This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Re: [ID 20000816.012] *foo = *_ is broken
[perl5.git] / pod / perlvar.pod
CommitLineData
a0d0e21e
LW
1=head1 NAME
2
3perlvar - Perl predefined variables
4
5=head1 DESCRIPTION
6
7=head2 Predefined Names
8
5a964f20 9The following names have special meaning to Perl. Most
14218588
GS
10punctuation names have reasonable mnemonics, or analogs in the
11shells. Nevertheless, if you wish to use long variable names,
12you need only say
a0d0e21e
LW
13
14 use English;
15
a1ce9542
JF
16at the top of your program. This aliases all the short names to the long
17names in the current package. Some even have medium names, generally
18borrowed from B<awk>. In general, it's best to use the
a0d0e21e 19
a1ce9542
JF
20 use English '-no_match_vars';
21
22invocation if you don't need $PREMATCH, $MATCH, or $POSTMATCH, as it avoids
23a certain performance hit with the use of regular expressions. See
24L<English>.
25
26Variables that depend on the currently selected filehandle may be set by
27calling an appropriate object method on the IO::Handle object, although
28this is less efficient than using the regular built-in variables. (Summary
29lines below for this contain the word HANDLE.) First you must say
a0d0e21e 30
19799a22 31 use IO::Handle;
a0d0e21e
LW
32
33after which you may use either
34
35 method HANDLE EXPR
36
5a964f20 37or more safely,
a0d0e21e
LW
38
39 HANDLE->method(EXPR)
40
14218588 41Each method returns the old value of the IO::Handle attribute.
a1ce9542 42The methods each take an optional EXPR, which, if supplied, specifies the
19799a22 43new value for the IO::Handle attribute in question. If not supplied,
14218588 44most methods do nothing to the current value--except for
a0d0e21e 45autoflush(), which will assume a 1 for you, just to be different.
a1ce9542 46
14218588 47Because loading in the IO::Handle class is an expensive operation, you should
19799a22 48learn how to use the regular built-in variables.
a0d0e21e 49
748a9306
LW
50A few of these variables are considered "read-only". This means that if
51you try to assign to this variable, either directly or indirectly through
52a reference, you'll raise a run-time exception.
a0d0e21e 53
22d0716c
SB
54You should be very careful when modifying the default values of most
55special variables described in this document. In most cases you want
56to localize these variables before changing them, since if you don't,
57the change may affect other modules which rely on the default values
58of the special variables that you have changed. This is one of the
59correct ways to read the whole file at once:
60
61 open my $fh, "foo" or die $!;
62 local $/; # enable localized slurp mode
63 my $content = <$fh>;
64 close $fh;
65
66But the following code is quite bad:
67
68 open my $fh, "foo" or die $!;
69 undef $/; # enable slurp mode
70 my $content = <$fh>;
71 close $fh;
72
73since some other module, may want to read data from some file in the
74default "line mode", so if the code we have just presented has been
75executed, the global value of C<$/> is now changed for any other code
76running inside the same Perl interpreter.
77
78Usually when a variable is localized you want to make sure that this
79change affects the shortest scope possible. So unless you are already
80inside some short C<{}> block, you should create one yourself. For
81example:
82
83 my $content = '';
84 open my $fh, "foo" or die $!;
85 {
86 local $/;
87 $content = <$fh>;
88 }
89 close $fh;
90
91Here is an example of how your own code can go broken:
92
93 for (1..5){
94 nasty_break();
95 print "$_ ";
96 }
97 sub nasty_break {
98 $_ = 5;
99 # do something with $_
100 }
101
102You probably expect this code to print:
103
104 1 2 3 4 5
105
106but instead you get:
107
108 5 5 5 5 5
109
110Why? Because nasty_break() modifies C<$_> without localizing it
111first. The fix is to add local():
112
113 local $_ = 5;
114
115It's easy to notice the problem in such a short example, but in more
116complicated code you are looking for trouble if you don't localize
117changes to the special variables.
118
fb73857a 119The following list is ordered by scalar variables first, then the
87275199 120arrays, then the hashes.
fb73857a 121
a0d0e21e
LW
122=over 8
123
124=item $ARG
125
126=item $_
127
128The default input and pattern-searching space. The following pairs are
129equivalent:
130
19799a22 131 while (<>) {...} # equivalent only in while!
54310121 132 while (defined($_ = <>)) {...}
a0d0e21e
LW
133
134 /^Subject:/
135 $_ =~ /^Subject:/
136
137 tr/a-z/A-Z/
138 $_ =~ tr/a-z/A-Z/
139
19799a22
GS
140 chomp
141 chomp($_)
a0d0e21e 142
54310121 143Here are the places where Perl will assume $_ even if you
cb1a09d0
AD
144don't use it:
145
146=over 3
147
148=item *
149
150Various unary functions, including functions like ord() and int(), as well
151as the all file tests (C<-f>, C<-d>) except for C<-t>, which defaults to
152STDIN.
153
154=item *
155
156Various list functions like print() and unlink().
157
158=item *
159
160The pattern matching operations C<m//>, C<s///>, and C<tr///> when used
161without an C<=~> operator.
162
54310121 163=item *
cb1a09d0
AD
164
165The default iterator variable in a C<foreach> loop if no other
166variable is supplied.
167
54310121 168=item *
cb1a09d0
AD
169
170The implicit iterator variable in the grep() and map() functions.
171
54310121 172=item *
cb1a09d0 173
c47ff5f1 174The default place to put an input record when a C<< <FH> >>
cb1a09d0 175operation's result is tested by itself as the sole criterion of a C<while>
14218588 176test. Outside a C<while> test, this will not happen.
cb1a09d0
AD
177
178=back
179
59f00321
RGS
180As C<$_> is a global variable, this may lead in some cases to unwanted
181side-effects. As of perl 5.9.1, you can now use a lexical version of
182C<$_> by declaring it in a file or in a block with C<my>. Moreover,
183declaring C<our $> restores the global C<$_> in the current scope.
184
a0d0e21e
LW
185(Mnemonic: underline is understood in certain operations.)
186
6e2995f4
PP
187=back
188
189=over 8
190
a1db74c9
JH
191=item $a
192
193=item $b
194
195Special package variables when using sort(), see L<perlfunc/sort>.
196Because of this specialness $a and $b don't need to be declared
f83912f2
J
197(using use vars, or our()) even when using the C<strict 'vars'> pragma.
198Don't lexicalize them with C<my $a> or C<my $b> if you want to be
199able to use them in the sort() comparison block or function.
a1db74c9
JH
200
201=back
202
203=over 8
204
c47ff5f1 205=item $<I<digits>>
a0d0e21e 206
19799a22
GS
207Contains the subpattern from the corresponding set of capturing
208parentheses from the last pattern match, not counting patterns
209matched in nested blocks that have been exited already. (Mnemonic:
210like \digits.) These variables are all read-only and dynamically
211scoped to the current BLOCK.
a0d0e21e
LW
212
213=item $MATCH
214
215=item $&
216
217The string matched by the last successful pattern match (not counting
218any matches hidden within a BLOCK or eval() enclosed by the current
19799a22
GS
219BLOCK). (Mnemonic: like & in some editors.) This variable is read-only
220and dynamically scoped to the current BLOCK.
a0d0e21e 221
19ddd453 222The use of this variable anywhere in a program imposes a considerable
667e1aea 223performance penalty on all regular expression matches. See L</BUGS>.
