5f05dabc |
381 | your pain program, or more likely, placing merely a BEGIN |
cb1a09d0 |
382 | sub around it to make sure it gets executed before your program |
383 | starts to run: |
384 | |
385 | sub BEGIN { |
386 | my $secret_val = 0; |
387 | sub gimme_another { |
388 | return ++$secret_val; |
389 | } |
390 | } |
391 | |
392 | See L<perlrun> about the BEGIN function. |
393 | |
394 | =head2 Temporary Values via local() |
395 | |
396 | B<NOTE>: In general, you should be using "my" instead of "local", because |
6d28dffb |
397 | it's faster and safer. Exceptions to this include the global punctuation |
cb1a09d0 |
398 | variables, filehandles and formats, and direct manipulation of the Perl |
399 | symbol table itself. Format variables often use "local" though, as do |
400 | other variables whose current value must be visible to called |
401 | subroutines. |
402 | |
403 | Synopsis: |
404 | |
405 | local $foo; # declare $foo dynamically local |
406 | local (@wid, %get); # declare list of variables local |
407 | local $foo = "flurp"; # declare $foo dynamic, and init it |
408 | local @oof = @bar; # declare @oof dynamic, and init it |
409 | |
410 | local *FH; # localize $FH, @FH, %FH, &FH ... |
411 | local *merlyn = *randal; # now $merlyn is really $randal, plus |
412 | # @merlyn is really @randal, etc |
413 | local *merlyn = 'randal'; # SAME THING: promote 'randal' to *randal |
414 | local *merlyn = \$randal; # just alias $merlyn, not @merlyn etc |
415 | |
416 | A local() modifies its listed variables to be local to the enclosing |
5f05dabc |
417 | block, (or subroutine, C<eval{}>, or C<do>) and I<any called from |
cb1a09d0 |
418 | within that block>. A local() just gives temporary values to global |
419 | (meaning package) variables. This is known as dynamic scoping. Lexical |
420 | scoping is done with "my", which works more like C's auto declarations. |
421 | |
422 | If more than one variable is given to local(), they must be placed in |
5f05dabc |
423 | parentheses. All listed elements must be legal lvalues. This operator works |
cb1a09d0 |
424 | by saving the current values of those variables in its argument list on a |
5f05dabc |
425 | hidden stack and restoring them upon exiting the block, subroutine, or |
cb1a09d0 |
426 | eval. This means that called subroutines can also reference the local |
427 | variable, but not the global one. The argument list may be assigned to if |
428 | desired, which allows you to initialize your local variables. (If no |
429 | initializer is given for a particular variable, it is created with an |
430 | undefined value.) Commonly this is used to name the parameters to a |
431 | subroutine. Examples: |
432 | |
433 | for $i ( 0 .. 9 ) { |
434 | $digits{$i} = $i; |
435 | } |
436 | # assume this function uses global %digits hash |
437 | parse_num(); |
438 | |
439 | # now temporarily add to %digits hash |
440 | if ($base12) { |
441 | # (NOTE: not claiming this is efficient!) |
442 | local %digits = (%digits, 't' => 10, 'e' => 11); |
443 | parse_num(); # parse_num gets this new %digits! |
444 | } |
445 | # old %digits restored here |
446 | |
1fef88e7 |
447 | Because local() is a run-time command, it gets executed every time |
cb1a09d0 |
448 | through a loop. In releases of Perl previous to 5.0, this used more stack |
449 | storage each time until the loop was exited. Perl now reclaims the space |
450 | each time through, but it's still more efficient to declare your variables |
451 | outside the loop. |
452 | |
453 | A local is simply a modifier on an lvalue expression. When you assign to |
454 | a localized variable, the local doesn't change whether its list is viewed |
455 | as a scalar or an array. So |
456 | |
457 | local($foo) = <STDIN>; |
458 | local @FOO = <STDIN>; |
459 | |
5f05dabc |
460 | both supply a list context to the right-hand side, while |
