3 perlfaq7 - Perl Language Issues ($Revision: 1.24 $, $Date: 1999/01/08 05:32:11 $)
7 This section deals with general Perl language issues that don't
8 clearly fit into any of the other sections.
10 =head2 Can I get a BNF/yacc/RE for the Perl language?
12 There is no BNF, but you can paw your way through the yacc grammar in
13 perly.y in the source distribution if you're particularly brave. The
14 grammar relies on very smart tokenizing code, so be prepared to
15 venture into toke.c as well.
17 In the words of Chaim Frenkel: "Perl's grammar can not be reduced to BNF.
18 The work of parsing perl is distributed between yacc, the lexer, smoke
21 =head2 What are all these $@%* punctuation signs, and how do I know when to use them?
23 They are type specifiers, as detailed in L<perldata>:
25 $ for scalar values (number, string or reference)
27 % for hashes (associative arrays)
28 * for all types of that symbol name. In version 4 you used them like
29 pointers, but in modern perls you can just use references.
31 While there are a few places where you don't actually need these type
32 specifiers, you should always use them.
34 A couple of others that you're likely to encounter that aren't
35 really type specifiers are:
37 <> are used for inputting a record from a filehandle.
38 \ takes a reference to something.
40 Note that E<lt>FILEE<gt> is I<neither> the type specifier for files
41 nor the name of the handle. It is the C<E<lt>E<gt>> operator applied
42 to the handle FILE. It reads one line (well, record - see
43 L<perlvar/$/>) from the handle FILE in scalar context, or I<all> lines
44 in list context. When performing open, close, or any other operation
45 besides C<E<lt>E<gt>> on files, or even talking about the handle, do
46 I<not> use the brackets. These are correct: C<eof(FH)>, C<seek(FH, 0,
47 2)> and "copying from STDIN to FILE".
49 =head2 Do I always/never have to quote my strings or use semicolons and commas?
51 Normally, a bareword doesn't need to be quoted, but in most cases
52 probably should be (and must be under C<use strict>). But a hash key
53 consisting of a simple word (that isn't the name of a defined
54 subroutine) and the left-hand operand to the C<=E<gt>> operator both
55 count as though they were quoted:
58 ------------ ---------------
59 $foo{line} $foo{"line"}
60 bar => stuff "bar" => stuff
62 The final semicolon in a block is optional, as is the final comma in a
63 list. Good style (see L<perlstyle>) says to put them in except for
66 if ($whoops) { exit 1 }
73 "There Beren came from mountains cold",
74 "And lost he wandered under leaves",
77 =head2 How do I skip some return values?
79 One way is to treat the return values as a list and index into it:
81 $dir = (getpwnam($user))[7];
83 Another way is to use undef as an element on the left-hand-side:
85 ($dev, $ino, undef, undef, $uid, $gid) = stat($file);
87 =head2 How do I temporarily block warnings?
89 The C<$^W> variable (documented in L<perlvar>) controls
90 runtime warnings for a block:
93 local $^W = 0; # temporarily turn off warnings
94 $a = $b + $c; # I know these might be undef
97 Note that like all the punctuation variables, you cannot currently
98 use my() on C<$^W>, only local().
100 A new C<use warnings> pragma is in the works to provide finer control
101 over all this. The curious should check the perl5-porters mailing list
102 archives for details.
104 =head2 What's an extension?
106 A way of calling compiled C code from Perl. Reading L<perlxstut>
107 is a good place to learn more about extensions.
109 =head2 Why do Perl operators have different precedence than C operators?
111 Actually, they don't. All C operators that Perl copies have the same
112 precedence in Perl as they do in C. The problem is with operators that C
113 doesn't have, especially functions that give a list context to everything
114 on their right, eg print, chmod, exec, and so on. Such functions are
115 called "list operators" and appear as such in the precedence table in
118 A common mistake is to write:
120 unlink $file || die "snafu";
122 This gets interpreted as:
124 unlink ($file || die "snafu");
126 To avoid this problem, either put in extra parentheses or use the
127 super low precedence C<or> operator:
129 (unlink $file) || die "snafu";
130 unlink $file or die "snafu";
132 The "English" operators (C<and>, C<or>, C<xor>, and C<not>)
133 deliberately have precedence lower than that of list operators for
134 just such situations as the one above.