19ddd453 224
a0d0e21e
LW
225=item $PREMATCH
226
227=item $`
228
229The string preceding whatever was matched by the last successful
230pattern match (not counting any matches hidden within a BLOCK or eval
a8f8344d 231enclosed by the current BLOCK). (Mnemonic: C<`> often precedes a quoted
a0d0e21e
LW
232string.) This variable is read-only.
233
19ddd453 234The use of this variable anywhere in a program imposes a considerable
667e1aea 235performance penalty on all regular expression matches. See L</BUGS>.
19ddd453 236
a0d0e21e
LW
237=item $POSTMATCH
238
239=item $'
240
241The string following whatever was matched by the last successful
242pattern match (not counting any matches hidden within a BLOCK or eval()
a8f8344d 243enclosed by the current BLOCK). (Mnemonic: C<'> often follows a quoted
a0d0e21e
LW
244string.) Example:
245
22d0716c 246 local $_ = 'abcdefghi';
a0d0e21e
LW
247 /def/;
248 print "$`:$&:$'\n"; # prints abc:def:ghi
249
19799a22 250This variable is read-only and dynamically scoped to the current BLOCK.
a0d0e21e 251
19ddd453 252The use of this variable anywhere in a program imposes a considerable
667e1aea 253performance penalty on all regular expression matches. See L</BUGS>.
19ddd453 254
a0d0e21e
LW
255=item $LAST_PAREN_MATCH
256
257=item $+
258
a01268b5
JH
259The text matched by the last bracket of the last successful search pattern.
260This is useful if you don't know which one of a set of alternative patterns
261matched. For example:
a0d0e21e
LW
262
263 /Version: (.*)|Revision: (.*)/ && ($rev = $+);
264
265(Mnemonic: be positive and forward looking.)
19799a22 266This variable is read-only and dynamically scoped to the current BLOCK.
a0d0e21e 267
a01268b5
JH
268=item $^N
269
270The text matched by the used group most-recently closed (i.e. the group
271with the rightmost closing parenthesis) of the last successful search
ad83b128
PN
272pattern. (Mnemonic: the (possibly) Nested parenthesis that most
273recently closed.)
274
210b36aa 275This is primarily used inside C<(?{...})> blocks for examining text
a01268b5
JH
276recently matched. For example, to effectively capture text to a variable
277(in addition to C<$1>, C<$2>, etc.), replace C<(...)> with
278
279 (?:(...)(?{ $var = $^N }))
280
281By setting and then using C<$var> in this way relieves you from having to
282worry about exactly which numbered set of parentheses they are.
283
284This variable is dynamically scoped to the current BLOCK.
285
fe307981
GS
286=item @LAST_MATCH_END
287
6cef1e77
IZ
288=item @+
289
4ba05bdc
GS
290This array holds the offsets of the ends of the last successful
291submatches in the currently active dynamic scope. C<$+[0]> is
292the offset into the string of the end of the entire match. This
293is the same value as what the C<pos> function returns when called
294on the variable that was matched against. The I<n>th element
295of this array holds the offset of the I<n>th submatch, so
296C<$+[1]> is the offset past where $1 ends, C<$+[2]> the offset
297past where $2 ends, and so on. You can use C<$#+> to determine
298how many subgroups were in the last successful match. See the
299examples given for the C<@-> variable.
6cef1e77 300
fcc7d916 301=item HANDLE->input_line_number(EXPR)
a0d0e21e
LW
302
303=item $INPUT_LINE_NUMBER
304
305=item $NR
306
307=item $.
308
fcc7d916
IK
309Current line number for the last filehandle accessed.
310
311Each filehandle in Perl counts the number of lines that have been read
312from it. (Depending on the value of C<$/>, Perl's idea of what
313constitutes a line may not match yours.) When a line is read from a
314filehandle (via readline() or C<< <> >>), or when tell() or seek() is
315called on it, C<$.> becomes an alias to the line counter for that
316filehandle.
317
318You can adjust the counter by assigning to C<$.>, but this will not
319actually move the seek pointer. I<Localizing C<$.> will not localize
320the filehandle's line count>. Instead, it will localize perl's notion
321of which filehandle C<$.> is currently aliased to.
322
323C<$.> is reset when the filehandle is closed, but B<not> when an open
324filehandle is reopened without an intervening close(). For more
e48df184 325details, see L<perlop/"IE<sol>O Operators">. Because C<< <> >> never does
fcc7d916
IK
326an explicit close, line numbers increase across ARGV files (but see
327examples in L<perlfunc/eof>).
328
329You can also use C<< HANDLE->input_line_number(EXPR) >> to access the
330line counter for a given filehandle without having to worry about
331which handle you last accessed.
332
333(Mnemonic: many programs use "." to mean the current line number.)
334
335=item IO::Handle->input_record_separator(EXPR)
a0d0e21e
LW
336
337=item $INPUT_RECORD_SEPARATOR
338
339=item $RS
340
341=item $/
342
14218588
GS
343The input record separator, newline by default. This
344influences Perl's idea of what a "line" is. Works like B<awk>'s RS
19799a22 345variable, including treating empty lines as a terminator if set to
14218588
GS
346the null string. (An empty line cannot contain any spaces
347or tabs.) You may set it to a multi-character string to match a
19799a22
GS
348multi-character terminator, or to C<undef> to read through the end
349of file. Setting it to C<"\n\n"> means something slightly
350different than setting to C<"">, if the file contains consecutive
351empty lines. Setting to C<""> will treat two or more consecutive
352empty lines as a single empty line. Setting to C<"\n\n"> will
353blindly assume that the next input character belongs to the next
14218588 354paragraph, even if it's a newline. (Mnemonic: / delimits
19799a22 355line boundaries when quoting poetry.)
a0d0e21e 356
22d0716c
SB
357 local $/; # enable "slurp" mode
358 local $_ = <FH>; # whole file now here
a0d0e21e
LW
359 s/\n[ \t]+/ /g;
360
19799a22
GS
361Remember: the value of C<$/> is a string, not a regex. B<awk> has to be
362better for something. :-)
68dc0745 363
19799a22
GS
364Setting C<$/> to a reference to an integer, scalar containing an integer, or
365scalar that's convertible to an integer will attempt to read records
5b2b9c68 366instead of lines, with the maximum record size being the referenced
19799a22 367integer. So this:
5b2b9c68 368
22d0716c
SB
369 local $/ = \32768; # or \"32768", or \$var_containing_32768
370 open my $fh, $myfile or die $!;
371 local $_ = <$fh>;
5b2b9c68 372
19799a22
GS
373will read a record of no more than 32768 bytes from FILE. If you're
374not reading from a record-oriented file (or your OS doesn't have
375record-oriented files), then you'll likely get a full chunk of data
376with every read. If a record is larger than the record size you've
acbd132f
JH
377set, you'll get the record back in pieces. Trying to set the record
378size to zero or less will cause reading in the (rest of the) whole file.
5b2b9c68 379
19799a22
GS
380On VMS, record reads are done with the equivalent of C<sysread>,
381so it's best not to mix record and non-record reads on the same
382file. (This is unlikely to be a problem, because any file you'd
83763826 383want to read in record mode is probably unusable in line mode.)
14218588 384Non-VMS systems do normal I/O, so it's safe to mix record and
19799a22 385non-record reads of a file.
5b2b9c68 386
14218588 387See also L<perlport/"Newlines">. Also see C<$.>.
