3 perlrun - how to execute the Perl interpreter
7 B<perl> S<[ B<-sTuU> ]>
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...'> ]>
15 S<[ B<-i>[I<extension>] ]>
16 S<[ B<-e> I<'command'> ] [ B<--> ] [ I<programfile> ] [ I<argument> ]...>
20 Upon startup, Perl looks for your script in one of the following
27 Specified line by line via B<-e> switches on the command line.
31 Contained in the file specified by the first filename on the command line.
32 (Note that systems supporting the #! notation invoke interpreters this
33 way. See L<Location of Perl>.)
37 Passed in implicitly via standard input. This works only if there are
38 no filename arguments--to pass arguments to a STDIN script you
39 must explicitly specify a "-" for the script name.
43 With methods 2 and 3, Perl starts parsing the input file from the
44 beginning, unless you've specified a B<-x> switch, in which case it
45 scans for the first line starting with #! and containing the word
46 "perl", and starts there instead. This is useful for running a script
47 embedded in a larger message. (In this case you would indicate the end
48 of the script using the C<__END__> token.)
50 The #! line is always examined for switches as the line is being
51 parsed. Thus, if you're on a machine that allows only one argument
52 with the #! line, or worse, doesn't even recognize the #! line, you
53 still can get consistent switch behavior regardless of how Perl was
54 invoked, even if B<-x> was used to find the beginning of the script.
56 Because many operating systems silently chop off kernel interpretation of
57 the #! line after 32 characters, some switches may be passed in on the
58 command line, and some may not; you could even get a "-" without its
59 letter, if you're not careful. You probably want to make sure that all
60 your switches fall either before or after that 32 character boundary.
61 Most switches don't actually care if they're processed redundantly, but
62 getting a - instead of a complete switch could cause Perl to try to
63 execute standard input instead of your script. And a partial B<-I> switch
64 could also cause odd results.
66 Some switches do care if they are processed twice, for instance combinations
67 of B<-l> and B<-0>. Either put all the switches after the 32 character
68 boundary (if applicable), or replace the use of B<-0>I<digits> by
69 C<BEGIN{ $/ = "\0digits"; }>.
71 Parsing of the #! switches starts wherever "perl" is mentioned in the line.
72 The sequences "-*" and "- " are specifically ignored so that you could,
73 if you were so inclined, say
75 #!/bin/sh -- # -*- perl -*- -p
76 eval 'exec /usr/bin/perl -wS $0 ${1+"$@"}'
77 if $running_under_some_shell;
79 to let Perl see the B<-p> switch.
81 If the #! line does not contain the word "perl", the program named after
82 the #! is executed instead of the Perl interpreter. This is slightly
83 bizarre, but it helps people on machines that don't do #!, because they
84 can tell a program that their SHELL is /usr/bin/perl, and Perl will then
85 dispatch the program to the correct interpreter for them.
87 After locating your script, Perl compiles the entire script to an
88 internal form. If there are any compilation errors, execution of the
89 script is not attempted. (This is unlike the typical shell script,
90 which might run part-way through before finding a syntax error.)
92 If the script is syntactically correct, it is executed. If the script
93 runs off the end without hitting an exit() or die() operator, an implicit
94 C<exit(0)> is provided to indicate successful completion.
96 =head2 #! and quoting on non-Unix systems
98 Unix's #! technique can be simulated on other systems:
106 extproc perl -S -your_switches
108 as the first line in C<*.cmd> file (C<-S> due to a bug in cmd.exe's
113 Create a batch file to run your script, and codify it in
114 C<ALTERNATIVE_SHEBANG> (see the F<dosish.h> file in the source
115 distribution for more information).
119 The Win95/NT installation, when using the Activeware port of Perl,
120 will modify the Registry to associate the F<.pl> extension with the perl
121 interpreter. If you install another port of Perl, including the one
122 in the Win32 directory of the Perl distribution, then you'll have to
123 modify the Registry yourself. Note that this means you can no
124 longer tell the difference between an executable Perl program
125 and a Perl library file.
