3 perlmod - Perl modules (packages and symbol tables)
8 X<package> X<namespace> X<variable, global> X<global variable> X<global>
10 Perl provides a mechanism for alternative namespaces to protect
11 packages from stomping on each other's variables. In fact, there's
12 really no such thing as a global variable in Perl. The package
13 statement declares the compilation unit as being in the given
14 namespace. The scope of the package declaration is from the
15 declaration itself through the end of the enclosing block, C<eval>,
16 or file, whichever comes first (the same scope as the my() and
17 local() operators). Unqualified dynamic identifiers will be in
18 this namespace, except for those few identifiers that if unqualified,
19 default to the main package instead of the current one as described
20 below. A package statement affects only dynamic variables--including
21 those you've used local() on--but I<not> lexical variables created
22 with my(). Typically it would be the first declaration in a file
23 included by the C<do>, C<require>, or C<use> operators. You can
24 switch into a package in more than one place; it merely influences
25 which symbol table is used by the compiler for the rest of that
26 block. You can refer to variables and filehandles in other packages
27 by prefixing the identifier with the package name and a double
28 colon: C<$Package::Variable>. If the package name is null, the
29 C<main> package is assumed. That is, C<$::sail> is equivalent to
32 The old package delimiter was a single quote, but double colon is now the
33 preferred delimiter, in part because it's more readable to humans, and
34 in part because it's more readable to B<emacs> macros. It also makes C++
35 programmers feel like they know what's going on--as opposed to using the
36 single quote as separator, which was there to make Ada programmers feel
37 like they knew what was going on. Because the old-fashioned syntax is still
38 supported for backwards compatibility, if you try to use a string like
39 C<"This is $owner's house">, you'll be accessing C<$owner::s>; that is,
40 the $s variable in package C<owner>, which is probably not what you meant.
41 Use braces to disambiguate, as in C<"This is ${owner}'s house">.
44 Packages may themselves contain package separators, as in
45 C<$OUTER::INNER::var>. This implies nothing about the order of
46 name lookups, however. There are no relative packages: all symbols
47 are either local to the current package, or must be fully qualified
48 from the outer package name down. For instance, there is nowhere
49 within package C<OUTER> that C<$INNER::var> refers to
50 C<$OUTER::INNER::var>. C<INNER> refers to a totally
51 separate global package.
53 Only identifiers starting with letters (or underscore) are stored
54 in a package's symbol table. All other symbols are kept in package
55 C<main>, including all punctuation variables, like $_. In addition,
56 when unqualified, the identifiers STDIN, STDOUT, STDERR, ARGV,
57 ARGVOUT, ENV, INC, and SIG are forced to be in package C<main>,
58 even when used for other purposes than their built-in ones. If you
59 have a package called C<m>, C<s>, or C<y>, then you can't use the
60 qualified form of an identifier because it would be instead interpreted
61 as a pattern match, a substitution, or a transliteration.
62 X<variable, punctuation>
64 Variables beginning with underscore used to be forced into package
65 main, but we decided it was more useful for package writers to be able
66 to use leading underscore to indicate private variables and method names.
67 However, variables and functions named with a single C<_>, such as
68 $_ and C<sub _>, are still forced into the package C<main>. See also
69 L<perlvar/"The Syntax of Variable Names">.
71 C<eval>ed strings are compiled in the package in which the eval() was
72 compiled. (Assignments to C<$SIG{}>, however, assume the signal
73 handler specified is in the C<main> package. Qualify the signal handler
74 name if you wish to have a signal handler in a package.) For an
75 example, examine F<perldb.pl> in the Perl library. It initially switches
76 to the C<DB> package so that the debugger doesn't interfere with variables
77 in the program you are trying to debug. At various points, however, it
78 temporarily switches back to the C<main> package to evaluate various
79 expressions in the context of the C<main> package (or wherever you came
80 from). See L<perldebug>.
82 The special symbol C<__PACKAGE__> contains the current package, but cannot
83 (easily) be used to construct variable names.
85 See L<perlsub> for other scoping issues related to my() and local(),
86 and L<perlref> regarding closures.
89 X<symbol table> X<stash> X<%::> X<%main::> X<typeglob> X<glob> X<alias>
91 The symbol table for a package happens to be stored in the hash of that
92 name with two colons appended. The main symbol table's name is thus
93 C<%main::>, or C<%::> for short. Likewise the symbol table for the nested
94 package mentioned earlier is named C<%OUTER::INNER::>.