883faa13 388
fcc7d916 389=item HANDLE->autoflush(EXPR)
a0d0e21e
LW
390
391=item $OUTPUT_AUTOFLUSH
392
393=item $|
394
19799a22
GS
395If set to nonzero, forces a flush right away and after every write
396or print on the currently selected output channel. Default is 0
14218588 397(regardless of whether the channel is really buffered by the
19799a22
GS
398system or not; C<$|> tells you only whether you've asked Perl
399explicitly to flush after each write). STDOUT will
400typically be line buffered if output is to the terminal and block
401buffered otherwise. Setting this variable is useful primarily when
402you are outputting to a pipe or socket, such as when you are running
403a Perl program under B<rsh> and want to see the output as it's
404happening. This has no effect on input buffering. See L<perlfunc/getc>
405for that. (Mnemonic: when you want your pipes to be piping hot.)
a0d0e21e 406
46550894 407=item IO::Handle->output_field_separator EXPR
a0d0e21e
LW
408
409=item $OUTPUT_FIELD_SEPARATOR
410
411=item $OFS
412
413=item $,
414
d6584ed8
XN
415The output field separator for the print operator. If defined, this
416value is printed between each of print's arguments. Default is C<undef>.
417(Mnemonic: what is printed when there is a "," in your print statement.)
a0d0e21e 418
46550894 419=item IO::Handle->output_record_separator EXPR
a0d0e21e
LW
420
421=item $OUTPUT_RECORD_SEPARATOR
422
423=item $ORS
424
425=item $\
426
d6584ed8
XN
427The output record separator for the print operator. If defined, this
428value is printed after the last of print's arguments. Default is C<undef>.
429(Mnemonic: you set C<$\> instead of adding "\n" at the end of the print.
430Also, it's just like C<$/>, but it's what you get "back" from Perl.)
a0d0e21e
LW
431
432=item $LIST_SEPARATOR
433
434=item $"
435
19799a22
GS
436This is like C<$,> except that it applies to array and slice values
437interpolated into a double-quoted string (or similar interpreted
438string). Default is a space. (Mnemonic: obvious, I think.)
a0d0e21e
LW
439
440=item $SUBSCRIPT_SEPARATOR
441
442=item $SUBSEP
443
444=item $;
445
54310121 446The subscript separator for multidimensional array emulation. If you
a0d0e21e
LW
447refer to a hash element as
448
449 $foo{$a,$b,$c}
450
451it really means
452
453 $foo{join($;, $a, $b, $c)}
454
455But don't put
456
457 @foo{$a,$b,$c} # a slice--note the @
458
459which means
460
461 ($foo{$a},$foo{$b},$foo{$c})
462
19799a22
GS
463Default is "\034", the same as SUBSEP in B<awk>. If your
464keys contain binary data there might not be any safe value for C<$;>.
a0d0e21e 465(Mnemonic: comma (the syntactic subscript separator) is a
19799a22 466semi-semicolon. Yeah, I know, it's pretty lame, but C<$,> is already
a0d0e21e
LW
467taken for something more important.)
468
19799a22
GS
469Consider using "real" multidimensional arrays as described
470in L<perllol>.
a0d0e21e 471
fcc7d916 472=item HANDLE->format_page_number(EXPR)
a0d0e21e
LW
473
474=item $FORMAT_PAGE_NUMBER
475
476=item $%
477
478The current page number of the currently selected output channel.
19799a22 479Used with formats.
a0d0e21e
LW
480(Mnemonic: % is page number in B<nroff>.)
481
fcc7d916 482=item HANDLE->format_lines_per_page(EXPR)
a0d0e21e
LW
483
484=item $FORMAT_LINES_PER_PAGE
485
486=item $=
487
488The current page length (printable lines) of the currently selected
19799a22
GS
489output channel. Default is 60.
490Used with formats.
491(Mnemonic: = has horizontal lines.)
a0d0e21e 492
fcc7d916 493=item HANDLE->format_lines_left(EXPR)
a0d0e21e
LW
494
495=item $FORMAT_LINES_LEFT
496
497=item $-
498
499The number of lines left on the page of the currently selected output
19799a22
GS
500channel.
501Used with formats.
502(Mnemonic: lines_on_page - lines_printed.)
a0d0e21e 503
fe307981
GS
504=item @LAST_MATCH_START
505
6cef1e77
IZ
506=item @-
507
19799a22 508$-[0] is the offset of the start of the last successful match.
6cef1e77 509C<$-[>I<n>C<]> is the offset of the start of the substring matched by
8f580fb8 510I<n>-th subpattern, or undef if the subpattern did not match.
6cef1e77
IZ
511
512Thus after a match against $_, $& coincides with C<substr $_, $-[0],
5060ef7b
RGS
513$+[0] - $-[0]>. Similarly, $I<n> coincides with C<substr $_, $-[n],
514$+[n] - $-[n]> if C<$-[n]> is defined, and $+ coincides with
515C<substr $_, $-[$#-], $+[$#-] - $-[$#-]>. One can use C<$#-> to find the last
14218588
GS
516matched subgroup in the last successful match. Contrast with
517C<$#+>, the number of subgroups in the regular expression. Compare
19799a22 518with C<@+>.
6cef1e77 519
4ba05bdc
GS
520This array holds the offsets of the beginnings of the last
521successful submatches in the currently active dynamic scope.
522C<$-[0]> is the offset into the string of the beginning of the
523entire match. The I<n>th element of this array holds the offset
0926d669
JP
524of the I<n>th submatch, so C<$-[1]> is the offset where $1
525begins, C<$-[2]> the offset where $2 begins, and so on.
4ba05bdc
GS
526
527After a match against some variable $var:
528
529=over 5
530
4375e838 531=item C<$`> is the same as C<substr($var, 0, $-[0])>
4ba05bdc 532
4375e838 533=item C<$&> is the same as C<substr($var, $-[0], $+[0] - $-[0])>
4ba05bdc 534
4375e838 535=item C<$'> is the same as C<substr($var, $+[0])>
4ba05bdc
GS
536
537=item C<$1> is the same as C<substr($var, $-[1], $+[1] - $-[1])>
538
539=item C<$2> is the same as C<substr($var, $-[2], $+[2] - $-[2])>
540
80dc6883 541=item C<$3> is the same as C<substr($var, $-[3], $+[3] - $-[3])>
4ba05bdc
GS
542
543=back
544
fcc7d916 545=item HANDLE->format_name(EXPR)
a0d0e21e
LW
546
547=item $FORMAT_NAME
548
549=item $~
550
551The name of the current report format for the currently selected output
14218588 552channel. Default is the name of the filehandle. (Mnemonic: brother to
19799a22 553C<$^>.)
a0d0e21e 554
fcc7d916 555=item HANDLE->format_top_name(EXPR)
a0d0e21e
LW
556
557=item $FORMAT_TOP_NAME
558
559=item $^
560
561The name of the current top-of-page format for the currently selected
14218588 562output channel. Default is the name of the filehandle with _TOP
a0d0e21e
LW
563appended. (Mnemonic: points to top of page.)
564
46550894 565=item IO::Handle->format_line_break_characters EXPR
a0d0e21e
LW
566
567=item $FORMAT_LINE_BREAK_CHARACTERS
568
569=item $:
570
571The current set of characters after which a string may be broken to
54310121 572fill continuation fields (starting with ^) in a format. Default is
a0d0e21e
LW
573S<" \n-">, to break on whitespace or hyphens. (Mnemonic: a "colon" in
574poetry is a part of a line.)