129 Macintosh perl scripts will have the appropriate Creator and
130 Type, so that double-clicking them will invoke the perl application.
134 Command-interpreters on non-Unix systems have rather different ideas
135 on quoting than Unix shells. You'll need to learn the special
136 characters in your command-interpreter (C<*>, C<\> and C<"> are
137 common) and how to protect whitespace and these characters to run
138 one-liners (see C<-e> below).
140 On some systems, you may have to change single-quotes to double ones,
141 which you must I<NOT> do on Unix or Plan9 systems. You might also
142 have to change a single % to a %%.
147 perl -e 'print "Hello world\n"'
150 perl -e "print \"Hello world\n\""
153 print "Hello world\n"
154 (then Run "Myscript" or Shift-Command-R)
157 perl -e "print ""Hello world\n"""
159 The problem is that none of this is reliable: it depends on the command
160 and it is entirely possible neither works. If 4DOS was the command shell, this would
161 probably work better:
163 perl -e "print <Ctrl-x>"Hello world\n<Ctrl-x>""
165 CMD.EXE in Windows NT slipped a lot of standard Unix functionality in
166 when nobody was looking, but just try to find documentation for its
169 Under the Macintosh, it depends which environment you are using. The MacPerl
170 shell, or MPW, is much like Unix shells in its support for several
171 quoting variants, except that it makes free use of the Macintosh's non-ASCII
172 characters as control characters.
174 There is no general solution to all of this. It's just a mess.
176 =head2 Location of Perl
178 It may seem obvious to say, but Perl is useful only when users can
179 easily find it. When possible, it's good for both B</usr/bin/perl> and
180 B</usr/local/bin/perl> to be symlinks to the actual binary. If that
181 can't be done, system administrators are strongly encouraged to put
182 (symlinks to) perl and its accompanying utilities, such as perldoc, into
183 a directory typically found along a user's PATH, or in another obvious
184 and convenient place.
186 In this documentation, C<#!/usr/bin/perl> on the first line of the script
187 will stand in for whatever method works on your system.
191 A single-character switch may be combined with the following switch, if
194 #!/usr/bin/perl -spi.bak # same as -s -p -i.bak
200 =item B<-0>[I<digits>]
202 specifies the input record separator (C<$/>) as an octal number. If there are
203 no digits, the null character is the separator. Other switches may
204 precede or follow the digits. For example, if you have a version of
205 B<find> which can print filenames terminated by the null character, you
208 find . -name '*.bak' -print0 | perl -n0e unlink
210 The special value 00 will cause Perl to slurp files in paragraph mode.
211 The value 0777 will cause Perl to slurp files whole because there is no
212 legal character with that value.
216 turns on autosplit mode when used with a B<-n> or B<-p>. An implicit
217 split command to the @F array is done as the first thing inside the
218 implicit while loop produced by the B<-n> or B<-p>.
220 perl -ane 'print pop(@F), "\n";'
229 An alternate delimiter may be specified using B<-F>.
233 causes Perl to check the syntax of the script and then exit without
234 executing it. Actually, it I<will> execute C<BEGIN>, C<END>, and C<use> blocks,
235 because these are considered as occurring outside the execution of
240 runs the script under the Perl debugger. See L<perldebug>.
244 runs the script under the control of a debugging or tracing module
245 installed as Devel::foo. E.g., B<-d:DProf> executes the script using the
246 Devel::DProf profiler. See L<perldebug>.
248 =item B<-D>I<letters>
252 sets debugging flags. To watch how it executes your script, use
253 B<-Dtls>. (This works only if debugging is compiled into your
254 Perl.) Another nice value is B<-Dx>, which lists your compiled
255 syntax tree. And B<-Dr> displays compiled regular expressions. As an
256 alternative, specify a number instead of list of letters (e.g., B<-D14> is
257 equivalent to B<-Dtls>):
259 1 p Tokenizing and parsing
261 4 l Context (loop) stack processing
263 16 o Method and overloading resolution
264 32 c String/numeric conversions
265 64 P Print preprocessor command for -P
266 128 m Memory allocation
267 256 f Format processing
268 512 r Regular expression parsing and execution
269 1024 x Syntax tree dump
270 2048 u Tainting checks
271 4096 L Memory leaks (needs C<-DLEAKTEST> when compiling Perl)
272 8192 H Hash dump -- usurps values()
273 16384 X Scratchpad allocation
275 65536 S Thread synchronization
277 All these flags require C<-DDEBUGGING> when you compile the Perl
278 executable. This flag is automatically set if you include C<-g>
279 option when C<Configure> asks you about optimizer/debugger flags.
