source: trunk/third/perl/pod/perlrun.pod @ 17035

Revision 17035, 30.6 KB checked in by zacheiss, 23 years ago (diff)
This commit was generated by cvs2svn to compensate for changes in r17034, which included commits to RCS files with non-trunk default branches.
Line 
1=head1 NAME
2
3perlrun - how to execute the Perl interpreter
4
5=head1 SYNOPSIS
6
7B<perl> S<[ B<-CsTuUWX> ]>
8        S<[ B<-hv> ] [ B<-V>[:I<configvar>] ]>
9        S<[ B<-cw> ] [ B<-d>[:I<debugger>] ] [ B<-D>[I<number/list>] ]>
10        S<[ B<-pna> ] [ B<-F>I<pattern> ] [ B<-l>[I<octal>] ] [ B<-0>[I<octal>] ]>
11        S<[ B<-I>I<dir> ] [ B<-m>[B<->]I<module> ] [ B<-M>[B<->]I<'module...'> ]>
12        S<[ B<-P> ]>
13        S<[ B<-S> ]>
14        S<[ B<-x>[I<dir>] ]>
15        S<[ B<-i>[I<extension>] ]>
16        S<[ B<-e> I<'command'> ] [ B<--> ] [ I<programfile> ] [ I<argument> ]...>
17
18=head1 DESCRIPTION
19
20The normal way to run a Perl program is by making it directly
21executable, or else by passing the name of the source file as an
22argument on the command line.  (An interactive Perl environment
23is also possible--see L<perldebug> for details on how to do that.)
24Upon startup, Perl looks for your program in one of the following
25places:
26
27=over 4
28
29=item 1.
30
31Specified line by line via B<-e> switches on the command line.
32
33=item 2.
34
35Contained in the file specified by the first filename on the command line.
36(Note that systems supporting the #! notation invoke interpreters this
37way. See L<Location of Perl>.)
38
39=item 3.
40
41Passed in implicitly via standard input.  This works only if there are
42no filename arguments--to pass arguments to a STDIN-read program you
43must explicitly specify a "-" for the program name.
44
45=back
46
47With methods 2 and 3, Perl starts parsing the input file from the
48beginning, unless you've specified a B<-x> switch, in which case it
49scans for the first line starting with #! and containing the word
50"perl", and starts there instead.  This is useful for running a program
51embedded in a larger message.  (In this case you would indicate the end
52of the program using the C<__END__> token.)
53
54The #! line is always examined for switches as the line is being
55parsed.  Thus, if you're on a machine that allows only one argument
56with the #! line, or worse, doesn't even recognize the #! line, you
57still can get consistent switch behavior regardless of how Perl was
58invoked, even if B<-x> was used to find the beginning of the program.
59
60Because historically some operating systems silently chopped off
61kernel interpretation of the #! line after 32 characters, some
62switches may be passed in on the command line, and some may not;
63you could even get a "-" without its letter, if you're not careful.
64You probably want to make sure that all your switches fall either
65before or after that 32-character boundary.  Most switches don't
66actually care if they're processed redundantly, but getting a "-"
67instead of a complete switch could cause Perl to try to execute
68standard input instead of your program.  And a partial B<-I> switch
69could also cause odd results.
70
71Some switches do care if they are processed twice, for instance
72combinations of B<-l> and B<-0>.  Either put all the switches after
73the 32-character boundary (if applicable), or replace the use of
74B<-0>I<digits> by C<BEGIN{ $/ = "\0digits"; }>.
75
76Parsing of the #! switches starts wherever "perl" is mentioned in the line.
77The sequences "-*" and "- " are specifically ignored so that you could,
78if you were so inclined, say
79
80    #!/bin/sh -- # -*- perl -*- -p
81    eval 'exec perl -wS $0 ${1+"$@"}'
82        if $running_under_some_shell;
83
84to let Perl see the B<-p> switch. 
85
86A similar trick involves the B<env> program, if you have it.
87
88    #!/usr/bin/env perl
89
90The examples above use a relative path to the perl interpreter,
91getting whatever version is first in the user's path.  If you want
92a specific version of Perl, say, perl5.005_57, you should place
93that directly in the #! line's path.