136 Another operator with surprising precedence is exponentiation. It
137 binds more tightly even than unary minus, making C<-2**2> product a
138 negative not a positive four. It is also right-associating, meaning
139 that C<2**3**2> is two raised to the ninth power, not eight squared.
141 Although it has the same precedence as in C, Perl's C<?:> operator
142 produces an lvalue. This assigns $x to either $a or $b, depending
143 on the trueness of $maybe:
145 ($maybe ? $a : $b) = $x;
147 =head2 How do I declare/create a structure?
149 In general, you don't "declare" a structure. Just use a (probably
150 anonymous) hash reference. See L<perlref> and L<perldsc> for details.
153 $person = {}; # new anonymous hash
154 $person->{AGE} = 24; # set field AGE to 24
155 $person->{NAME} = "Nat"; # set field NAME to "Nat"
157 If you're looking for something a bit more rigorous, try L<perltoot>.
159 =head2 How do I create a module?
161 A module is a package that lives in a file of the same name. For
162 example, the Hello::There module would live in Hello/There.pm. For
163 details, read L<perlmod>. You'll also find L<Exporter> helpful. If
164 you're writing a C or mixed-language module with both C and Perl, then
165 you should study L<perlxstut>.
167 Here's a convenient template you might wish you use when starting your
168 own module. Make sure to change the names appropriately.
170 package Some::Module; # assumes Some/Module.pm
176 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS);
178 ## set the version for version checking; uncomment to use
181 # if using RCS/CVS, this next line may be preferred,
182 # but beware two-digit versions.
183 $VERSION = do{my@r=q$Revision: 1.24 $=~/\d+/g;sprintf '%d.'.'%02d'x$#r,@r};
186 @EXPORT = qw(&func1 &func2 &func3);
187 %EXPORT_TAGS = ( ); # eg: TAG => [ qw!name1 name2! ],
189 # your exported package globals go here,
190 # as well as any optionally exported functions
191 @EXPORT_OK = qw($Var1 %Hashit);
195 # non-exported package globals go here
196 use vars qw( @more $stuff );
198 # initialize package globals, first exported ones
202 # then the others (which are still accessible as $Some::Module::stuff)
206 # all file-scoped lexicals must be created before
207 # the functions below that use them.
209 # file-private lexicals go here
211 my %secret_hash = ();
213 # here's a file-private function as a closure,
214 # callable as &$priv_func; it cannot be prototyped.
215 my $priv_func = sub {
219 # make all your functions, whether exported or not;
220 # remember to put something interesting in the {} stubs
221 sub func1 {} # no prototype
222 sub func2() {} # proto'd void
223 sub func3($$) {} # proto'd to 2 scalars
225 # this one isn't exported, but could be called!
226 sub func4(\%) {} # proto'd to 1 hash ref
228 END { } # module clean-up code here (global destructor)
230 1; # modules must return true
232 The h2xs program will create stubs for all the important stuff for you:
234 % h2xs -XA -n My::Module
236 =head2 How do I create a class?
238 See L<perltoot> for an introduction to classes and objects, as well as
239 L<perlobj> and L<perlbot>.
241 =head2 How can I tell if a variable is tainted?
243 See L<perlsec/"Laundering and Detecting Tainted Data">. Here's an
244 example (which doesn't use any system calls, because the kill()
245 is given no processes to signal):
248 return ! eval { join('',@_), kill 0; 1; };
251 This is not C<-w> clean, however. There is no C<-w> clean way to
252 detect taintedness - take this as a hint that you should untaint
253 all possibly-tainted data.
255 =head2 What's a closure?
257 Closures are documented in L<perlref>.
259 I<Closure> is a computer science term with a precise but
260 hard-to-explain meaning. Closures are implemented in Perl as anonymous
261 subroutines with lasting references to lexical variables outside their
262 own scopes. These lexicals magically refer to the variables that were
263 around when the subroutine was defined (deep binding).
265 Closures make sense in any programming language where you can have the
266 return value of a function be itself a function, as you can in Perl.