281 =item B<-e> I<commandline>
283 may be used to enter one line of script.
284 If B<-e> is given, Perl
285 will not look for a script filename in the argument list.
286 Multiple B<-e> commands may
287 be given to build up a multi-line script.
288 Make sure to use semicolons where you would in a normal program.
290 =item B<-F>I<pattern>
292 specifies the pattern to split on if B<-a> is also in effect. The
293 pattern may be surrounded by C<//>, C<"">, or C<''>, otherwise it will be
294 put in single quotes.
298 prints a summary of the options.
300 =item B<-i>[I<extension>]
302 specifies that files processed by the C<E<lt>E<gt>> construct are to be
303 edited in-place. It does this by renaming the input file, opening the
304 output file by the original name, and selecting that output file as the
305 default for print() statements. The extension, if supplied, is used to
306 modify the name of the old file to make a backup copy, following these
309 If no extension is supplied, no backup is made and the current file is
312 If the extension doesn't contain a C<*> then it is appended to the end
313 of the current filename as a suffix.
315 If the extension does contain one or more C<*> characters, then each C<*>
316 is replaced with the current filename. In perl terms you could think of
319 ($backup = $extension) =~ s/\*/$file_name/g;
321 This allows you to add a prefix to the backup file, instead of (or in
322 addition to) a suffix:
324 $ perl -pi'bak_*' -e 's/bar/baz/' fileA # backup to 'bak_fileA'
326 Or even to place backup copies of the original files into another
327 directory (provided the directory already exists):
329 $ perl -pi'old/*.bak' -e 's/bar/baz/' fileA # backup to 'old/fileA.bak'
331 These sets of one-liners are equivalent:
333 $ perl -pi -e 's/bar/baz/' fileA # overwrite current file
334 $ perl -pi'*' -e 's/bar/baz/' fileA # overwrite current file
336 $ perl -pi'.bak' -e 's/bar/baz/' fileA # backup to 'fileA.bak'
337 $ perl -pi'*.bak' -e 's/bar/baz/' fileA # backup to 'fileA.bak'
339 From the shell, saying
341 $ perl -p -i.bak -e "s/foo/bar/; ... "
343 is the same as using the script:
345 #!/usr/bin/perl -pi.bak
348 which is equivalent to
353 if ($ARGV ne $oldargv) {
354 if ($extension !~ /\*/) {
355 $backup = $ARGV . $extension;
358 ($backup = $extension) =~ s/\*/$ARGV/g;
360 rename($ARGV, $backup);
361 open(ARGVOUT, ">$ARGV");
368 print; # this prints to original filename
372 except that the B<-i> form doesn't need to compare $ARGV to $oldargv to
373 know when the filename has changed. It does, however, use ARGVOUT for
374 the selected filehandle. Note that STDOUT is restored as the default
375 output filehandle after the loop.
377 As shown above, Perl creates the backup file whether or not any output
378 is actually changed. So this is just a fancy way to copy files:
380 $ perl -p -i'/some/file/path/*' -e 1 file1 file2 file3...
382 $ perl -p -i'.bak' -e 1 file1 file2 file3...
384 You can use C<eof> without parentheses to locate the end of each input
385 file, in case you want to append to each file, or reset line numbering
386 (see example in L<perlfunc/eof>).
388 If, for a given file, Perl is unable to create the backup file as
389 specified in the extension then it will skip that file and continue on
390 with the next one (if it exists).