cb1a09d0 |
461 | |
462 | local $foo = <STDIN>; |
463 | |
464 | supplies a scalar context. |
465 | |
466 | =head2 Passing Symbol Table Entries (typeglobs) |
467 | |
468 | [Note: The mechanism described in this section was originally the only |
469 | way to simulate pass-by-reference in older versions of Perl. While it |
470 | still works fine in modern versions, the new reference mechanism is |
471 | generally easier to work with. See below.] |
a0d0e21e |
472 | |
473 | Sometimes you don't want to pass the value of an array to a subroutine |
474 | but rather the name of it, so that the subroutine can modify the global |
475 | copy of it rather than working with a local copy. In perl you can |
cb1a09d0 |
476 | refer to all objects of a particular name by prefixing the name |
5f05dabc |
477 | with a star: C<*foo>. This is often known as a "typeglob", because the |
a0d0e21e |
478 | star on the front can be thought of as a wildcard match for all the |
479 | funny prefix characters on variables and subroutines and such. |
480 | |
55497cff |
481 | When evaluated, the typeglob produces a scalar value that represents |
5f05dabc |
482 | all the objects of that name, including any filehandle, format, or |
a0d0e21e |
483 | subroutine. When assigned to, it causes the name mentioned to refer to |
484 | whatever "*" value was assigned to it. Example: |
485 | |
486 | sub doubleary { |
487 | local(*someary) = @_; |
488 | foreach $elem (@someary) { |
489 | $elem *= 2; |
490 | } |
491 | } |
492 | doubleary(*foo); |
493 | doubleary(*bar); |
494 | |
495 | Note that scalars are already passed by reference, so you can modify |
496 | scalar arguments without using this mechanism by referring explicitly |
1fef88e7 |
497 | to C<$_[0]> etc. You can modify all the elements of an array by passing |
a0d0e21e |
498 | all the elements as scalars, but you have to use the * mechanism (or |
5f05dabc |
499 | the equivalent reference mechanism) to push, pop, or change the size of |
a0d0e21e |
500 | an array. It will certainly be faster to pass the typeglob (or reference). |
501 | |
502 | Even if you don't want to modify an array, this mechanism is useful for |
5f05dabc |
503 | passing multiple arrays in a single LIST, because normally the LIST |
a0d0e21e |
504 | mechanism will merge all the array values so that you can't extract out |
55497cff |
505 | the individual arrays. For more on typeglobs, see |
506 | L<perldata/"Typeglobs and FileHandles">. |
cb1a09d0 |
507 | |
508 | =head2 Pass by Reference |
509 | |
55497cff |
510 | If you want to pass more than one array or hash into a function--or |
511 | return them from it--and have them maintain their integrity, then |
512 | you're going to have to use an explicit pass-by-reference. Before you |
513 | do that, you need to understand references as detailed in L<perlref>. |
c07a80fd |
514 | This section may not make much sense to you otherwise. |
cb1a09d0 |
515 | |
516 | Here are a few simple examples. First, let's pass in several |
517 | arrays to a function and have it pop all of then, return a new |
518 | list of all their former last elements: |
519 | |
520 | @tailings = popmany ( \@a, \@b, \@c, \@d ); |
521 | |
522 | sub popmany { |
523 | my $aref; |
524 | my @retlist = (); |
525 | foreach $aref ( @_ ) { |
526 | push @retlist, pop @$aref; |
527 | } |
528 | return @retlist; |
529 | } |
530 | |
531 | Here's how you might write a function that returns a |
532 | list of keys occurring in all the hashes passed to it: |
533 | |
534 | @common = inter( \%foo, \%bar, \%joe ); |
535 | sub inter { |
536 | my ($k, $href, %seen); # locals |
537 | foreach $href (@_) { |
538 | while ( $k = each %$href ) { |
539 | $seen{$k}++; |
540 | } |
541 | } |
542 | return grep { $seen{$_} == @_ } keys %seen; |
543 | } |
544 | |
5f05dabc |