267 Note that some languages provide anonymous functions but are not
268 capable of providing proper closures; the Python language, for
269 example. For more information on closures, check out any textbook on
270 functional programming. Scheme is a language that not only supports
271 but encourages closures.
273 Here's a classic function-generating function:
275 sub add_function_generator {
276 return sub { shift + shift };
279 $add_sub = add_function_generator();
280 $sum = $add_sub->(4,5); # $sum is 9 now.
282 The closure works as a I<function template> with some customization
283 slots left out to be filled later. The anonymous subroutine returned
284 by add_function_generator() isn't technically a closure because it
285 refers to no lexicals outside its own scope.
287 Contrast this with the following make_adder() function, in which the
288 returned anonymous function contains a reference to a lexical variable
289 outside the scope of that function itself. Such a reference requires
290 that Perl return a proper closure, thus locking in for all time the
291 value that the lexical had when the function was created.
294 my $addpiece = shift;
295 return sub { shift + $addpiece };
298 $f1 = make_adder(20);
299 $f2 = make_adder(555);
301 Now C<&$f1($n)> is always 20 plus whatever $n you pass in, whereas
302 C<&$f2($n)> is always 555 plus whatever $n you pass in. The $addpiece
303 in the closure sticks around.
305 Closures are often used for less esoteric purposes. For example, when
306 you want to pass in a bit of code into a function:
309 timeout( 30, sub { $line = <STDIN> } );
311 If the code to execute had been passed in as a string, C<'$line =
312 E<lt>STDINE<gt>'>, there would have been no way for the hypothetical
313 timeout() function to access the lexical variable $line back in its
316 =head2 What is variable suicide and how can I prevent it?
318 Variable suicide is when you (temporarily or permanently) lose the
319 value of a variable. It is caused by scoping through my() and local()
320 interacting with either closures or aliased foreach() iterator
321 variables and subroutine arguments. It used to be easy to
322 inadvertently lose a variable's value this way, but now it's much
323 harder. Take this code:
327 while ($i++ < 3) { my $f = $f; $f .= "bar"; print $f, "\n" }
330 print "Finally $f\n";
332 The $f that has "bar" added to it three times should be a new C<$f>
333 (C<my $f> should create a new local variable each time through the
334 loop). It isn't, however. This is a bug, and will be fixed.
336 =head2 How can I pass/return a {Function, FileHandle, Array, Hash, Method, Regexp}?
338 With the exception of regexps, you need to pass references to these
339 objects. See L<perlsub/"Pass by Reference"> for this particular
340 question, and L<perlref> for information on references.
344 =item Passing Variables and Functions
346 Regular variables and functions are quite easy: just pass in a
347 reference to an existing or anonymous variable or function:
349 func( \$some_scalar );
351 func( \@some_array );
355 func( { this => 10, that => 20 } );
358 func( sub { $_[0] ** $_[1] } );
360 =item Passing Filehandles
362 To pass filehandles to subroutines, use the C<*FH> or C<\*FH> notations.
363 These are "typeglobs" - see L<perldata/"Typeglobs and Filehandles">
364 and especially L<perlsub/"Pass by Reference"> for more information.
368 If you're passing around filehandles, you could usually just use the bare
369 typeglob, like *STDOUT, but typeglobs references would be better because
370 they'll still work properly under C<use strict 'refs'>. For example:
375 print $fh "her um well a hmmm\n";
378 $rec = get_rec(\*STDIN);
384 If you're planning on generating new filehandles, you could do this:
389 return open (FH, $path) ? *FH : undef;
391 $fh = openit('< /etc/motd');
394 =item Passing Regexps
396 To pass regexps around, you'll need to either use one of the highly
397 experimental regular expression modules from CPAN (Nick Ing-Simmons's
398 Regexp or Ilya Zakharevich's Devel::Regexp), pass around strings
399 and use an exception-trapping eval, or else be very, very clever.