392 For a discussion of issues surrounding file permissions and C<-i>, see
393 L<perlfaq5/Why does Perl let me delete read-only files? Why does -i clobber protected files? Isn't this a bug in Perl?>.
395 You cannot use B<-i> to create directories or to strip extensions from
398 Perl does not expand C<~>, so don't do that.
400 Finally, note that the B<-i> switch does not impede execution when no
401 files are given on the command line. In this case, no backup is made
402 (the original file cannot, of course, be determined) and processing
403 proceeds from STDIN to STDOUT as might be expected.
405 =item B<-I>I<directory>
407 Directories specified by B<-I> are prepended to the search path for
408 modules (C<@INC>), and also tells the C preprocessor where to search for
409 include files. The C preprocessor is invoked with B<-P>; by default it
410 searches /usr/include and /usr/lib/perl.
412 =item B<-l>[I<octnum>]
414 enables automatic line-ending processing. It has two effects: first,
415 it automatically chomps "C<$/>" (the input record separator) when used
416 with B<-n> or B<-p>, and second, it assigns "C<$\>"
417 (the output record separator) to have the value of I<octnum> so that
418 any print statements will have that separator added back on. If
419 I<octnum> is omitted, sets "C<$\>" to the current value of "C<$/>". For
420 instance, to trim lines to 80 columns:
422 perl -lpe 'substr($_, 80) = ""'
424 Note that the assignment C<$\ = $/> is done when the switch is processed,
425 so the input record separator can be different than the output record
426 separator if the B<-l> switch is followed by a B<-0> switch:
428 gnufind / -print0 | perl -ln0e 'print "found $_" if -p'
430 This sets C<$\> to newline and then sets C<$/> to the null character.
432 =item B<-m>[B<->]I<module>
434 =item B<-M>[B<->]I<module>
436 =item B<-M>[B<->]I<'module ...'>
438 =item B<-[mM]>[B<->]I<module=arg[,arg]...>
440 C<-m>I<module> executes C<use> I<module> C<();> before executing your
443 C<-M>I<module> executes C<use> I<module> C<;> before executing your
444 script. You can use quotes to add extra code after the module name,
445 e.g., C<-M'module qw(foo bar)'>.
447 If the first character after the C<-M> or C<-m> is a dash (C<->)
448 then the 'use' is replaced with 'no'.
450 A little builtin syntactic sugar means you can also say
451 C<-mmodule=foo,bar> or C<-Mmodule=foo,bar> as a shortcut for
452 C<-M'module qw(foo bar)'>. This avoids the need to use quotes when
453 importing symbols. The actual code generated by C<-Mmodule=foo,bar> is
454 C<use module split(/,/,q{foo,bar})>. Note that the C<=> form
455 removes the distinction between C<-m> and C<-M>.
459 causes Perl to assume the following loop around your script, which
460 makes it iterate over filename arguments somewhat like B<sed -n> or
464 ... # your script goes here
467 Note that the lines are not printed by default. See B<-p> to have
468 lines printed. If a file named by an argument cannot be opened for
469 some reason, Perl warns you about it, and moves on to the next file.
471 Here is an efficient way to delete all files older than a week:
473 find . -mtime +7 -print | perl -nle 'unlink;'
475 This is faster than using the C<-exec> switch of B<find> because you don't
476 have to start a process on every filename found.
478 C<BEGIN> and C<END> blocks may be used to capture control before or after
479 the implicit loop, just as in B<awk>.
483 causes Perl to assume the following loop around your script, which
484 makes it iterate over filename arguments somewhat like B<sed>:
488 ... # your script goes here
490 print or die "-p destination: $!\n";
493 If a file named by an argument cannot be opened for some reason, Perl
494 warns you about it, and moves on to the next file. Note that the
495 lines are printed automatically. An error occuring during printing is
496 treated as fatal. To suppress printing use the B<-n> switch. A B<-p>
497 overrides a B<-n> switch.
499 C<BEGIN> and C<END> blocks may be used to capture control before or after
500 the implicit loop, just as in awk.