94
95If the #! line does not contain the word "perl", the program named after
96the #! is executed instead of the Perl interpreter.  This is slightly
97bizarre, but it helps people on machines that don't do #!, because they
98can tell a program that their SHELL is F</usr/bin/perl>, and Perl will then
99dispatch the program to the correct interpreter for them.
100
101After locating your program, Perl compiles the entire program to an
102internal form.  If there are any compilation errors, execution of the
103program is not attempted.  (This is unlike the typical shell script,
104which might run part-way through before finding a syntax error.)
105
106If the program is syntactically correct, it is executed.  If the program
107runs off the end without hitting an exit() or die() operator, an implicit
108C<exit(0)> is provided to indicate successful completion.
109
110=head2 #! and quoting on non-Unix systems
111
112Unix's #! technique can be simulated on other systems:
113
114=over 4
115
116=item OS/2
117
118Put
119
120    extproc perl -S -your_switches
121
122as the first line in C<*.cmd> file (B<-S> due to a bug in cmd.exe's
123`extproc' handling).
124
125=item MS-DOS
126
127Create a batch file to run your program, and codify it in
128C<ALTERNATIVE_SHEBANG> (see the F<dosish.h> file in the source
129distribution for more information).
130
131=item Win95/NT
132
133The Win95/NT installation, when using the ActiveState installer for Perl,
134will modify the Registry to associate the F<.pl> extension with the perl
135interpreter.  If you install Perl by other means (including building from
136the sources), you may have to modify the Registry yourself.  Note that
137this means you can no longer tell the difference between an executable
138Perl program and a Perl library file.
139
140=item Macintosh
141
142A Macintosh perl program will have the appropriate Creator and
143Type, so that double-clicking them will invoke the perl application.
144
145=item VMS
146
147Put
148
149    $ perl -mysw 'f$env("procedure")' 'p1' 'p2' 'p3' 'p4' 'p5' 'p6' 'p7' 'p8' !
150    $ exit++ + ++$status != 0 and $exit = $status = undef;
151
152at the top of your program, where B<-mysw> are any command line switches you
153want to pass to Perl.  You can now invoke the program directly, by saying
154C<perl program>, or as a DCL procedure, by saying C<@program> (or implicitly
155via F<DCL$PATH> by just using the name of the program).
156
157This incantation is a bit much to remember, but Perl will display it for
158you if you say C<perl "-V:startperl">.
159
160=back
161
162Command-interpreters on non-Unix systems have rather different ideas
163on quoting than Unix shells.  You'll need to learn the special
164characters in your command-interpreter (C<*>, C<\> and C<"> are
165common) and how to protect whitespace and these characters to run
166one-liners (see B<-e> below).
167
168On some systems, you may have to change single-quotes to double ones,
169which you must I<not> do on Unix or Plan9 systems.  You might also
170have to change a single % to a %%.
171
172For example:
173
174    # Unix
175    perl -e 'print "Hello world\n"'
176
177    # MS-DOS, etc.
178    perl -e "print \"Hello world\n\""
179
180    # Macintosh
181    print "Hello world\n"
182     (then Run "Myscript" or Shift-Command-R)
183
184    # VMS
185    perl -e "print ""Hello world\n"""
186
187The problem is that none of this is reliable: it depends on the
188command and it is entirely possible neither works.  If B<4DOS> were
189the command shell, this would probably work better:
190
191    perl -e "print <Ctrl-x>"Hello world\n<Ctrl-x>""
192
193B<CMD.EXE> in Windows NT slipped a lot of standard Unix functionality in
194when nobody was looking, but just try to find documentation for its
195quoting rules.
196
197Under the Macintosh, it depends which environment you are using.  The MacPerl
198shell, or MPW, is much like Unix shells in its support for several
199quoting variants, except that it makes free use of the Macintosh's non-ASCII
200characters as control characters.
201
202There is no general solution to all of this.  It's just a mess.