545 | So far, we're using just the normal list return mechanism. |
cb1a09d0 |
546 | What happens if you want to pass or return a hash? Well, |
5f05dabc |
547 | if you're using only one of them, or you don't mind them |
cb1a09d0 |
548 | concatenating, then the normal calling convention is ok, although |
549 | a little expensive. |
550 | |
551 | Where people get into trouble is here: |
552 | |
553 | (@a, @b) = func(@c, @d); |
554 | or |
555 | (%a, %b) = func(%c, %d); |
556 | |
5f05dabc |
557 | That syntax simply won't work. It sets just @a or %a and clears the @b or |
cb1a09d0 |
558 | %b. Plus the function didn't get passed into two separate arrays or |
559 | hashes: it got one long list in @_, as always. |
560 | |
561 | If you can arrange for everyone to deal with this through references, it's |
562 | cleaner code, although not so nice to look at. Here's a function that |
563 | takes two array references as arguments, returning the two array elements |
564 | in order of how many elements they have in them: |
565 | |
566 | ($aref, $bref) = func(\@c, \@d); |
567 | print "@$aref has more than @$bref\n"; |
568 | sub func { |
569 | my ($cref, $dref) = @_; |
570 | if (@$cref > @$dref) { |
571 | return ($cref, $dref); |
572 | } else { |
c07a80fd |
573 | return ($dref, $cref); |
cb1a09d0 |
574 | } |
575 | } |
576 | |
577 | It turns out that you can actually do this also: |
578 | |
579 | (*a, *b) = func(\@c, \@d); |
580 | print "@a has more than @b\n"; |
581 | sub func { |
582 | local (*c, *d) = @_; |
583 | if (@c > @d) { |
584 | return (\@c, \@d); |
585 | } else { |
586 | return (\@d, \@c); |
587 | } |
588 | } |
589 | |
590 | Here we're using the typeglobs to do symbol table aliasing. It's |
591 | a tad subtle, though, and also won't work if you're using my() |
5f05dabc |
592 | variables, because only globals (well, and local()s) are in the symbol table. |
593 | |
594 | If you're passing around filehandles, you could usually just use the bare |
595 | typeglob, like *STDOUT, but typeglobs references would be better because |
596 | they'll still work properly under C<use strict 'refs'>. For example: |
597 | |
598 | splutter(\*STDOUT); |
599 | sub splutter { |
600 | my $fh = shift; |
601 | print $fh "her um well a hmmm\n"; |
602 | } |
603 | |
604 | $rec = get_rec(\*STDIN); |
605 | sub get_rec { |
606 | my $fh = shift; |
607 | return scalar <$fh>; |
608 | } |
609 | |
610 | Another way to do this is using *HANDLE{IO}, see L<perlref> for usage |
611 | and caveats. |
612 | |
613 | If you're planning on generating new filehandles, you could do this: |
614 | |
615 | sub openit { |
616 | my $name = shift; |
617 | local *FH; |
e05a3a1e |
618 | return open (FH, $path) ? *FH : undef; |
5f05dabc |
619 | } |
620 | |
621 | Although that will actually produce a small memory leak. See the bottom |
622 | of L<perlfunc/open()> for a somewhat cleaner way using the IO::Handle |
623 | package. |
cb1a09d0 |
624 | |
cb1a09d0 |
625 | =head2 Prototypes |
626 | |
627 | As of the 5.002 release of perl, if you declare |
628 | |
629 | sub mypush (\@@) |
630 | |
c07a80fd |
631 | then mypush() takes arguments exactly like push() does. The declaration |
632 | of the function to be called must be visible at compile time. The prototype |
5f05dabc |
633 | affects only the interpretation of new-style calls to the function, where |
c07a80fd |
634 | new-style is defined as not using the C<&> character. In other words, |
635 | if you call it like a builtin function, then it behaves like a builtin |
636 | function. If you call it like an old-fashioned subroutine, then it |
637 | behaves like an old-fashioned subroutine. It naturally falls out from |
638 | this rule that prototypes have no influence on subroutine references |
639 | like C<\&foo> or on indirect subroutine calls like C<&{$subref}>. |
640 | |