575
46550894 576=item IO::Handle->format_formfeed EXPR
a0d0e21e
LW
577
578=item $FORMAT_FORMFEED
579
580=item $^L
581
14218588 582What formats output as a form feed. Default is \f.
a0d0e21e
LW
583
584=item $ACCUMULATOR
585
586=item $^A
587
588The current value of the write() accumulator for format() lines. A format
19799a22 589contains formline() calls that put their result into C<$^A>. After
a0d0e21e 590calling its format, write() prints out the contents of C<$^A> and empties.
14218588 591So you never really see the contents of C<$^A> unless you call
a0d0e21e
LW
592formline() yourself and then look at it. See L<perlform> and
593L<perlfunc/formline()>.
594
595=item $CHILD_ERROR
596
597=item $?
598
54310121 599The status returned by the last pipe close, backtick (C<``>) command,
19799a22
GS
600successful call to wait() or waitpid(), or from the system()
601operator. This is just the 16-bit status word returned by the
e5218da5 602traditional Unix wait() system call (or else is made up to look like it). Thus, the
c47ff5f1 603exit value of the subprocess is really (C<<< $? >> 8 >>>), and
19799a22
GS
604C<$? & 127> gives which signal, if any, the process died from, and
605C<$? & 128> reports whether there was a core dump. (Mnemonic:
606similar to B<sh> and B<ksh>.)
a0d0e21e 607
7b8d334a 608Additionally, if the C<h_errno> variable is supported in C, its value
14218588 609is returned via $? if any C<gethost*()> function fails.
7b8d334a 610
19799a22 611If you have installed a signal handler for C<SIGCHLD>, the
aa689395
PP
612value of C<$?> will usually be wrong outside that handler.
613
a8f8344d
PP
614Inside an C<END> subroutine C<$?> contains the value that is going to be
615given to C<exit()>. You can modify C<$?> in an C<END> subroutine to
19799a22
GS
616change the exit status of your program. For example:
617
618 END {
619 $? = 1 if $? == 255; # die would make it 255
620 }
a8f8344d 621
aa689395 622Under VMS, the pragma C<use vmsish 'status'> makes C<$?> reflect the
ff0cee69 623actual VMS exit status, instead of the default emulation of POSIX
9bc98430 624status; see L<perlvms/$?> for details.
f86702cc 625
55602bd2
IZ
626Also see L<Error Indicators>.
627
e5218da5
GA
628=item ${^CHILD_ERROR_NATIVE}
629
630The native status returned by the last pipe close, backtick (C<``>)
631command, successful call to wait() or waitpid(), or from the system()
632operator. On POSIX-like systems this value can be decoded with the
633WIFEXITED, WEXITSTATUS, WIFSIGNALED, WTERMSIG, WIFSTOPPED, WSTOPSIG
634and WIFCONTINUED functions provided by the L<POSIX> module.
635
636Under VMS this reflects the actual VMS exit status; i.e. it is the same
637as $? when the pragma C<use vmsish 'status'> is in effect.
638
0a378802
JH
639=item ${^ENCODING}
640
740bd165
PN
641The I<object reference> to the Encode object that is used to convert
642the source code to Unicode. Thanks to this variable your perl script
643does not have to be written in UTF-8. Default is I<undef>. The direct
644manipulation of this variable is highly discouraged. See L<encoding>
048c20cb 645for more details.
0a378802 646
a0d0e21e
LW
647=item $OS_ERROR
648
649=item $ERRNO
650
651=item $!
652
19799a22 653If used numerically, yields the current value of the C C<errno>
6ab308ee
JH
654variable, or in other words, if a system or library call fails, it
655sets this variable. This means that the value of C<$!> is meaningful
656only I<immediately> after a B<failure>:
657
658 if (open(FH, $filename)) {
659 # Here $! is meaningless.
660 ...
661 } else {
662 # ONLY here is $! meaningful.
663 ...
664 # Already here $! might be meaningless.
665 }
666 # Since here we might have either success or failure,
667 # here $! is meaningless.
668
669In the above I<meaningless> stands for anything: zero, non-zero,
670C<undef>. A successful system or library call does B<not> set
671the variable to zero.
672
271df126 673If used as a string, yields the corresponding system error string.
19799a22
GS
674You can assign a number to C<$!> to set I<errno> if, for instance,
675you want C<"$!"> to return the string for error I<n>, or you want
676to set the exit value for the die() operator. (Mnemonic: What just
677went bang?)
a0d0e21e 678
55602bd2
IZ
679Also see L<Error Indicators>.
680
4c5cef9b
MJD
681=item %!
682
683Each element of C<%!> has a true value only if C<$!> is set to that
684value. For example, C<$!{ENOENT}> is true if and only if the current
3be065a1
JH
685value of C<$!> is C<ENOENT>; that is, if the most recent error was
686"No such file or directory" (or its moral equivalent: not all operating
687systems give that exact error, and certainly not all languages).
688To check if a particular key is meaningful on your system, use
689C<exists $!{the_key}>; for a list of legal keys, use C<keys %!>.
690See L<Errno> for more information, and also see above for the
691validity of C<$!>.
4c5cef9b 692
5c055ba3
PP
693=item $EXTENDED_OS_ERROR
694
695=item $^E
696
22fae026
TM
697Error information specific to the current operating system. At
698the moment, this differs from C<$!> under only VMS, OS/2, and Win32
699(and for MacPerl). On all other platforms, C<$^E> is always just
700the same as C<$!>.
701
702Under VMS, C<$^E> provides the VMS status value from the last
703system error. This is more specific information about the last
704system error than that provided by C<$!>. This is particularly
d516a115 705important when C<$!> is set to B<EVMSERR>.
22fae026 706
1c1c7f20
GS
707Under OS/2, C<$^E> is set to the error code of the last call to
708OS/2 API either via CRT, or directly from perl.
22fae026
TM
709
710Under Win32, C<$^E> always returns the last error information
711reported by the Win32 call C<GetLastError()> which describes
712the last error from within the Win32 API. Most Win32-specific
19799a22 713code will report errors via C<$^E>. ANSI C and Unix-like calls
22fae026
TM
714set C<errno> and so most portable Perl code will report errors
715via C<$!>.
716
717Caveats mentioned in the description of C<$!> generally apply to
718C<$^E>, also. (Mnemonic: Extra error explanation.)
5c055ba3 719
55602bd2
IZ
720Also see L<Error Indicators>.
721
a0d0e21e
LW
722=item $EVAL_ERROR
723
724=item $@
725
4a280ebe
JG
726The Perl syntax error message from the last eval() operator.
727If $@ is the null string, the last eval() parsed and executed
728correctly (although the operations you invoked may have failed in the
729normal fashion). (Mnemonic: Where was the syntax error "at"?)
a0d0e21e 730
19799a22 731Warning messages are not collected in this variable. You can,
a8f8344d 732however, set up a routine to process warnings by setting C<$SIG{__WARN__}>
54310121 733as described below.
748a9306 734
55602bd2
IZ
735Also see L<Error Indicators>.
736
a0d0e21e
LW
737=item $PROCESS_ID
738
739=item $PID
740
741=item $$
742
19799a22
GS
743The process number of the Perl running this script. You should
744consider this variable read-only, although it will be altered
745across fork() calls. (Mnemonic: same as shells.)
a0d0e21e 746
4d76a344
RGS
747Note for Linux users: on Linux, the C functions C<getpid()> and
748C<getppid()> return different values from different threads. In order to
749be portable, this behavior is not reflected by C<$$>, whose value remains
750consistent across threads. If you want to call the underlying C<getpid()>,
e3256f86 751you may use the CPAN module C<Linux::Pid>.