203
204=head2 Location of Perl
205
206It may seem obvious to say, but Perl is useful only when users can
207easily find it.  When possible, it's good for both F</usr/bin/perl>
208and F</usr/local/bin/perl> to be symlinks to the actual binary.  If
209that can't be done, system administrators are strongly encouraged
210to put (symlinks to) perl and its accompanying utilities into a
211directory typically found along a user's PATH, or in some other
212obvious and convenient place.
213
214In this documentation, C<#!/usr/bin/perl> on the first line of the program
215will stand in for whatever method works on your system.  You are
216advised to use a specific path if you care about a specific version.
217
218    #!/usr/local/bin/perl5.00554
219
220or if you just want to be running at least version, place a statement
221like this at the top of your program:
222
223    use 5.005_54;
224
225=head2 Command Switches
226
227As with all standard commands, a single-character switch may be
228clustered with the following switch, if any.
229
230    #!/usr/bin/perl -spi.orig   # same as -s -p -i.orig
231
232Switches include:
233
234=over 5
235
236=item B<-0>[I<digits>]
237
238specifies the input record separator (C<$/>) as an octal number.  If there are
239no digits, the null character is the separator.  Other switches may
240precede or follow the digits.  For example, if you have a version of
241B<find> which can print filenames terminated by the null character, you
242can say this:
243
244    find . -name '*.orig' -print0 | perl -n0e unlink
245
246The special value 00 will cause Perl to slurp files in paragraph mode.
247The value 0777 will cause Perl to slurp files whole because there is no
248legal character with that value.
249
250=item B<-a>
251
252turns on autosplit mode when used with a B<-n> or B<-p>.  An implicit
253split command to the @F array is done as the first thing inside the
254implicit while loop produced by the B<-n> or B<-p>.
255
256    perl -ane 'print pop(@F), "\n";'
257
258is equivalent to
259
260    while (<>) {
261        @F = split(' ');
262        print pop(@F), "\n";
263    }
264
265An alternate delimiter may be specified using B<-F>.
266
267=item B<-C>
268
269enables Perl to use the native wide character APIs on the target system.
270The magic variable C<${^WIDE_SYSTEM_CALLS}> reflects the state of
271this switch.  See L<perlvar/"${^WIDE_SYSTEM_CALLS}">.
272
273This feature is currently only implemented on the Win32 platform.
274
275=item B<-c>
276
277causes Perl to check the syntax of the program and then exit without
278executing it.  Actually, it I<will> execute C<BEGIN>, C<CHECK>, and
279C<use> blocks, because these are considered as occurring outside the
280execution of your program.  C<INIT> and C<END> blocks, however, will
281be skipped.
282
283=item B<-d>
284
285runs the program under the Perl debugger.  See L<perldebug>.
286
287=item B<-d:>I<foo[=bar,baz]>
288
289runs the program under the control of a debugging, profiling, or
290tracing module installed as Devel::foo. E.g., B<-d:DProf> executes
291the program using the Devel::DProf profiler.  As with the B<-M>
292flag, options may be passed to the Devel::foo package where they
293will be received and interpreted by the Devel::foo::import routine.
294The comma-separated list of options must follow a C<=> character.
295See L<perldebug>.
296
297=item B<-D>I<letters>
298
299=item B<-D>I<number>
300
301sets debugging flags.  To watch how it executes your program, use
302B<-Dtls>.  (This works only if debugging is compiled into your
303Perl.)  Another nice value is B<-Dx>, which lists your compiled
304syntax tree.  And B<-Dr> displays compiled regular expressions. As an
305alternative, specify a number instead of list of letters (e.g., B<-D14> is
306equivalent to B<-Dtls>):
307
308        1  p  Tokenizing and parsing
309        2  s  Stack snapshots
310        4  l  Context (loop) stack processing
311        8  t  Trace execution
312       16  o  Method and overloading resolution
313       32  c  String/numeric conversions
314       64  P  Print preprocessor command for -P, source file input state
315      128  m  Memory allocation
316      256  f  Format processing
317      512  r  Regular expression parsing and execution
318     1024  x  Syntax tree dump
319     2048  u  Tainting checks
320     4096  L  Memory leaks (needs -DLEAKTEST when compiling Perl)
321     8192  H  Hash dump -- usurps values()
322    16384  X  Scratchpad allocation
323    32768  D  Cleaning up
324    65536  S  Thread synchronization
325   131072  T  Tokenising
326
327All these flags require B<-DDEBUGGING> when you compile the Perl
328executable.  See the F<INSTALL> file in the Perl source distribution
329for how to do this.  This flag is automatically set if you include B<-g>
330option when C<Configure> asks you about optimizer/debugger flags.