641 | Method calls are not influenced by prototypes either, because the |
5f05dabc |
642 | function to be called is indeterminate at compile time, because it depends |
c07a80fd |
643 | on inheritance. |
cb1a09d0 |
644 | |
5f05dabc |
645 | Because the intent is primarily to let you define subroutines that work |
c07a80fd |
646 | like builtin commands, here are the prototypes for some other functions |
647 | that parse almost exactly like the corresponding builtins. |
cb1a09d0 |
648 | |
649 | Declared as Called as |
650 | |
651 | sub mylink ($$) mylink $old, $new |
652 | sub myvec ($$$) myvec $var, $offset, 1 |
653 | sub myindex ($$;$) myindex &getstring, "substr" |
654 | sub mysyswrite ($$$;$) mysyswrite $buf, 0, length($buf) - $off, $off |
655 | sub myreverse (@) myreverse $a,$b,$c |
656 | sub myjoin ($@) myjoin ":",$a,$b,$c |
657 | sub mypop (\@) mypop @array |
658 | sub mysplice (\@$$@) mysplice @array,@array,0,@pushme |
659 | sub mykeys (\%) mykeys %{$hashref} |
660 | sub myopen (*;$) myopen HANDLE, $name |
661 | sub mypipe (**) mypipe READHANDLE, WRITEHANDLE |
662 | sub mygrep (&@) mygrep { /foo/ } $a,$b,$c |
663 | sub myrand ($) myrand 42 |
664 | sub mytime () mytime |
665 | |
c07a80fd |
666 | Any backslashed prototype character represents an actual argument |
6e47f808 |
667 | that absolutely must start with that character. The value passed |
668 | to the subroutine (as part of C<@_>) will be a reference to the |
669 | actual argument given in the subroutine call, obtained by applying |
670 | C<\> to that argument. |
c07a80fd |
671 | |
672 | Unbackslashed prototype characters have special meanings. Any |
673 | unbackslashed @ or % eats all the rest of the arguments, and forces |
674 | list context. An argument represented by $ forces scalar context. An |
675 | & requires an anonymous subroutine, which, if passed as the first |
676 | argument, does not require the "sub" keyword or a subsequent comma. A |
677 | * does whatever it has to do to turn the argument into a reference to a |
678 | symbol table entry. |
679 | |
680 | A semicolon separates mandatory arguments from optional arguments. |
681 | (It is redundant before @ or %.) |
cb1a09d0 |
682 | |
c07a80fd |
683 | Note how the last three examples above are treated specially by the parser. |
cb1a09d0 |
684 | mygrep() is parsed as a true list operator, myrand() is parsed as a |
685 | true unary operator with unary precedence the same as rand(), and |
5f05dabc |
686 | mytime() is truly without arguments, just like time(). That is, if you |
cb1a09d0 |
687 | say |
688 | |
689 | mytime +2; |
690 | |
691 | you'll get mytime() + 2, not mytime(2), which is how it would be parsed |
692 | without the prototype. |
693 | |
694 | The interesting thing about & is that you can generate new syntax with it: |
695 | |
6d28dffb |
696 | sub try (&@) { |
cb1a09d0 |
697 | my($try,$catch) = @_; |
698 | eval { &$try }; |
699 | if ($@) { |
700 | local $_ = $@; |
701 | &$catch; |
702 | } |
703 | } |
55497cff |
704 | sub catch (&) { $_[0] } |
cb1a09d0 |
705 | |
706 | try { |
707 | die "phooey"; |
708 | } catch { |
709 | /phooey/ and print "unphooey\n"; |
710 | }; |
711 | |
712 | That prints "unphooey". (Yes, there are still unresolved |
713 | issues having to do with the visibility of @_. I'm ignoring that |
714 | question for the moment. (But note that if we make @_ lexically |
715 | scoped, those anonymous subroutines can act like closures... (Gee, |
5f05dabc |
716 | is this sounding a little Lispish? (Never mind.)))) |
cb1a09d0 |
717 | |
718 | And here's a reimplementation of grep: |
719 | |
720 | sub mygrep (&@) { |
721 | my $code = shift; |
722 | my @result; |
723 | foreach $_ (@_) { |
6e47f808 |
724 | push(@result, $_) if &$code; |
cb1a09d0 |
725 | } |
726 | @result; |
727 | } |
a0d0e21e |
728 | |
cb1a09d0 |