400 Here's an example of how to pass in a string to be regexp compared:
403 my ($val1, $regexp) = @_;
404 my $retval = eval { $val =~ /$regexp/ };
409 $match = compare("old McDonald", q/d.*D/);
411 Make sure you never say something like this:
413 return eval "\$val =~ /$regexp/"; # WRONG
415 or someone can sneak shell escapes into the regexp due to the double
416 interpolation of the eval and the double-quoted string. For example:
418 $pattern_of_evil = 'danger ${ system("rm -rf * &") } danger';
420 eval "\$string =~ /$pattern_of_evil/";
422 Those preferring to be very, very clever might see the O'Reilly book,
423 I<Mastering Regular Expressions>, by Jeffrey Friedl. Page 273's
424 Build_MatchMany_Function() is particularly interesting. A complete
425 citation of this book is given in L<perlfaq2>.
427 =item Passing Methods
429 To pass an object method into a subroutine, you can do this:
431 call_a_lot(10, $some_obj, "methname")
433 my ($count, $widget, $trick) = @_;
434 for (my $i = 0; $i < $count; $i++) {
439 Or you can use a closure to bundle up the object and its method call
442 my $whatnot = sub { $some_obj->obfuscate(@args) };
449 You could also investigate the can() method in the UNIVERSAL class
450 (part of the standard perl distribution).
454 =head2 How do I create a static variable?
456 As with most things in Perl, TMTOWTDI. What is a "static variable" in
457 other languages could be either a function-private variable (visible
458 only within a single function, retaining its value between calls to
459 that function), or a file-private variable (visible only to functions
460 within the file it was declared in) in Perl.
462 Here's code to implement a function-private variable:
466 sub prev_counter { return --$counter }
467 sub next_counter { return $counter++ }
470 Now prev_counter() and next_counter() share a private variable $counter
471 that was initialized at compile time.
473 To declare a file-private variable, you'll still use a my(), putting
474 it at the outer scope level at the top of the file. Assume this is in
478 my $started = scalar(localtime(time()));
480 sub begun { return $started }
482 When C<use Pax> or C<require Pax> loads this module, the variable will
483 be initialized. It won't get garbage-collected the way most variables
484 going out of scope do, because the begun() function cares about it,
485 but no one else can get it. It is not called $Pax::started because
486 its scope is unrelated to the package. It's scoped to the file. You
487 could conceivably have several packages in that same file all
488 accessing the same private variable, but another file with the same
489 package couldn't get to it.
491 See L<perlsub/"Peristent Private Variables"> for details.
493 =head2 What's the difference between dynamic and lexical (static) scoping? Between local() and my()?
495 C<local($x)> saves away the old value of the global variable C<$x>,
496 and assigns a new value for the duration of the subroutine, I<which is
497 visible in other functions called from that subroutine>. This is done
498 at run-time, so is called dynamic scoping. local() always affects global
499 variables, also called package variables or dynamic variables.
501 C<my($x)> creates a new variable that is only visible in the current
502 subroutine. This is done at compile-time, so is called lexical or
503 static scoping. my() always affects private variables, also called
504 lexical variables or (improperly) static(ly scoped) variables.
509 print "var has value $var\n";
513 local $var = 'local'; # new temporary value for the still-global
514 visible(); # variable called $var
518 my $var = 'private'; # new private variable, $var
519 visible(); # (invisible outside of sub scope)
524 visible(); # prints global
525 dynamic(); # prints local
526 lexical(); # prints global
528 Notice how at no point does the value "private" get printed. That's
529 because $var only has that value within the block of the lexical()
530 function, and it is hidden from called subroutine.
532 In summary, local() doesn't make what you think of as private, local
533 variables. It gives a global variable a temporary value. my() is
534 what you're looking for if you want private variables.
536 See L<perlsub/"Private Variables via my()"> and L<perlsub/"Temporary
537 Values via local()"> for excruciating details.
539 =head2 How can I access a dynamic variable while a similarly named lexical is in scope?
541 You can do this via symbolic references, provided you haven't set
542 C<use strict "refs">. So instead of $var, use C<${'var'}>.
544 local $var = "global";
547 print "lexical is $var\n";
550 print "global is ${'var'}\n";
552 If you know your package, you can just mention it explicitly, as in
553 $Some_Pack::var. Note that the notation $::var is I<not> the dynamic
554 $var in the current package, but rather the one in the C<main>
555 package, as though you had written $main::var. Specifying the package
556 directly makes you hard-code its name, but it executes faster and
557 avoids running afoul of C<use strict "refs">.