331
332If you're just trying to get a print out of each line of Perl code
333as it executes, the way that C<sh -x> provides for shell scripts,
334you can't use Perl's B<-D> switch.  Instead do this
335
336  # Bourne shell syntax
337  $ PERLDB_OPTS="NonStop=1 AutoTrace=1 frame=2" perl -dS program
338
339  # csh syntax
340  % (setenv PERLDB_OPTS "NonStop=1 AutoTrace=1 frame=2"; perl -dS program)
341
342See L<perldebug> for details and variations.
343
344=item B<-e> I<commandline>
345
346may be used to enter one line of program.  If B<-e> is given, Perl
347will not look for a filename in the argument list.  Multiple B<-e>
348commands may be given to build up a multi-line script.  Make sure
349to use semicolons where you would in a normal program.
350
351=item B<-F>I<pattern>
352
353specifies the pattern to split on if B<-a> is also in effect.  The
354pattern may be surrounded by C<//>, C<"">, or C<''>, otherwise it will be
355put in single quotes.
356
357=item B<-h>
358
359prints a summary of the options.
360
361=item B<-i>[I<extension>]
362
363specifies that files processed by the C<E<lt>E<gt>> construct are to be
364edited in-place.  It does this by renaming the input file, opening the
365output file by the original name, and selecting that output file as the
366default for print() statements.  The extension, if supplied, is used to
367modify the name of the old file to make a backup copy, following these
368rules:
369
370If no extension is supplied, no backup is made and the current file is
371overwritten.
372
373If the extension doesn't contain a C<*>, then it is appended to the
374end of the current filename as a suffix.  If the extension does
375contain one or more C<*> characters, then each C<*> is replaced
376with the current filename.  In Perl terms, you could think of this
377as:
378
379    ($backup = $extension) =~ s/\*/$file_name/g;
380
381This allows you to add a prefix to the backup file, instead of (or in
382addition to) a suffix:
383
384    $ perl -pi 'orig_*' -e 's/bar/baz/' fileA   # backup to 'orig_fileA'
385
386Or even to place backup copies of the original files into another
387directory (provided the directory already exists):
388
389    $ perl -pi 'old/*.orig' -e 's/bar/baz/' fileA # backup to 'old/fileA.orig'
390
391These sets of one-liners are equivalent:
392
393    $ perl -pi -e 's/bar/baz/' fileA            # overwrite current file
394    $ perl -pi '*' -e 's/bar/baz/' fileA        # overwrite current file
395
396    $ perl -pi '.orig' -e 's/bar/baz/' fileA    # backup to 'fileA.orig'
397    $ perl -pi '*.orig' -e 's/bar/baz/' fileA   # backup to 'fileA.orig'
398
399From the shell, saying
400
401    $ perl -p -i.orig -e "s/foo/bar/; ... "
402
403is the same as using the program:
404
405    #!/usr/bin/perl -pi.orig
406    s/foo/bar/;
407
408which is equivalent to
409
410    #!/usr/bin/perl
411    $extension = '.orig';
412    LINE: while (<>) {
413        if ($ARGV ne $oldargv) {
414            if ($extension !~ /\*/) {
415                $backup = $ARGV . $extension;
416            }
417            else {
418                ($backup = $extension) =~ s/\*/$ARGV/g;
419            }
420            rename($ARGV, $backup);
421            open(ARGVOUT, ">$ARGV");
422            select(ARGVOUT);
423            $oldargv = $ARGV;
424        }
425        s/foo/bar/;
426    }
427    continue {
428        print;  # this prints to original filename
429    }
430    select(STDOUT);
431
432except that the B<-i> form doesn't need to compare $ARGV to $oldargv to
433know when the filename has changed.  It does, however, use ARGVOUT for
434the selected filehandle.  Note that STDOUT is restored as the default
435output filehandle after the loop.