504 causes your script to be run through the C preprocessor before
505 compilation by Perl. (Because both comments and cpp directives begin
506 with the # character, you should avoid starting comments with any words
507 recognized by the C preprocessor such as "if", "else", or "define".)
511 enables some rudimentary switch parsing for switches on the command
512 line after the script name but before any filename arguments (or before
513 a B<-->). Any switch found there is removed from @ARGV and sets the
514 corresponding variable in the Perl script. The following script
515 prints "true" if and only if the script is invoked with a B<-xyz> switch.
518 if ($xyz) { print "true\n"; }
522 makes Perl use the PATH environment variable to search for the
523 script (unless the name of the script contains directory separators).
524 On some platforms, this also makes Perl append suffixes to the
525 filename while searching for it. For example, on Win32 platforms,
526 the ".bat" and ".cmd" suffixes are appended if a lookup for the
527 original name fails, and if the name does not already end in one
528 of those suffixes. If your Perl was compiled with DEBUGGING turned
529 on, using the -Dp switch to Perl shows how the search progresses.
531 If the filename supplied contains directory separators (i.e. it is an
532 absolute or relative pathname), and if the file is not found,
533 platforms that append file extensions will do so and try to look
534 for the file with those extensions added, one by one.
536 On DOS-like platforms, if the script does not contain directory
537 separators, it will first be searched for in the current directory
538 before being searched for on the PATH. On Unix platforms, the
539 script will be searched for strictly on the PATH.
541 Typically this is used to emulate #! startup on platforms that
542 don't support #!. This example works on many platforms that
543 have a shell compatible with Bourne shell:
546 eval 'exec /usr/bin/perl -wS $0 ${1+"$@"}'
547 if $running_under_some_shell;
549 The system ignores the first line and feeds the script to /bin/sh,
550 which proceeds to try to execute the Perl script as a shell script.
551 The shell executes the second line as a normal shell command, and thus
552 starts up the Perl interpreter. On some systems $0 doesn't always
553 contain the full pathname, so the B<-S> tells Perl to search for the
554 script if necessary. After Perl locates the script, it parses the
555 lines and ignores them because the variable $running_under_some_shell
556 is never true. If the script will be interpreted by csh, you will need
557 to replace C<${1+"$@"}> with C<$*>, even though that doesn't understand
558 embedded spaces (and such) in the argument list. To start up sh rather
559 than csh, some systems may have to replace the #! line with a line
560 containing just a colon, which will be politely ignored by Perl. Other
561 systems can't control that, and need a totally devious construct that
562 will work under any of csh, sh, or Perl, such as the following:
564 eval '(exit $?0)' && eval 'exec /usr/bin/perl -wS $0 ${1+"$@"}'
565 & eval 'exec /usr/bin/perl -wS $0 $argv:q'
566 if $running_under_some_shell;
570 forces "taint" checks to be turned on so you can test them. Ordinarily
571 these checks are done only when running setuid or setgid. It's a good
572 idea to turn them on explicitly for programs run on another's behalf,
573 such as CGI programs. See L<perlsec>. Note that (for security reasons)
574 this option must be seen by Perl quite early; usually this means it must
575 appear early on the command line or in the #! line (for systems which
580 causes Perl to dump core after compiling your script. You can then
581 in theory take this core dump and turn it into an executable file by using the
582 B<undump> program (not supplied). This speeds startup at the expense of
583 some disk space (which you can minimize by stripping the executable).
584 (Still, a "hello world" executable comes out to about 200K on my
585 machine.) If you want to execute a portion of your script before dumping,
586 use the dump() operator instead. Note: availability of B<undump> is
587 platform specific and may not be available for a specific port of
588 Perl. It has been superseded by the new perl-to-C compiler, which is more
589 portable, even though it's still only considered beta.
593 allows Perl to do unsafe operations. Currently the only "unsafe"
594 operations are the unlinking of directories while running as superuser,
595 and running setuid programs with fatal taint checks turned into
596 warnings. Note that the B<-w> switch (or the C<$^W> variable) must
597 be used along with this option to actually B<generate> the
598 taint-check warnings.