729 | Some folks would prefer full alphanumeric prototypes. Alphanumerics have |
730 | been intentionally left out of prototypes for the express purpose of |
731 | someday in the future adding named, formal parameters. The current |
732 | mechanism's main goal is to let module writers provide better diagnostics |
733 | for module users. Larry feels the notation quite understandable to Perl |
734 | programmers, and that it will not intrude greatly upon the meat of the |
735 | module, nor make it harder to read. The line noise is visually |
736 | encapsulated into a small pill that's easy to swallow. |
737 | |
738 | It's probably best to prototype new functions, not retrofit prototyping |
739 | into older ones. That's because you must be especially careful about |
740 | silent impositions of differing list versus scalar contexts. For example, |
741 | if you decide that a function should take just one parameter, like this: |
742 | |
743 | sub func ($) { |
744 | my $n = shift; |
745 | print "you gave me $n\n"; |
746 | } |
747 | |
748 | and someone has been calling it with an array or expression |
749 | returning a list: |
750 | |
751 | func(@foo); |
752 | func( split /:/ ); |
753 | |
754 | Then you've just supplied an automatic scalar() in front of their |
755 | argument, which can be more than a bit surprising. The old @foo |
756 | which used to hold one thing doesn't get passed in. Instead, |
5f05dabc |
757 | the func() now gets passed in 1, that is, the number of elements |
cb1a09d0 |
758 | in @foo. And the split() gets called in a scalar context and |
759 | starts scribbling on your @_ parameter list. |
760 | |
5f05dabc |
761 | This is all very powerful, of course, and should be used only in moderation |
cb1a09d0 |
762 | to make the world a better place. |
44a8e56a |
763 | |
764 | =head2 Constant Functions |
765 | |
766 | Functions with a prototype of C<()> are potential candidates for |
767 | inlining. If the result after optimization and constant folding is a |
768 | constant then it will be used in place of new-style calls to the |
769 | function. Old-style calls (that is, calls made using C<&>) are not |
770 | affected. |
771 | |
772 | All of the following functions would be inlined. |
773 | |
774 | sub PI () { 3.14159 } |
775 | sub ST_DEV () { 0 } |
776 | sub ST_INO () { 1 } |
777 | |
778 | sub FLAG_FOO () { 1 << 8 } |
779 | sub FLAG_BAR () { 1 << 9 } |
780 | sub FLAG_MASK () { FLAG_FOO | FLAG_BAR } |
781 | |
782 | sub OPT_BAZ () { 1 } |
783 | sub BAZ_VAL () { |
784 | if (OPT_BAZ) { |
785 | return 23; |
786 | } |
787 | else { |
788 | return 42; |
789 | } |
790 | } |
cb1a09d0 |
791 | |
4cee8e80 |
792 | If you redefine a subroutine which was eligible for inlining you'll get |
793 | a mandatory warning. (You can use this warning to tell whether or not a |
794 | particular subroutine is considered constant.) The warning is |
795 | considered severe enough not to be optional because previously compiled |
796 | invocations of the function will still be using the old value of the |
797 | function. If you need to be able to redefine the subroutine you need to |
798 | ensure that it isn't inlined, either by dropping the C<()> prototype |
799 | (which changes the calling semantics, so beware) or by thwarting the |
800 | inlining mechanism in some other way, such as |
801 | |
802 | my $dummy; |
803 | sub not_inlined () { |
804 | $dummy || 23 |
805 | } |
806 | |
cb1a09d0 |
807 | =head2 Overriding Builtin Functions |
a0d0e21e |
808 | |
5f05dabc |
809 | Many builtin functions may be overridden, though this should be tried |
810 | only occasionally and for good reason. Typically this might be |
a0d0e21e |
811 | done by a package attempting to emulate missing builtin functionality |
812 | on a non-Unix system. |
813 | |
5f05dabc |
814 | Overriding may be done only by importing the name from a |