436
437As shown above, Perl creates the backup file whether or not any output
438is actually changed.  So this is just a fancy way to copy files:
439
440    $ perl -p -i '/some/file/path/*' -e 1 file1 file2 file3...
441or
442    $ perl -p -i '.orig' -e 1 file1 file2 file3...
443
444You can use C<eof> without parentheses to locate the end of each input
445file, in case you want to append to each file, or reset line numbering
446(see example in L<perlfunc/eof>).
447
448If, for a given file, Perl is unable to create the backup file as
449specified in the extension then it will skip that file and continue on
450with the next one (if it exists).
451
452For a discussion of issues surrounding file permissions and B<-i>,
453see L<perlfaq5/Why does Perl let me delete read-only files?  Why does -i clobber protected files?  Isn't this a bug in Perl?>.
454
455You cannot use B<-i> to create directories or to strip extensions from
456files.
457
458Perl does not expand C<~> in filenames, which is good, since some
459folks use it for their backup files:
460
461    $ perl -pi~ -e 's/foo/bar/' file1 file2 file3...
462
463Finally, the B<-i> switch does not impede execution when no
464files are given on the command line.  In this case, no backup is made
465(the original file cannot, of course, be determined) and processing
466proceeds from STDIN to STDOUT as might be expected.
467
468=item B<-I>I<directory>
469
470Directories specified by B<-I> are prepended to the search path for
471modules (C<@INC>), and also tells the C preprocessor where to search for
472include files.  The C preprocessor is invoked with B<-P>; by default it
473searches /usr/include and /usr/lib/perl.
474
475=item B<-l>[I<octnum>]
476
477enables automatic line-ending processing.  It has two separate
478effects.  First, it automatically chomps C<$/> (the input record
479separator) when used with B<-n> or B<-p>.  Second, it assigns C<$\>
480(the output record separator) to have the value of I<octnum> so
481that any print statements will have that separator added back on.
482If I<octnum> is omitted, sets C<$\> to the current value of
483C<$/>.  For instance, to trim lines to 80 columns:
484
485    perl -lpe 'substr($_, 80) = ""'
486
487Note that the assignment C<$\ = $/> is done when the switch is processed,
488so the input record separator can be different than the output record
489separator if the B<-l> switch is followed by a B<-0> switch:
490
491    gnufind / -print0 | perl -ln0e 'print "found $_" if -p'
492
493This sets C<$\> to newline and then sets C<$/> to the null character.
494
495=item B<-m>[B<->]I<module>
496
497=item B<-M>[B<->]I<module>
498
499=item B<-M>[B<->]I<'module ...'>
500
501=item B<-[mM]>[B<->]I<module=arg[,arg]...>
502
503B<-m>I<module> executes C<use> I<module> C<();> before executing your
504program.
505
506B<-M>I<module> executes C<use> I<module> C<;> before executing your
507program.  You can use quotes to add extra code after the module name,
508e.g., C<'-Mmodule qw(foo bar)'>.
509
510If the first character after the B<-M> or B<-m> is a dash (C<->)
511then the 'use' is replaced with 'no'.
512
513A little builtin syntactic sugar means you can also say
514B<-mmodule=foo,bar> or B<-Mmodule=foo,bar> as a shortcut for
515C<'-Mmodule qw(foo bar)'>.  This avoids the need to use quotes when
516importing symbols.  The actual code generated by B<-Mmodule=foo,bar> is
517C<use module split(/,/,q{foo,bar})>.  Note that the C<=> form
518removes the distinction between B<-m> and B<-M>.
519
520=item B<-n>
521
522causes Perl to assume the following loop around your program, which
523makes it iterate over filename arguments somewhat like B<sed -n> or
524B<awk>:
525
526  LINE:
527    while (<>) {
528        ...             # your program goes here
529    }
530
531Note that the lines are not printed by default.  See B<-p> to have
532lines printed.  If a file named by an argument cannot be opened for
533some reason, Perl warns you about it and moves on to the next file.