96 The value in each entry of the hash is what you are referring to when you
97 use the C<*name> typeglob notation.
99 local *main::foo = *main::bar;
101 You can use this to print out all the variables in a package, for
102 instance. The standard but antiquated F<dumpvar.pl> library and
103 the CPAN module Devel::Symdump make use of this.
105 The results of creating new symbol table entries directly or modifying any
106 entries that are not already typeglobs are undefined and subject to change
107 between releases of perl.
109 Assignment to a typeglob performs an aliasing operation, i.e.,
113 causes variables, subroutines, formats, and file and directory handles
114 accessible via the identifier C<richard> also to be accessible via the
115 identifier C<dick>. If you want to alias only a particular variable or
116 subroutine, assign a reference instead:
120 Which makes $richard and $dick the same variable, but leaves
121 @richard and @dick as separate arrays. Tricky, eh?
123 There is one subtle difference between the following statements:
128 C<*foo = *bar> makes the typeglobs themselves synonymous while
129 C<*foo = \$bar> makes the SCALAR portions of two distinct typeglobs
130 refer to the same scalar value. This means that the following code:
133 *foo = \$bar; # Make $foo an alias for $bar
136 local $bar = 2; # Restrict changes to block
137 print $foo; # Prints '1'!
140 Would print '1', because C<$foo> holds a reference to the I<original>
141 C<$bar>. The one that was stuffed away by C<local()> and which will be
142 restored when the block ends. Because variables are accessed through the
143 typeglob, you can use C<*foo = *bar> to create an alias which can be
144 localized. (But be aware that this means you can't have a separate
145 C<@foo> and C<@bar>, etc.)
147 What makes all of this important is that the Exporter module uses glob
148 aliasing as the import/export mechanism. Whether or not you can properly
149 localize a variable that has been exported from a module depends on how
152 @EXPORT = qw($FOO); # Usual form, can't be localized
153 @EXPORT = qw(*FOO); # Can be localized
155 You can work around the first case by using the fully qualified name
156 (C<$Package::FOO>) where you need a local value, or by overriding it
157 by saying C<*FOO = *Package::FOO> in your script.
159 The C<*x = \$y> mechanism may be used to pass and return cheap references
160 into or from subroutines if you don't want to copy the whole
161 thing. It only works when assigning to dynamic variables, not
164 %some_hash = (); # can't be my()
165 *some_hash = fn( \%another_hash );
167 local *hashsym = shift;
168 # now use %hashsym normally, and you
169 # will affect the caller's %another_hash
170 my %nhash = (); # do what you want
174 On return, the reference will overwrite the hash slot in the
175 symbol table specified by the *some_hash typeglob. This
176 is a somewhat tricky way of passing around references cheaply
177 when you don't want to have to remember to dereference variables
180 Another use of symbol tables is for making "constant" scalars.
181 X<constant> X<scalar, constant>
183 *PI = \3.14159265358979;
185 Now you cannot alter C<$PI>, which is probably a good thing all in all.
186 This isn't the same as a constant subroutine, which is subject to
187 optimization at compile-time. A constant subroutine is one prototyped
188 to take no arguments and to return a constant expression. See
189 L<perlsub> for details on these. The C<use constant> pragma is a
190 convenient shorthand for these.
192 You can say C<*foo{PACKAGE}> and C<*foo{NAME}> to find out what name and
193 package the *foo symbol table entry comes from. This may be useful
194 in a subroutine that gets passed typeglobs as arguments:
196 sub identify_typeglob {
198 print 'You gave me ', *{$glob}{PACKAGE}, '::', *{$glob}{NAME}, "\n";
200 identify_typeglob *foo;
201 identify_typeglob *bar::baz;
205 You gave me main::foo
208 The C<*foo{THING}> notation can also be used to obtain references to the
209 individual elements of *foo. See L<perlref>.