4d76a344 752
a0d0e21e
LW
753=item $REAL_USER_ID
754
755=item $UID
756
757=item $<
758
19799a22 759The real uid of this process. (Mnemonic: it's the uid you came I<from>,
a043a685 760if you're running setuid.) You can change both the real uid and
a537debe
SP
761the effective uid at the same time by using POSIX::setuid(). Since
762changes to $< require a system call, check $! after a change attempt to
763detect any possible errors.
a0d0e21e
LW
764
765=item $EFFECTIVE_USER_ID
766
767=item $EUID
768
769=item $>
770
771The effective uid of this process. Example:
772
773 $< = $>; # set real to effective uid
774 ($<,$>) = ($>,$<); # swap real and effective uid
775
a043a685 776You can change both the effective uid and the real uid at the same
a537debe
SP
777time by using POSIX::setuid(). Changes to $> require a check to $!
778to detect any possible errors after an attempted change.
a043a685 779
19799a22 780(Mnemonic: it's the uid you went I<to>, if you're running setuid.)
c47ff5f1 781C<< $< >> and C<< $> >> can be swapped only on machines
8cc95fdb 782supporting setreuid().
a0d0e21e
LW
783
784=item $REAL_GROUP_ID
785
786=item $GID
787
788=item $(
789
790The real gid of this process. If you are on a machine that supports
791membership in multiple groups simultaneously, gives a space separated
792list of groups you are in. The first number is the one returned by
793getgid(), and the subsequent ones by getgroups(), one of which may be
8cc95fdb
PP
794the same as the first number.
795
19799a22
GS
796However, a value assigned to C<$(> must be a single number used to
797set the real gid. So the value given by C<$(> should I<not> be assigned
798back to C<$(> without being forced numeric, such as by adding zero.
8cc95fdb 799
a043a685 800You can change both the real gid and the effective gid at the same
a537debe
SP
801time by using POSIX::setgid(). Changes to $( require a check to $!
802to detect any possible errors after an attempted change.
a043a685 803
19799a22
GS
804(Mnemonic: parentheses are used to I<group> things. The real gid is the
805group you I<left>, if you're running setgid.)
a0d0e21e
LW
806
807=item $EFFECTIVE_GROUP_ID
808
809=item $EGID
810
811=item $)
812
813The effective gid of this process. If you are on a machine that
814supports membership in multiple groups simultaneously, gives a space
815separated list of groups you are in. The first number is the one
816returned by getegid(), and the subsequent ones by getgroups(), one of
8cc95fdb
PP
817which may be the same as the first number.
818
19799a22 819Similarly, a value assigned to C<$)> must also be a space-separated
14218588 820list of numbers. The first number sets the effective gid, and
8cc95fdb
PP
821the rest (if any) are passed to setgroups(). To get the effect of an
822empty list for setgroups(), just repeat the new effective gid; that is,
823to force an effective gid of 5 and an effectively empty setgroups()
824list, say C< $) = "5 5" >.
825
a043a685
GW
826You can change both the effective gid and the real gid at the same
827time by using POSIX::setgid() (use only a single numeric argument).
a537debe
SP
828Changes to $) require a check to $! to detect any possible errors
829after an attempted change.
a043a685 830
19799a22
GS
831(Mnemonic: parentheses are used to I<group> things. The effective gid
832is the group that's I<right> for you, if you're running setgid.)
a0d0e21e 833
c47ff5f1 834C<< $< >>, C<< $> >>, C<$(> and C<$)> can be set only on
19799a22
GS
835machines that support the corresponding I<set[re][ug]id()> routine. C<$(>
836and C<$)> can be swapped only on machines supporting setregid().
a0d0e21e
LW
837
838=item $PROGRAM_NAME
839
840=item $0
841
80bca1b4
JH
842Contains the name of the program being executed.
843
844On some (read: not all) operating systems assigning to C<$0> modifies
845the argument area that the C<ps> program sees. On some platforms you
846may have to use special C<ps> options or a different C<ps> to see the
847changes. Modifying the $0 is more useful as a way of indicating the
848current program state than it is for hiding the program you're
849running. (Mnemonic: same as B<sh> and B<ksh>.)
f9cbb277 850
cf525c36 851Note that there are platform specific limitations on the maximum
f9cbb277
JH
852length of C<$0>. In the most extreme case it may be limited to the
853space occupied by the original C<$0>.
a0d0e21e 854
80bca1b4
JH
855In some platforms there may be arbitrary amount of padding, for
856example space characters, after the modified name as shown by C<ps>.
dda345b7 857In some platforms this padding may extend all the way to the original
c80e2480
JH
858length of the argument area, no matter what you do (this is the case
859for example with Linux 2.2).
80bca1b4 860
4bc88a62 861Note for BSD users: setting C<$0> does not completely remove "perl"
6a4647a3
JH
862from the ps(1) output. For example, setting C<$0> to C<"foobar"> may
863result in C<"perl: foobar (perl)"> (whether both the C<"perl: "> prefix
864and the " (perl)" suffix are shown depends on your exact BSD variant
865and version). This is an operating system feature, Perl cannot help it.
4bc88a62 866
e2975953
JH
867In multithreaded scripts Perl coordinates the threads so that any
868thread may modify its copy of the C<$0> and the change becomes visible
cf525c36 869to ps(1) (assuming the operating system plays along). Note that
80bca1b4
JH
870the view of C<$0> the other threads have will not change since they
871have their own copies of it.
e2975953 872
a0d0e21e
LW
873=item $[
874
875The index of the first element in an array, and of the first character
19799a22
GS
876in a substring. Default is 0, but you could theoretically set it
877to 1 to make Perl behave more like B<awk> (or Fortran) when
878subscripting and when evaluating the index() and substr() functions.
879(Mnemonic: [ begins subscripts.)
a0d0e21e 880
19799a22
GS
881As of release 5 of Perl, assignment to C<$[> is treated as a compiler
882directive, and cannot influence the behavior of any other file.
f83ed198 883(That's why you can only assign compile-time constants to it.)
19799a22 884Its use is highly discouraged.
a0d0e21e 885
f83ed198 886Note that, unlike other compile-time directives (such as L<strict>),
af7a0647
RGS
887assignment to C<$[> can be seen from outer lexical scopes in the same file.
888However, you can use local() on it to strictly bind its value to a
f83ed198
RGS
889lexical block.
890
a0d0e21e
LW
891=item $]
892
54310121
PP
893The version + patchlevel / 1000 of the Perl interpreter. This variable
894can be used to determine whether the Perl interpreter executing a
895script is in the right range of versions. (Mnemonic: Is this version
896of perl in the right bracket?) Example:
a0d0e21e
LW
897
898 warn "No checksumming!\n" if $] < 3.019;
899
54310121 900See also the documentation of C<use VERSION> and C<require VERSION>
19799a22 901for a convenient way to fail if the running Perl interpreter is too old.
a0d0e21e 902
0c8d858b
MS
903The floating point representation can sometimes lead to inaccurate
904numeric comparisons. See C<$^V> for a more modern representation of
905the Perl version that allows accurate string comparisons.
16070b82 906
305aace0
NIS
907=item $COMPILING
908
909=item $^C
910
19799a22
GS
911The current value of the flag associated with the B<-c> switch.
912Mainly of use with B<-MO=...> to allow code to alter its behavior
913when being compiled, such as for example to AUTOLOAD at compile
914time rather than normal, deferred loading. See L<perlcc>. Setting
915C<$^C = 1> is similar to calling C<B::minus_c>.