534
535Here is an efficient way to delete all files older than a week:
536
537    find . -mtime +7 -print | perl -nle unlink
538
539This is faster than using the B<-exec> switch of B<find> because you don't
540have to start a process on every filename found.  It does suffer from
541the bug of mishandling newlines in pathnames, which you can fix if
542you
543
544C<BEGIN> and C<END> blocks may be used to capture control before or after
545the implicit program loop, just as in B<awk>.
546
547=item B<-p>
548
549causes Perl to assume the following loop around your program, which
550makes it iterate over filename arguments somewhat like B<sed>:
551
552
553  LINE:
554    while (<>) {
555        ...             # your program goes here
556    } continue {
557        print or die "-p destination: $!\n";
558    }
559
560If a file named by an argument cannot be opened for some reason, Perl
561warns you about it, and moves on to the next file.  Note that the
562lines are printed automatically.  An error occurring during printing is
563treated as fatal.  To suppress printing use the B<-n> switch.  A B<-p>
564overrides a B<-n> switch.
565
566C<BEGIN> and C<END> blocks may be used to capture control before or after
567the implicit loop, just as in B<awk>.
568
569=item B<-P>
570
571causes your program to be run through the C preprocessor before
572compilation by Perl.  Because both comments and B<cpp> directives begin
573with the # character, you should avoid starting comments with any words
574recognized by the C preprocessor such as C<"if">, C<"else">, or C<"define">.
575Also, in some platforms the C preprocessor knows too much: it knows
576about the C++ -style until-end-of-line comments starting with C<"//">.
577This will cause problems with common Perl constructs like
578
579    s/foo//;
580
581because after -P this will became illegal code
582
583    s/foo
584
585The workaround is to use some other quoting separator than C<"/">,
586like for example C<"!">:
587
588    s!foo!!;
589
590=item B<-s>
591
592enables rudimentary switch parsing for switches on the command
593line after the program name but before any filename arguments (or before
594an argument of B<-->).  This means you can have switches with two leading
595dashes (B<--help>).  Any switch found there is removed from @ARGV and sets the
596corresponding variable in the Perl program.  The following program
597prints "1" if the program is invoked with a B<-xyz> switch, and "abc"
598if it is invoked with B<-xyz=abc>.
599
600    #!/usr/bin/perl -s
601    if ($xyz) { print "$xyz\n" }
602
603Do note that B<--help> creates the variable ${-help}, which is not compliant
604with C<strict refs>.
605
606=item B<-S>
607
608makes Perl use the PATH environment variable to search for the
609program (unless the name of the program contains directory separators).
610
611On some platforms, this also makes Perl append suffixes to the
612filename while searching for it.  For example, on Win32 platforms,
613the ".bat" and ".cmd" suffixes are appended if a lookup for the
614original name fails, and if the name does not already end in one
615of those suffixes.  If your Perl was compiled with DEBUGGING turned
616on, using the -Dp switch to Perl shows how the search progresses.
617
618Typically this is used to emulate #! startup on platforms that
619don't support #!.  This example works on many platforms that
620have a shell compatible with Bourne shell:
621
622    #!/usr/bin/perl
623    eval 'exec /usr/bin/perl -wS $0 ${1+"$@"}'
624            if $running_under_some_shell;
625
626The system ignores the first line and feeds the program to F</bin/sh>,
627which proceeds to try to execute the Perl program as a shell script.
628The shell executes the second line as a normal shell command, and thus
629starts up the Perl interpreter.  On some systems $0 doesn't always
630contain the full pathname, so the B<-S> tells Perl to search for the
631program if necessary.  After Perl locates the program, it parses the
632lines and ignores them because the variable $running_under_some_shell
633is never true.  If the program will be interpreted by csh, you will need
634to replace C<${1+"$@"}> with C<$*>, even though that doesn't understand
635embedded spaces (and such) in the argument list.  To start up sh rather
636than csh, some systems may have to replace the #! line with a line
637containing just a colon, which will be politely ignored by Perl.  Other
638systems can't control that, and need a totally devious construct that
639will work under any of B<csh>, B<sh>, or Perl, such as the following:
640
641        eval '(exit $?0)' && eval 'exec perl -wS $0 ${1+"$@"}'
642        & eval 'exec /usr/bin/perl -wS $0 $argv:q'
643                if $running_under_some_shell;
644
645If the filename supplied contains directory separators (i.e., is an
646absolute or relative pathname), and if that file is not found,
647platforms that append file extensions will do so and try to look
648for the file with those extensions added, one by one.