559 =head2 What's the difference between deep and shallow binding?
561 In deep binding, lexical variables mentioned in anonymous subroutines
562 are the same ones that were in scope when the subroutine was created.
563 In shallow binding, they are whichever variables with the same names
564 happen to be in scope when the subroutine is called. Perl always uses
565 deep binding of lexical variables (i.e., those created with my()).
566 However, dynamic variables (aka global, local, or package variables)
567 are effectively shallowly bound. Consider this just one more reason
568 not to use them. See the answer to L<"What's a closure?">.
570 =head2 Why doesn't "my($foo) = E<lt>FILEE<gt>;" work right?
572 C<my()> and C<local()> give list context to the right hand side
573 of C<=>. The E<lt>FHE<gt> read operation, like so many of Perl's
574 functions and operators, can tell which context it was called in and
575 behaves appropriately. In general, the scalar() function can help.
576 This function does nothing to the data itself (contrary to popular myth)
577 but rather tells its argument to behave in whatever its scalar fashion is.
578 If that function doesn't have a defined scalar behavior, this of course
579 doesn't help you (such as with sort()).
581 To enforce scalar context in this particular case, however, you need
582 merely omit the parentheses:
584 local($foo) = <FILE>; # WRONG
585 local($foo) = scalar(<FILE>); # ok
586 local $foo = <FILE>; # right
588 You should probably be using lexical variables anyway, although the
589 issue is the same here:
591 my($foo) = <FILE>; # WRONG
592 my $foo = <FILE>; # right
594 =head2 How do I redefine a builtin function, operator, or method?
596 Why do you want to do that? :-)
598 If you want to override a predefined function, such as open(),
599 then you'll have to import the new definition from a different
600 module. See L<perlsub/"Overriding Builtin Functions">. There's
601 also an example in L<perltoot/"Class::Template">.
603 If you want to overload a Perl operator, such as C<+> or C<**>,
604 then you'll want to use the C<use overload> pragma, documented
607 If you're talking about obscuring method calls in parent classes,
608 see L<perltoot/"Overridden Methods">.
610 =head2 What's the difference between calling a function as &foo and foo()?
612 When you call a function as C<&foo>, you allow that function access to
613 your current @_ values, and you by-pass prototypes. That means that
614 the function doesn't get an empty @_, it gets yours! While not
615 strictly speaking a bug (it's documented that way in L<perlsub>), it
616 would be hard to consider this a feature in most cases.
618 When you call your function as C<&foo()>, then you I<do> get a new @_,
619 but prototyping is still circumvented.
621 Normally, you want to call a function using C<foo()>. You may only
622 omit the parentheses if the function is already known to the compiler
623 because it already saw the definition (C<use> but not C<require>),
624 or via a forward reference or C<use subs> declaration. Even in this
625 case, you get a clean @_ without any of the old values leaking through
626 where they don't belong.
628 =head2 How do I create a switch or case statement?
630 This is explained in more depth in the L<perlsyn>. Briefly, there's
631 no official case statement, because of the variety of tests possible
632 in Perl (numeric comparison, string comparison, glob comparison,
633 regexp matching, overloaded comparisons, ...). Larry couldn't decide
634 how best to do this, so he left it out, even though it's been on the
635 wish list since perl1.
637 The general answer is to write a construct like this:
639 for ($variable_to_test) {
640 if (/pat1/) { } # do something
641 elsif (/pat2/) { } # do something else
642 elsif (/pat3/) { } # do something else
646 Here's a simple example of a switch based on pattern matching, this
647 time lined up in a way to make it look more like a switch statement.
648 We'll do a multi-way conditional based on the type of reference stored
651 SWITCH: for (ref $whatchamacallit) {
653 /^$/ && die "not a reference";
671 warn "can't print function ref";
677 warn "User defined type skipped";
681 See C<perlsyn/"Basic BLOCKs and Switch Statements"> for many other
682 examples in this style.
684 Sometimes you should change the positions of the constant and the variable.
685 For example, let's say you wanted to test which of many answers you were
686 given, but in a case-insensitive way that also allows abbreviations.