305aace0 916
a0d0e21e
LW
917=item $DEBUGGING
918
919=item $^D
920
921The current value of the debugging flags. (Mnemonic: value of B<-D>
b4ab917c
DM
922switch.) May be read or set. Like its command-line equivalent, you can use
923numeric or symbolic values, eg C<$^D = 10> or C<$^D = "st">.
a0d0e21e 924
a3621e74
YO
925=item ${^RE_DEBUG_FLAGS}
926
927The current value of the regex debugging flags. Set to 0 for no debug output
928even when the re 'debug' module is loaded. See L<re> for details.
929
0111c4fd 930=item ${^RE_TRIE_MAXBUF}
a3621e74
YO
931
932Controls how certain regex optimisations are applied and how much memory they
933utilize. This value by default is 65536 which corresponds to a 512kB temporary
934cache. Set this to a higher value to trade memory for speed when matching
935large alternations. Set it to a lower value if you want the optimisations to
936be as conservative of memory as possible but still occur, and set it to a
937negative value to prevent the optimisation and conserve the most memory.
938Under normal situations this variable should be of no interest to you.
939
a0d0e21e
LW
940=item $SYSTEM_FD_MAX
941
942=item $^F
943
944The maximum system file descriptor, ordinarily 2. System file
945descriptors are passed to exec()ed processes, while higher file
946descriptors are not. Also, during an open(), system file descriptors are
947preserved even if the open() fails. (Ordinary file descriptors are
19799a22 948closed before the open() is attempted.) The close-on-exec
a0d0e21e 949status of a file descriptor will be decided according to the value of
8d2a6795
GS
950C<$^F> when the corresponding file, pipe, or socket was opened, not the
951time of the exec().
a0d0e21e 952
6e2995f4
PP
953=item $^H
954
0462a1ab
GS
955WARNING: This variable is strictly for internal use only. Its availability,
956behavior, and contents are subject to change without notice.
957
958This variable contains compile-time hints for the Perl interpreter. At the
959end of compilation of a BLOCK the value of this variable is restored to the
960value when the interpreter started to compile the BLOCK.
961
962When perl begins to parse any block construct that provides a lexical scope
963(e.g., eval body, required file, subroutine body, loop body, or conditional
964block), the existing value of $^H is saved, but its value is left unchanged.
965When the compilation of the block is completed, it regains the saved value.
966Between the points where its value is saved and restored, code that
967executes within BEGIN blocks is free to change the value of $^H.
968
969This behavior provides the semantic of lexical scoping, and is used in,
970for instance, the C<use strict> pragma.
971
972The contents should be an integer; different bits of it are used for
973different pragmatic flags. Here's an example:
974
975 sub add_100 { $^H |= 0x100 }
976
977 sub foo {
978 BEGIN { add_100() }
979 bar->baz($boon);
980 }
981
982Consider what happens during execution of the BEGIN block. At this point
983the BEGIN block has already been compiled, but the body of foo() is still
984being compiled. The new value of $^H will therefore be visible only while
985the body of foo() is being compiled.
986
987Substitution of the above BEGIN block with:
988
989 BEGIN { require strict; strict->import('vars') }
990
991demonstrates how C<use strict 'vars'> is implemented. Here's a conditional
992version of the same lexical pragma:
993
994 BEGIN { require strict; strict->import('vars') if $condition }
995
996=item %^H
997
0462a1ab 998The %^H hash provides the same scoping semantic as $^H. This makes it
46e5f5f4 999useful for implementation of lexically scoped pragmas. See L<perlpragma>.
6e2995f4 1000
a0d0e21e
LW
1001=item $INPLACE_EDIT
1002
1003=item $^I
1004
1005The current value of the inplace-edit extension. Use C<undef> to disable
1006inplace editing. (Mnemonic: value of B<-i> switch.)
1007
fb73857a
PP
1008=item $^M
1009
19799a22
GS
1010By default, running out of memory is an untrappable, fatal error.
1011However, if suitably built, Perl can use the contents of C<$^M>
1012as an emergency memory pool after die()ing. Suppose that your Perl
0acca065 1013were compiled with C<-DPERL_EMERGENCY_SBRK> and used Perl's malloc.
19799a22 1014Then
fb73857a 1015
19799a22 1016 $^M = 'a' x (1 << 16);
fb73857a 1017
51ee6500 1018would allocate a 64K buffer for use in an emergency. See the
19799a22 1019F<INSTALL> file in the Perl distribution for information on how to
0acca065
RGS
1020add custom C compilation flags when compiling perl. To discourage casual
1021use of this advanced feature, there is no L<English|English> long name for
1022this variable.
fb73857a 1023
5c055ba3 1024=item $OSNAME
6e2995f4 1025
5c055ba3
PP
1026=item $^O
1027
1028The name of the operating system under which this copy of Perl was
1029built, as determined during the configuration process. The value
19799a22
GS
1030is identical to C<$Config{'osname'}>. See also L<Config> and the
1031B<-V> command-line switch documented in L<perlrun>.
5c055ba3 1032
443f6d01 1033In Windows platforms, $^O is not very helpful: since it is always
7f510801
GS
1034C<MSWin32>, it doesn't tell the difference between
103595/98/ME/NT/2000/XP/CE/.NET. Use Win32::GetOSName() or
1036Win32::GetOSVersion() (see L<Win32> and L<perlport>) to distinguish
1037between the variants.
916d64a3 1038
e2e27056
JH
1039=item ${^OPEN}
1040
1041An internal variable used by PerlIO. A string in two parts, separated
fae2c0fb
RGS
1042by a C<\0> byte, the first part describes the input layers, the second
1043part describes the output layers.
e2e27056 1044
a0d0e21e
LW
1045=item $PERLDB
1046
1047=item $^P
1048
19799a22
GS
1049The internal variable for debugging support. The meanings of the
1050various bits are subject to change, but currently indicate:
84902520
TB
1051
1052=over 6
1053
1054=item 0x01
1055
1056Debug subroutine enter/exit.
1057
1058=item 0x02
1059
1060Line-by-line debugging.
1061
1062=item 0x04
1063
1064Switch off optimizations.
1065
1066=item 0x08
1067
1068Preserve more data for future interactive inspections.
1069
1070=item 0x10
1071
1072Keep info about source lines on which a subroutine is defined.
1073
1074=item 0x20
1075
1076Start with single-step on.
1077
83ee9e09
GS
1078=item 0x40
1079
1080Use subroutine address instead of name when reporting.
1081
1082=item 0x80
1083
1084Report C<goto &subroutine> as well.
1085
1086=item 0x100
1087
1088Provide informative "file" names for evals based on the place they were compiled.
1089
1090=item 0x200
1091
1092Provide informative names to anonymous subroutines based on the place they
1093were compiled.
1094
7619c85e
RG
1095=item 0x400
1096
1097Debug assertion subroutines enter/exit.
1098
84902520
TB
1099=back
1100
19799a22
GS
1101Some bits may be relevant at compile-time only, some at
1102run-time only. This is a new mechanism and the details may change.
a0d0e21e 1103
66558a10
GS
1104=item $LAST_REGEXP_CODE_RESULT
1105
b9ac3b5b
GS
1106=item $^R
1107
19799a22
GS
1108The result of evaluation of the last successful C<(?{ code })>
1109regular expression assertion (see L<perlre>). May be written to.
b9ac3b5b 1110
66558a10
GS
1111=item $EXCEPTIONS_BEING_CAUGHT
1112
fb73857a
PP
1113=item $^S
1114
fa05a9fd
IST
1115Current state of the interpreter.