211 Subroutine definitions (and declarations, for that matter) need
212 not necessarily be situated in the package whose symbol table they
213 occupy. You can define a subroutine outside its package by
214 explicitly qualifying the name of the subroutine:
217 sub Some_package::foo { ... } # &foo defined in Some_package
219 This is just a shorthand for a typeglob assignment at compile time:
221 BEGIN { *Some_package::foo = sub { ... } }
223 and is I<not> the same as writing:
226 package Some_package;
230 In the first two versions, the body of the subroutine is
231 lexically in the main package, I<not> in Some_package. So
236 $Some_package::name = "fred";
237 $main::name = "barney";
239 sub Some_package::foo {
240 print "in ", __PACKAGE__, ": \$name is '$name'\n";
247 in main: $name is 'barney'
251 in Some_package: $name is 'fred'
253 This also has implications for the use of the SUPER:: qualifier
256 =head2 BEGIN, UNITCHECK, CHECK, INIT and END
257 X<BEGIN> X<UNITCHECK> X<CHECK> X<INIT> X<END>
259 Five specially named code blocks are executed at the beginning and at
260 the end of a running Perl program. These are the C<BEGIN>,
261 C<UNITCHECK>, C<CHECK>, C<INIT>, and C<END> blocks.
263 These code blocks can be prefixed with C<sub> to give the appearance of a
264 subroutine (although this is not considered good style). One should note
265 that these code blocks don't really exist as named subroutines (despite
266 their appearance). The thing that gives this away is the fact that you can
267 have B<more than one> of these code blocks in a program, and they will get
268 B<all> executed at the appropriate moment. So you can't execute any of
269 these code blocks by name.
271 A C<BEGIN> code block is executed as soon as possible, that is, the moment
272 it is completely defined, even before the rest of the containing file (or
273 string) is parsed. You may have multiple C<BEGIN> blocks within a file (or
274 eval'ed string); they will execute in order of definition. Because a C<BEGIN>
275 code block executes immediately, it can pull in definitions of subroutines
276 and such from other files in time to be visible to the rest of the compile
277 and run time. Once a C<BEGIN> has run, it is immediately undefined and any
278 code it used is returned to Perl's memory pool.
280 An C<END> code block is executed as late as possible, that is, after
281 perl has finished running the program and just before the interpreter
282 is being exited, even if it is exiting as a result of a die() function.
283 (But not if it's morphing into another program via C<exec>, or
284 being blown out of the water by a signal--you have to trap that yourself
285 (if you can).) You may have multiple C<END> blocks within a file--they
286 will execute in reverse order of definition; that is: last in, first
287 out (LIFO). C<END> blocks are not executed when you run perl with the
288 C<-c> switch, or if compilation fails.
290 Note that C<END> code blocks are B<not> executed at the end of a string
291 C<eval()>: if any C<END> code blocks are created in a string C<eval()>,
292 they will be executed just as any other C<END> code block of that package
293 in LIFO order just before the interpreter is being exited.
295 Inside an C<END> code block, C<$?> contains the value that the program is
296 going to pass to C<exit()>. You can modify C<$?> to change the exit
297 value of the program. Beware of changing C<$?> by accident (e.g. by
298 running something via C<system>).
301 Inside of a C<END> block, the value of C<${^GLOBAL_PHASE}> will be
304 C<UNITCHECK>, C<CHECK> and C<INIT> code blocks are useful to catch the
305 transition between the compilation phase and the execution phase of
308 C<UNITCHECK> blocks are run just after the unit which defined them has
309 been compiled. The main program file and each module it loads are
310 compilation units, as are string C<eval>s, code compiled using the
311 C<(?{ })> construct in a regex, calls to C<do FILE>, C<require FILE>,
312 and code after the C<-e> switch on the command line.
314 C<BEGIN> and C<UNITCHECK> blocks are not directly related to the phase of
315 the interpreter. They can be created and executed during any phase.
317 C<CHECK> code blocks are run just after the B<initial> Perl compile phase ends
318 and before the run time begins, in LIFO order. C<CHECK> code blocks are used
319 in the Perl compiler suite to save the compiled state of the program.
321 Inside of a C<CHECK> block, the value of C<${^GLOBAL_PHASE}> will be
324 C<INIT> blocks are run just before the Perl runtime begins execution, in
325 "first in, first out" (FIFO) order.
327 Inside of an C<INIT> block, the value of C<${^GLOBAL_PHASE}> will be C<"INIT">.
329 The C<CHECK> and C<INIT> blocks in code compiled by C<require>, string C<do>,
330 or string C<eval> will not be executed if they occur after the end of the
331 main compilation phase; that can be a problem in mod_perl and other persistent
332 environments which use those functions to load code at runtime.