649
650On DOS-like platforms, if the program does not contain directory
651separators, it will first be searched for in the current directory
652before being searched for on the PATH.  On Unix platforms, the
653program will be searched for strictly on the PATH.
654
655=item B<-T>
656
657forces "taint" checks to be turned on so you can test them.  Ordinarily
658these checks are done only when running setuid or setgid.  It's a
659good idea to turn them on explicitly for programs that run on behalf
660of someone else whom you might not necessarily trust, such as CGI
661programs or any internet servers you might write in Perl.  See
662L<perlsec> for details.  For security reasons, this option must be
663seen by Perl quite early; usually this means it must appear early
664on the command line or in the #! line for systems which support
665that construct.
666
667=item B<-u>
668
669This obsolete switch causes Perl to dump core after compiling your
670program.  You can then in theory take this core dump and turn it
671into an executable file by using the B<undump> program (not supplied).
672This speeds startup at the expense of some disk space (which you
673can minimize by stripping the executable).  (Still, a "hello world"
674executable comes out to about 200K on my machine.)  If you want to
675execute a portion of your program before dumping, use the dump()
676operator instead.  Note: availability of B<undump> is platform
677specific and may not be available for a specific port of Perl.
678
679This switch has been superseded in favor of the new Perl code
680generator backends to the compiler.  See L<B> and L<B::Bytecode>
681for details.
682
683=item B<-U>
684
685allows Perl to do unsafe operations.  Currently the only "unsafe"
686operations are the unlinking of directories while running as superuser,
687and running setuid programs with fatal taint checks turned into
688warnings.  Note that the B<-w> switch (or the C<$^W> variable) must
689be used along with this option to actually I<generate> the
690taint-check warnings.
691
692=item B<-v>
693
694prints the version and patchlevel of your perl executable.
695
696=item B<-V>
697
698prints summary of the major perl configuration values and the current
699values of @INC.
700
701=item B<-V:>I<name>
702
703Prints to STDOUT the value of the named configuration variable.
704For example,
705
706    $ perl -V:man.dir
707
708will provide strong clues about what your MANPATH variable should
709be set to in order to access the Perl documentation.
710
711=item B<-w>
712
713prints warnings about dubious constructs, such as variable names
714that are mentioned only once and scalar variables that are used
715before being set, redefined subroutines, references to undefined
716filehandles or filehandles opened read-only that you are attempting
717to write on, values used as a number that doesn't look like numbers,
718using an array as though it were a scalar, if your subroutines
719recurse more than 100 deep, and innumerable other things.
720
721This switch really just enables the internal C<^$W> variable.  You
722can disable or promote into fatal errors specific warnings using
723C<__WARN__> hooks, as described in L<perlvar> and L<perlfunc/warn>.
724See also L<perldiag> and L<perltrap>.  A new, fine-grained warning
725facility is also available if you want to manipulate entire classes
726of warnings; see L<warnings> or L<perllexwarn>.
727
728=item B<-W>
729
730Enables all warnings regardless of C<no warnings> or C<$^W>.
731See L<perllexwarn>.
732
733=item B<-X>
734
735Disables all warnings regardless of C<use warnings> or C<$^W>.
736See L<perllexwarn>.
737
738=item B<-x> I<directory>
739
740tells Perl that the program is embedded in a larger chunk of unrelated
741ASCII text, such as in a mail message.  Leading garbage will be
742discarded until the first line that starts with #! and contains the
743string "perl".  Any meaningful switches on that line will be applied.
744If a directory name is specified, Perl will switch to that directory
745before running the program.  The B<-x> switch controls only the
746disposal of leading garbage.  The program must be terminated with
747C<__END__> if there is trailing garbage to be ignored (the program
748can process any or all of the trailing garbage via the DATA filehandle
749if desired).