1116
1117 $^S State
1118 --------- -------------------
1119 undef Parsing module/eval
1120 true (1) Executing an eval
1121 false (0) Otherwise
1122
1123The first state may happen in $SIG{__DIE__} and $SIG{__WARN__} handlers.
fb73857a 1124
a0d0e21e
LW
1125=item $BASETIME
1126
1127=item $^T
1128
19799a22 1129The time at which the program began running, in seconds since the
5f05dabc 1130epoch (beginning of 1970). The values returned by the B<-M>, B<-A>,
19799a22 1131and B<-C> filetests are based on this value.
a0d0e21e 1132
7c36658b
MS
1133=item ${^TAINT}
1134
9aa05f58
RGS
1135Reflects if taint mode is on or off. 1 for on (the program was run with
1136B<-T>), 0 for off, -1 when only taint warnings are enabled (i.e. with
18e8c5b0 1137B<-t> or B<-TU>). This variable is read-only.
7c36658b 1138
a05d7ebb
JH
1139=item ${^UNICODE}
1140
ab9e1bb7
JH
1141Reflects certain Unicode settings of Perl. See L<perlrun>
1142documentation for the C<-C> switch for more information about
1143the possible values. This variable is set during Perl startup
1144and is thereafter read-only.
fde18df1 1145
e07ea26a
NC
1146=item ${^UTF8CACHE}
1147
1148This variable controls the state of the internal UTF-8 offset caching code.
16d9fe92
NC
11491 for on (the default), 0 for off, -1 to debug the caching code by checking
1150all its results against linear scans, and panicking on any discrepancy.
e07ea26a 1151
ea8eae40
RGS
1152=item ${^UTF8LOCALE}
1153
1154This variable indicates whether an UTF-8 locale was detected by perl at
1155startup. This information is used by perl when it's in
1156adjust-utf8ness-to-locale mode (as when run with the C<-CL> command-line
1157switch); see L<perlrun> for more info on this.
1158
44dcb63b 1159=item $PERL_VERSION
b459063d 1160
16070b82
GS
1161=item $^V
1162
1163The revision, version, and subversion of the Perl interpreter, represented
da2094fd 1164as a string composed of characters with those ordinals. Thus in Perl v5.6.0
44dcb63b
GS
1165it equals C<chr(5) . chr(6) . chr(0)> and will return true for
1166C<$^V eq v5.6.0>. Note that the characters in this string value can
1167potentially be in Unicode range.
16070b82 1168
7d2b1222
DM
1169This variable first appeared in perl 5.6.0; earlier versions of perl will
1170see an undefined value.
1171
16070b82
GS
1172This can be used to determine whether the Perl interpreter executing a
1173script is in the right range of versions. (Mnemonic: use ^V for Version
44dcb63b 1174Control.) Example:
16070b82 1175
7d2b1222 1176 warn "Hashes not randomized!\n" if !$^V or $^V lt v5.8.1
16070b82 1177
aa2f2a36
AMS
1178To convert C<$^V> into its string representation use sprintf()'s
1179C<"%vd"> conversion:
1180
1181 printf "version is v%vd\n", $^V; # Perl's version
1182
44dcb63b 1183See the documentation of C<use VERSION> and C<require VERSION>
16070b82
GS
1184for a convenient way to fail if the running Perl interpreter is too old.
1185
1186See also C<$]> for an older representation of the Perl version.
1187
a0d0e21e
LW
1188=item $WARNING
1189
1190=item $^W
1191
19799a22
GS
1192The current value of the warning switch, initially true if B<-w>
1193was used, false otherwise, but directly modifiable. (Mnemonic:
4438c4b7
JH
1194related to the B<-w> switch.) See also L<warnings>.
1195
6a818117 1196=item ${^WARNING_BITS}
4438c4b7
JH
1197
1198The current set of warning checks enabled by the C<use warnings> pragma.
1199See the documentation of C<warnings> for more details.
a0d0e21e 1200
2a8c8378
JD
1201=item ${^WIN32_SLOPPY_STAT}
1202
1203If this variable is set to a true value, then stat() on Windows will
1204not try to open the file. This means that the link count cannot be
1205determined and file attributes may be out of date if additional
1206hardlinks to the file exist. On the other hand, not opening the file
1207is considerably faster, especially for files on network drives.
1208
1209This variable could be set in the F<sitecustomize.pl> file to
1210configure the local Perl installation to use "sloppy" stat() by
1211default. See L<perlrun> for more information about site
1212customization.
1213
a0d0e21e
LW
1214=item $EXECUTABLE_NAME
1215
1216=item $^X
1217
e71940de 1218The name used to execute the current copy of Perl, from C's
21c1191d 1219C<argv[0]> or (where supported) F</proc/self/exe>.
38e4f4ae 1220
e71940de
PG
1221Depending on the host operating system, the value of $^X may be
1222a relative or absolute pathname of the perl program file, or may
1223be the string used to invoke perl but not the pathname of the
1224perl program file. Also, most operating systems permit invoking
1225programs that are not in the PATH environment variable, so there
a10d74f3
PG
1226is no guarantee that the value of $^X is in PATH. For VMS, the
1227value may or may not include a version number.
38e4f4ae 1228
e71940de
PG
1229You usually can use the value of $^X to re-invoke an independent
1230copy of the same perl that is currently running, e.g.,
1231
1232 @first_run = `$^X -le "print int rand 100 for 1..100"`;
1233
1234But recall that not all operating systems support forking or
1235capturing of the output of commands, so this complex statement
1236may not be portable.
38e4f4ae 1237
e71940de
PG
1238It is not safe to use the value of $^X as a path name of a file,
1239as some operating systems that have a mandatory suffix on
1240executable files do not require use of the suffix when invoking
1241a command. To convert the value of $^X to a path name, use the
1242following statements:
1243
304dea91 1244 # Build up a set of file names (not command names).
e71940de 1245 use Config;
68fb0eb7
PG
1246 $this_perl = $^X;
1247 if ($^O ne 'VMS')
1248 {$this_perl .= $Config{_exe}
1249 unless $this_perl =~ m/$Config{_exe}$/i;}
e71940de
PG
1250
1251Because many operating systems permit anyone with read access to
1252the Perl program file to make a copy of it, patch the copy, and
1253then execute the copy, the security-conscious Perl programmer
1254should take care to invoke the installed copy of perl, not the
1255copy referenced by $^X. The following statements accomplish
1256this goal, and produce a pathname that can be invoked as a
1257command or referenced as a file.
38e4f4ae
SB
1258
1259 use Config;
68fb0eb7
PG
1260 $secure_perl_path = $Config{perlpath};
1261 if ($^O ne 'VMS')
1262 {$secure_perl_path .= $Config{_exe}
1263 unless $secure_perl_path =~ m/$Config{_exe}$/i;}
a0d0e21e 1264
2d84a16a
DM
1265=item ARGV
1266
1267The special filehandle that iterates over command-line filenames in
1268C<@ARGV>. Usually written as the null filehandle in the angle operator
1269C<< <> >>. Note that currently C<ARGV> only has its magical effect
1270within the C<< <> >> operator; elsewhere it is just a plain filehandle
1271corresponding to the last file opened by C<< <> >>. In particular,
1272passing C<\*ARGV> as a parameter to a function that expects a filehandle
1273may not cause your function to automatically read the contents of all the
1274files in C<@ARGV>.