602 prints the version and patchlevel of your Perl executable.
606 prints summary of the major perl configuration values and the current
611 Prints to STDOUT the value of the named configuration variable.
615 prints warnings about variable names that are mentioned only once, and
616 scalar variables that are used before being set. Also warns about
617 redefined subroutines, and references to undefined filehandles or
618 filehandles opened read-only that you are attempting to write on. Also
619 warns you if you use values as a number that doesn't look like numbers,
620 using an array as though it were a scalar, if your subroutines recurse
621 more than 100 deep, and innumerable other things.
623 You can disable specific warnings using C<__WARN__> hooks, as described
624 in L<perlvar> and L<perlfunc/warn>. See also L<perldiag> and L<perltrap>.
626 =item B<-x> I<directory>
628 tells Perl that the script is embedded in a message. Leading
629 garbage will be discarded until the first line that starts with #! and
630 contains the string "perl". Any meaningful switches on that line will
631 be applied. If a directory name is specified, Perl will switch to
632 that directory before running the script. The B<-x> switch controls
633 only the disposal of leading garbage. The script must be
634 terminated with C<__END__> if there is trailing garbage to be ignored (the
635 script can process any or all of the trailing garbage via the DATA
636 filehandle if desired).
646 Used if chdir has no argument.
650 Used if chdir has no argument and HOME is not set.
654 Used in executing subprocesses, and in finding the script if B<-S> is
659 A colon-separated list of directories in which to look for Perl library
660 files before looking in the standard library and the current
661 directory. If PERL5LIB is not defined, PERLLIB is used. When running
662 taint checks (because the script was running setuid or setgid, or the
663 B<-T> switch was used), neither variable is used. The script should
666 use lib "/my/directory";
670 Command-line options (switches). Switches in this variable are taken
671 as if they were on every Perl command line. Only the B<-[DIMUdmw]>
672 switches are allowed. When running taint checks (because the script
673 was running setuid or setgid, or the B<-T> switch was used), this
678 A colon-separated list of directories in which to look for Perl library
679 files before looking in the standard library and the current directory.
680 If PERL5LIB is defined, PERLLIB is not used.
684 The command used to load the debugger code. The default is:
686 BEGIN { require 'perl5db.pl' }
688 =item PERL5SHELL (specific to WIN32 port)
690 May be set to an alternative shell that perl must use internally for
691 executing "backtick" commands or system(). Default is C<cmd.exe /x/c>
692 on WindowsNT and C<command.com /c> on Windows95. The value is considered
693 to be space delimited. Precede any character that needs to be protected
694 (like a space or backslash) with a backslash.
696 Note that Perl doesn't use COMSPEC for this purpose because
697 COMSPEC has a high degree of variability among users, leading to
698 portability concerns. Besides, perl can use a shell that may not be
699 fit for interactive use, and setting COMSPEC to such a shell may
700 interfere with the proper functioning of other programs (which usually
701 look in COMSPEC to find a shell fit for interactive use).
703 =item PERL_DEBUG_MSTATS
705 Relevant only if perl is compiled with the malloc included with the perl
706 distribution (that is, if C<perl -V:d_mymalloc> is 'define').
707 If set, this causes memory statistics to be dumped after execution. If set
708 to an integer greater than one, also causes memory statistics to be dumped
711 =item PERL_DESTRUCT_LEVEL
713 Relevant only if your perl executable was built with B<-DDEBUGGING>,
714 this controls the behavior of global destruction of objects and other
719 Perl also has environment variables that control how Perl handles data
720 specific to particular natural languages. See L<perllocale>.
722 Apart from these, Perl uses no other environment variables, except
723 to make them available to the script being executed, and to child
724 processes. However, scripts running setuid would do well to execute
725 the following lines before doing anything else, just to keep people
728 $ENV{PATH} = '/bin:/usr/bin'; # or whatever you need
729 $ENV{SHELL} = '/bin/sh' if exists $ENV{SHELL};
730 delete @ENV{qw(IFS CDPATH ENV BASH_ENV)};