334 When you use the B<-n> and B<-p> switches to Perl, C<BEGIN> and
335 C<END> work just as they do in B<awk>, as a degenerate case.
336 Both C<BEGIN> and C<CHECK> blocks are run when you use the B<-c>
337 switch for a compile-only syntax check, although your main code
340 The B<begincheck> program makes it all clear, eventually:
346 print "10. Ordinary code runs at runtime.\n";
348 END { print "16. So this is the end of the tale.\n" }
349 INIT { print " 7. INIT blocks run FIFO just before runtime.\n" }
351 print " 4. And therefore before any CHECK blocks.\n"
353 CHECK { print " 6. So this is the sixth line.\n" }
355 print "11. It runs in order, of course.\n";
357 BEGIN { print " 1. BEGIN blocks run FIFO during compilation.\n" }
358 END { print "15. Read perlmod for the rest of the story.\n" }
359 CHECK { print " 5. CHECK blocks run LIFO after all compilation.\n" }
360 INIT { print " 8. Run this again, using Perl's -c switch.\n" }
362 print "12. This is anti-obfuscated code.\n";
364 END { print "14. END blocks run LIFO at quitting time.\n" }
365 BEGIN { print " 2. So this line comes out second.\n" }
367 print " 3. UNITCHECK blocks run LIFO after each file is compiled.\n"
369 INIT { print " 9. You'll see the difference right away.\n" }
371 print "13. It merely _looks_ like it should be confusing.\n";
378 There is no special class syntax in Perl, but a package may act
379 as a class if it provides subroutines to act as methods. Such a
380 package may also derive some of its methods from another class (package)
381 by listing the other package name(s) in its global @ISA array (which
382 must be a package global, not a lexical).
384 For more on this, see L<perlootut> and L<perlobj>.
389 A module is just a set of related functions in a library file, i.e.,
390 a Perl package with the same name as the file. It is specifically
391 designed to be reusable by other modules or programs. It may do this
392 by providing a mechanism for exporting some of its symbols into the
393 symbol table of any package using it, or it may function as a class
394 definition and make its semantics available implicitly through
395 method calls on the class and its objects, without explicitly
396 exporting anything. Or it can do a little of both.
398 For example, to start a traditional, non-OO module called Some::Module,
399 create a file called F<Some/Module.pm> and start with this template:
401 package Some::Module; # assumes Some/Module.pm
409 # set the version for version checking
412 # Inherit from Exporter to export functions and variables
413 our @ISA = qw(Exporter);
415 # Functions and variables which are exported by default
416 our @EXPORT = qw(func1 func2);
418 # Functions and variables which can be optionally exported
419 our @EXPORT_OK = qw($Var1 %Hashit func3);
422 # exported package globals go here
426 # non-exported package globals go here
427 # (they are still accessible as $Some::Module::stuff)
431 # file-private lexicals go here, before any functions which use them
433 my %secret_hash = ();
435 # here's a file-private function as a closure,
436 # callable as $priv_func->();
437 my $priv_func = sub {
441 # make all your functions, whether exported or not;
442 # remember to put something interesting in the {} stubs
446 # this one isn't exported, but could be called directly
447 # as Some::Module::func3()
450 END { ... } # module clean-up code here (global destructor)
452 1; # don't forget to return a true value from the file
454 Then go on to declare and use your variables in functions without
455 any qualifications. See L<Exporter> and the L<perlmodlib> for
456 details on mechanics and style issues in module creation.
458 Perl modules are included into your program by saying
466 This is exactly equivalent to
468 BEGIN { require 'Module.pm'; 'Module'->import; }
472 BEGIN { require 'Module.pm'; 'Module'->import( LIST ); }
478 is exactly equivalent to
480 BEGIN { require 'Module.pm'; }
482 All Perl module files have the extension F<.pm>. The C<use> operator
483 assumes this so you don't have to spell out "F<Module.pm>" in quotes.
484 This also helps to differentiate new modules from old F<.pl> and
485 F<.ph> files. Module names are also capitalized unless they're
486 functioning as pragmas; pragmas are in effect compiler directives,
487 and are sometimes called "pragmatic modules" (or even "pragmata"
488 if you're a classicist).
493 require "SomeModule.pm";
495 differ from each other in two ways. In the first case, any double
496 colons in the module name, such as C<Some::Module>, are translated
497 into your system's directory separator, usually "/". The second
498 case does not, and would have to be specified literally. The other
499 difference is that seeing the first C<require> clues in the compiler
500 that uses of indirect object notation involving "SomeModule", as
501 in C<$ob = purge SomeModule>, are method calls, not function calls.
502 (Yes, this really can make a difference.)