a0d0e21e |
815 | module--ordinary predeclaration isn't good enough. However, the |
5f05dabc |
816 | C<subs> pragma (compiler directive) lets you, in effect, pre-declare subs |
a0d0e21e |
817 | via the import syntax, and these names may then override the builtin ones: |
818 | |
819 | use subs 'chdir', 'chroot', 'chmod', 'chown'; |
820 | chdir $somewhere; |
821 | sub chdir { ... } |
822 | |
823 | Library modules should not in general export builtin names like "open" |
5f05dabc |
824 | or "chdir" as part of their default @EXPORT list, because these may |
a0d0e21e |
825 | sneak into someone else's namespace and change the semantics unexpectedly. |
826 | Instead, if the module adds the name to the @EXPORT_OK list, then it's |
827 | possible for a user to import the name explicitly, but not implicitly. |
828 | That is, they could say |
829 | |
830 | use Module 'open'; |
831 | |
832 | and it would import the open override, but if they said |
833 | |
834 | use Module; |
835 | |
836 | they would get the default imports without the overrides. |
837 | |
838 | =head2 Autoloading |
839 | |
840 | If you call a subroutine that is undefined, you would ordinarily get an |
841 | immediate fatal error complaining that the subroutine doesn't exist. |
842 | (Likewise for subroutines being used as methods, when the method |
843 | doesn't exist in any of the base classes of the class package.) If, |
844 | however, there is an C<AUTOLOAD> subroutine defined in the package or |
845 | packages that were searched for the original subroutine, then that |
846 | C<AUTOLOAD> subroutine is called with the arguments that would have been |
847 | passed to the original subroutine. The fully qualified name of the |
848 | original subroutine magically appears in the $AUTOLOAD variable in the |
849 | same package as the C<AUTOLOAD> routine. The name is not passed as an |
850 | ordinary argument because, er, well, just because, that's why... |
851 | |
852 | Most C<AUTOLOAD> routines will load in a definition for the subroutine in |
853 | question using eval, and then execute that subroutine using a special |
854 | form of "goto" that erases the stack frame of the C<AUTOLOAD> routine |
855 | without a trace. (See the standard C<AutoLoader> module, for example.) |
856 | But an C<AUTOLOAD> routine can also just emulate the routine and never |
cb1a09d0 |
857 | define it. For example, let's pretend that a function that wasn't defined |
858 | should just call system() with those arguments. All you'd do is this: |
859 | |
860 | sub AUTOLOAD { |
861 | my $program = $AUTOLOAD; |
862 | $program =~ s/.*:://; |
863 | system($program, @_); |
864 | } |
865 | date(); |
6d28dffb |
866 | who('am', 'i'); |
cb1a09d0 |
867 | ls('-l'); |
868 | |
5f05dabc |
869 | In fact, if you pre-declare the functions you want to call that way, you don't |
cb1a09d0 |
870 | even need the parentheses: |
871 | |
872 | use subs qw(date who ls); |
873 | date; |
874 | who "am", "i"; |
875 | ls -l; |
876 | |
877 | A more complete example of this is the standard Shell module, which |
a0d0e21e |
878 | can treat undefined subroutine calls as calls to Unix programs. |
879 | |
cb1a09d0 |
880 | Mechanisms are available for modules writers to help split the modules |
6d28dffb |
881 | up into autoloadable files. See the standard AutoLoader module |
882 | described in L<AutoLoader> and in L<AutoSplit>, the standard |
883 | SelfLoader modules in L<SelfLoader>, and the document on adding C |
884 | functions to perl code in L<perlxs>. |
cb1a09d0 |
885 | |
886 | =head1 SEE ALSO |
a0d0e21e |
887 | |
cb1a09d0 |
888 | See L<perlref> for more on references. See L<perlxs> if you'd |
889 | like to learn about calling C subroutines from perl. See |
890 | L<perlmod> to learn about bundling up your functions in |
891 | separate files. |