750
751=back
752
753=head1 ENVIRONMENT
754
755=over 12
756
757=item HOME
758
759Used if chdir has no argument.
760
761=item LOGDIR
762
763Used if chdir has no argument and HOME is not set.
764
765=item PATH
766
767Used in executing subprocesses, and in finding the program if B<-S> is
768used.
769
770=item PERL5LIB
771
772A colon-separated list of directories in which to look for Perl library
773files before looking in the standard library and the current
774directory.  Any architecture-specific directories under the specified
775locations are automatically included if they exist.  If PERL5LIB is not
776defined, PERLLIB is used.
777
778When running taint checks (either because the program was running setuid
779or setgid, or the B<-T> switch was used), neither variable is used.
780The program should instead say:
781
782    use lib "/my/directory";
783
784=item PERL5OPT
785
786Command-line options (switches).  Switches in this variable are taken
787as if they were on every Perl command line.  Only the B<-[DIMUdmw]>
788switches are allowed.  When running taint checks (because the program
789was running setuid or setgid, or the B<-T> switch was used), this
790variable is ignored.  If PERL5OPT begins with B<-T>, tainting will be
791enabled, and any subsequent options ignored.
792
793=item PERLLIB
794
795A colon-separated list of directories in which to look for Perl library
796files before looking in the standard library and the current directory.
797If PERL5LIB is defined, PERLLIB is not used.
798
799=item PERL5DB
800
801The command used to load the debugger code.  The default is:
802
803        BEGIN { require 'perl5db.pl' }
804
805=item PERL5SHELL (specific to the Win32 port)
806
807May be set to an alternative shell that perl must use internally for
808executing "backtick" commands or system().  Default is C<cmd.exe /x/c>
809on WindowsNT and C<command.com /c> on Windows95.  The value is considered
810to be space-separated.  Precede any character that needs to be protected
811(like a space or backslash) with a backslash.
812
813Note that Perl doesn't use COMSPEC for this purpose because
814COMSPEC has a high degree of variability among users, leading to
815portability concerns.  Besides, perl can use a shell that may not be
816fit for interactive use, and setting COMSPEC to such a shell may
817interfere with the proper functioning of other programs (which usually
818look in COMSPEC to find a shell fit for interactive use).
819
820=item PERL_DEBUG_MSTATS
821
822Relevant only if perl is compiled with the malloc included with the perl
823distribution (that is, if C<perl -V:d_mymalloc> is 'define').
824If set, this causes memory statistics to be dumped after execution.  If set
825to an integer greater than one, also causes memory statistics to be dumped
826after compilation.
827
828=item PERL_DESTRUCT_LEVEL
829
830Relevant only if your perl executable was built with B<-DDEBUGGING>,
831this controls the behavior of global destruction of objects and other
832references.
833
834=item PERL_ROOT (specific to the VMS port)
835
836A translation concealed rooted logical name that contains perl and the
837logical device for the @INC path on VMS only.  Other logical names that
838affect perl on VMS include PERLSHR, PERL_ENV_TABLES, and
839SYS$TIMEZONE_DIFFERENTIAL but are optional and discussed further in
840L<perlvms> and in F<README.vms> in the Perl source distribution.
841
842=item SYS$LOGIN (specific to the VMS port)
843
844Used if chdir has no argument and HOME and LOGDIR are not set.
845
846=back
847
848Perl also has environment variables that control how Perl handles data
849specific to particular natural languages.  See L<perllocale>.
850
851Apart from these, Perl uses no other environment variables, except
852to make them available to the program being executed, and to child
853processes.  However, programs running setuid would do well to execute
854the following lines before doing anything else, just to keep people
855honest:
856
857    $ENV{PATH}  = '/bin:/usr/bin';    # or whatever you need
858    $ENV{SHELL} = '/bin/sh' if exists $ENV{SHELL};
859    delete @ENV{qw(IFS CDPATH ENV BASH_ENV)};
Note: See TracBrowser for help on using the repository browser.