504 Because the C<use> statement implies a C<BEGIN> block, the importing
505 of semantics happens as soon as the C<use> statement is compiled,
506 before the rest of the file is compiled. This is how it is able
507 to function as a pragma mechanism, and also how modules are able to
508 declare subroutines that are then visible as list or unary operators for
509 the rest of the current file. This will not work if you use C<require>
510 instead of C<use>. With C<require> you can get into this problem:
512 require Cwd; # make Cwd:: accessible
513 $here = Cwd::getcwd();
515 use Cwd; # import names from Cwd::
518 require Cwd; # make Cwd:: accessible
519 $here = getcwd(); # oops! no main::getcwd()
521 In general, C<use Module ()> is recommended over C<require Module>,
522 because it determines module availability at compile time, not in the
523 middle of your program's execution. An exception would be if two modules
524 each tried to C<use> each other, and each also called a function from
525 that other module. In that case, it's easy to use C<require> instead.
527 Perl packages may be nested inside other package names, so we can have
528 package names containing C<::>. But if we used that package name
529 directly as a filename it would make for unwieldy or impossible
530 filenames on some systems. Therefore, if a module's name is, say,
531 C<Text::Soundex>, then its definition is actually found in the library
532 file F<Text/Soundex.pm>.
534 Perl modules always have a F<.pm> file, but there may also be
535 dynamically linked executables (often ending in F<.so>) or autoloaded
536 subroutine definitions (often ending in F<.al>) associated with the
537 module. If so, these will be entirely transparent to the user of
538 the module. It is the responsibility of the F<.pm> file to load
539 (or arrange to autoload) any additional functionality. For example,
540 although the POSIX module happens to do both dynamic loading and
541 autoloading, the user can say just C<use POSIX> to get it all.
543 =head2 Making your module threadsafe
544 X<threadsafe> X<thread safe>
545 X<module, threadsafe> X<module, thread safe>
546 X<CLONE> X<CLONE_SKIP> X<thread> X<threads> X<ithread>
548 Since 5.6.0, Perl has had support for a new type of threads called
549 interpreter threads (ithreads). These threads can be used explicitly
552 Ithreads work by cloning the data tree so that no data is shared
553 between different threads. These threads can be used by using the C<threads>
554 module or by doing fork() on win32 (fake fork() support). When a
555 thread is cloned all Perl data is cloned, however non-Perl data cannot
556 be cloned automatically. Perl after 5.7.2 has support for the C<CLONE>
557 special subroutine. In C<CLONE> you can do whatever
559 like for example handle the cloning of non-Perl data, if necessary.
560 C<CLONE> will be called once as a class method for every package that has it
561 defined (or inherits it). It will be called in the context of the new thread,
562 so all modifications are made in the new area. Currently CLONE is called with
563 no parameters other than the invocant package name, but code should not assume
564 that this will remain unchanged, as it is likely that in future extra parameters
565 will be passed in to give more information about the state of cloning.
567 If you want to CLONE all objects you will need to keep track of them per
568 package. This is simply done using a hash and Scalar::Util::weaken().
570 Perl after 5.8.7 has support for the C<CLONE_SKIP> special subroutine.
571 Like C<CLONE>, C<CLONE_SKIP> is called once per package; however, it is
572 called just before cloning starts, and in the context of the parent
573 thread. If it returns a true value, then no objects of that class will
574 be cloned; or rather, they will be copied as unblessed, undef values.
575 For example: if in the parent there are two references to a single blessed
576 hash, then in the child there will be two references to a single undefined
577 scalar value instead.
578 This provides a simple mechanism for making a module threadsafe; just add
579 C<sub CLONE_SKIP { 1 }> at the top of the class, and C<DESTROY()> will
580 now only be called once per object. Of course, if the child thread needs
581 to make use of the objects, then a more sophisticated approach is
584 Like C<CLONE>, C<CLONE_SKIP> is currently called with no parameters other
585 than the invocant package name, although that may change. Similarly, to
586 allow for future expansion, the return value should be a single C<0> or
591 See L<perlmodlib> for general style issues related to building Perl
592 modules and classes, as well as descriptions of the standard library
593 and CPAN, L<Exporter> for how Perl's standard import/export mechanism
594 works, L<perlootut> and L<perlobj> for in-depth information on
595 creating classes, L<perlobj> for a hard-core reference document on
596 objects, L<perlsub> for an explanation of functions and scoping,
597 and L<perlxstut> and L<perlguts> for more information on writing