687 You can use the following technique if the strings all start with
688 different characters, or if you want to arrange the matches so that
689 one takes precedence over another, as C<"SEND"> has precedence over
693 if ("SEND" =~ /^\Q$answer/i) { print "Action is send\n" }
694 elsif ("STOP" =~ /^\Q$answer/i) { print "Action is stop\n" }
695 elsif ("ABORT" =~ /^\Q$answer/i) { print "Action is abort\n" }
696 elsif ("LIST" =~ /^\Q$answer/i) { print "Action is list\n" }
697 elsif ("EDIT" =~ /^\Q$answer/i) { print "Action is edit\n" }
699 A totally different approach is to create a hash of function references.
704 "done" => sub { die "See ya!" },
708 print "How are you? ";
709 chomp($string = <STDIN>);
710 if ($commands{$string}) {
711 $commands{$string}->();
713 print "No such command: $string\n";
716 =head2 How can I catch accesses to undefined variables/functions/methods?
718 The AUTOLOAD method, discussed in L<perlsub/"Autoloading"> and
719 L<perltoot/"AUTOLOAD: Proxy Methods">, lets you capture calls to
720 undefined functions and methods.
722 When it comes to undefined variables that would trigger a warning
723 under C<-w>, you can use a handler to trap the pseudo-signal
724 C<__WARN__> like this:
726 $SIG{__WARN__} = sub {
728 for ( $_[0] ) { # voici un switch statement
730 /Use of uninitialized value/ && do {
731 # promote warning to a fatal
735 # other warning cases to catch could go here;
742 =head2 Why can't a method included in this same file be found?
744 Some possible reasons: your inheritance is getting confused, you've
745 misspelled the method name, or the object is of the wrong type. Check
746 out L<perltoot> for details on these. You may also use C<print
747 ref($object)> to find out the class C<$object> was blessed into.
749 Another possible reason for problems is because you've used the
750 indirect object syntax (eg, C<find Guru "Samy">) on a class name
751 before Perl has seen that such a package exists. It's wisest to make
752 sure your packages are all defined before you start using them, which
753 will be taken care of if you use the C<use> statement instead of
754 C<require>. If not, make sure to use arrow notation (eg,
755 C<Guru-E<gt>find("Samy")>) instead. Object notation is explained in
758 Make sure to read about creating modules in L<perlmod> and
759 the perils of indirect objects in L<perlobj/"WARNING">.
761 =head2 How can I find out my current package?
763 If you're just a random program, you can do this to find
764 out what the currently compiled package is:
766 my $packname = __PACKAGE__;
768 But if you're a method and you want to print an error message
769 that includes the kind of object you were called on (which is
770 not necessarily the same as the one in which you were compiled):
774 my $class = ref($self) || $self;
775 warn "called me from a $class object";
778 =head2 How can I comment out a large block of perl code?
780 Use embedded POD to discard it:
785 This paragraph is commented out
800 This can't go just anywhere. You have to put a pod directive where
801 the parser is expecting a new statement, not just in the middle
802 of an expression or some other arbitrary yacc grammar production.
804 =head2 How do I clear a package?
806 Use this code, provided by Mark-Jason Dominus:
811 die "Shouldn't delete main package"
812 if $pack eq "" || $pack eq "main";
813 my $stash = *{$pack . '::'}{HASH};
815 foreach $name (keys %$stash) {
816 my $fullname = $pack . '::' . $name;
817 # Get rid of everything with that name.
826 Or, if you're using a recent release of Perl, you can
827 just use the Symbol::delete_package() function instead.
829 =head1 AUTHOR AND COPYRIGHT
831 Copyright (c) 1997-1999 Tom Christiansen and Nathan Torkington.
834 When included as part of the Standard Version of Perl, or as part of
835 its complete documentation whether printed or otherwise, this work
836 may be distributed only under the terms of Perl's Artistic License.
837 Any distribution of this file or derivatives thereof I<outside>
838 of that package require that special arrangements be made with
841 Irrespective of its distribution, all code examples in this file
842 are hereby placed into the public domain. You are permitted and
843 encouraged to use this code in your own programs for fun
844 or for profit as you see fit. A simple comment in the code giving
845 credit would be courteous but is not required.