1275
a0d0e21e
LW
1276=item $ARGV
1277
c47ff5f1 1278contains the name of the current file when reading from <>.
a0d0e21e
LW
1279
1280=item @ARGV
1281
19799a22 1282The array @ARGV contains the command-line arguments intended for
14218588 1283the script. C<$#ARGV> is generally the number of arguments minus
19799a22
GS
1284one, because C<$ARGV[0]> is the first argument, I<not> the program's
1285command name itself. See C<$0> for the command name.
a0d0e21e 1286
5ccee41e
JA
1287=item ARGVOUT
1288
1289The special filehandle that points to the currently open output file
1290when doing edit-in-place processing with B<-i>. Useful when you have
1291to do a lot of inserting and don't want to keep modifying $_. See
1292L<perlrun> for the B<-i> switch.
1293
9b0e6e7a
JP
1294=item @F
1295
1296The array @F contains the fields of each line read in when autosplit
1297mode is turned on. See L<perlrun> for the B<-a> switch. This array
1298is package-specific, and must be declared or given a full package name
1299if not in package main when running under C<strict 'vars'>.
1300
a0d0e21e
LW
1301=item @INC
1302
19799a22
GS
1303The array @INC contains the list of places that the C<do EXPR>,
1304C<require>, or C<use> constructs look for their library files. It
1305initially consists of the arguments to any B<-I> command-line
1306switches, followed by the default Perl library, probably
1307F</usr/local/lib/perl>, followed by ".", to represent the current
e48df184
RGS
1308directory. ("." will not be appended if taint checks are enabled, either by
1309C<-T> or by C<-t>.) If you need to modify this at runtime, you should use
19799a22
GS
1310the C<use lib> pragma to get the machine-dependent library properly
1311loaded also:
a0d0e21e 1312
cb1a09d0
AD
1313 use lib '/mypath/libdir/';
1314 use SomeMod;
303f2f76 1315
d54b56d5
RGS
1316You can also insert hooks into the file inclusion system by putting Perl
1317code directly into @INC. Those hooks may be subroutine references, array
1318references or blessed objects. See L<perlfunc/require> for details.
1319
314d39ce
MG
1320=item @ARG
1321
fb73857a
PP
1322=item @_
1323
1324Within a subroutine the array @_ contains the parameters passed to that
19799a22 1325subroutine. See L<perlsub>.
fb73857a 1326
a0d0e21e
LW
1327=item %INC
1328
19799a22
GS
1329The hash %INC contains entries for each filename included via the
1330C<do>, C<require>, or C<use> operators. The key is the filename
1331you specified (with module names converted to pathnames), and the
14218588 1332value is the location of the file found. The C<require>
87275199 1333operator uses this hash to determine whether a particular file has
19799a22 1334already been included.
a0d0e21e 1335
89ccab8c
RGS
1336If the file was loaded via a hook (e.g. a subroutine reference, see
1337L<perlfunc/require> for a description of these hooks), this hook is
9ae8cd5b
RGS
1338by default inserted into %INC in place of a filename. Note, however,
1339that the hook may have set the %INC entry by itself to provide some more
1340specific info.
44f0be63 1341
b687b08b
TC
1342=item %ENV
1343
1344=item $ENV{expr}
a0d0e21e
LW
1345
1346The hash %ENV contains your current environment. Setting a
19799a22
GS
1347value in C<ENV> changes the environment for any child processes
1348you subsequently fork() off.
a0d0e21e 1349
b687b08b
TC
1350=item %SIG
1351
1352=item $SIG{expr}
a0d0e21e 1353
14218588 1354The hash %SIG contains signal handlers for signals. For example:
a0d0e21e
LW
1355
1356 sub handler { # 1st argument is signal name
fb73857a 1357 my($sig) = @_;
a0d0e21e
LW
1358 print "Caught a SIG$sig--shutting down\n";
1359 close(LOG);
1360 exit(0);
1361 }
1362
fb73857a
PP
1363 $SIG{'INT'} = \&handler;
1364 $SIG{'QUIT'} = \&handler;
a0d0e21e 1365 ...
19799a22 1366 $SIG{'INT'} = 'DEFAULT'; # restore default action
a0d0e21e
LW
1367 $SIG{'QUIT'} = 'IGNORE'; # ignore SIGQUIT
1368
f648820c
GS
1369Using a value of C<'IGNORE'> usually has the effect of ignoring the
1370signal, except for the C<CHLD> signal. See L<perlipc> for more about
1371this special case.
1372
19799a22 1373Here are some other examples:
a0d0e21e 1374
fb73857a 1375 $SIG{"PIPE"} = "Plumber"; # assumes main::Plumber (not recommended)
a0d0e21e 1376 $SIG{"PIPE"} = \&Plumber; # just fine; assume current Plumber
19799a22 1377 $SIG{"PIPE"} = *Plumber; # somewhat esoteric
a0d0e21e
LW
1378 $SIG{"PIPE"} = Plumber(); # oops, what did Plumber() return??
1379
19799a22
GS
1380Be sure not to use a bareword as the name of a signal handler,
1381lest you inadvertently call it.
748a9306 1382
44a8e56a 1383If your system has the sigaction() function then signal handlers are
9ce5b4ad 1384installed using it. This means you get reliable signal handling.
44a8e56a 1385
9ce5b4ad
SG
1386The default delivery policy of signals changed in Perl 5.8.0 from
1387immediate (also known as "unsafe") to deferred, also known as
1388"safe signals". See L<perlipc> for more information.
45c0772f 1389
748a9306 1390Certain internal hooks can be also set using the %SIG hash. The
a8f8344d 1391routine indicated by C<$SIG{__WARN__}> is called when a warning message is
748a9306
LW
1392about to be printed. The warning message is passed as the first
1393argument. The presence of a __WARN__ hook causes the ordinary printing
1394of warnings to STDERR to be suppressed. You can use this to save warnings
1395in a variable, or turn warnings into fatal errors, like this:
1396
1397 local $SIG{__WARN__} = sub { die $_[0] };
1398 eval $proggie;
1399
a8f8344d 1400The routine indicated by C<$SIG{__DIE__}> is called when a fatal exception
748a9306
LW
1401is about to be thrown. The error message is passed as the first
1402argument. When a __DIE__ hook routine returns, the exception
1403processing continues as it would have in the absence of the hook,
cb1a09d0 1404unless the hook routine itself exits via a C<goto>, a loop exit, or a die().
774d564b 1405The C<__DIE__> handler is explicitly disabled during the call, so that you
fb73857a
PP
1406can die from a C<__DIE__> handler. Similarly for C<__WARN__>.
1407
19799a22
GS
1408Due to an implementation glitch, the C<$SIG{__DIE__}> hook is called
1409even inside an eval(). Do not use this to rewrite a pending exception
1410in C<$@>, or as a bizarre substitute for overriding CORE::GLOBAL::die().
1411This strange action at a distance may be fixed in a future release
1412so that C<$SIG{__DIE__}> is only called if your program is about
1413to exit, as was the original intent. Any other use is deprecated.
1414
1415C<__DIE__>/C<__WARN__> handlers are very special in one respect:
1416they may be called to report (probable) errors found by the parser.
1417In such a case the parser may be in inconsistent state, so any
1418attempt to evaluate Perl code from such a handler will probably
1419result in a segfault. This means that warnings or errors that
1420result from parsing Perl should be used with extreme caution, like
1421this:
fb73857a
PP
1422
1423 require Carp if defined $^S;
1424 Carp::confess("Something wrong") if defined &Carp::confess;
1425 die "Something wrong, but could not load Carp to give backtrace...
1426 To see backtrace try starting Perl with -MCarp switch";
1427
1428Here the first line will load Carp I<unless> it is the parser who
1429called the handler. The second line will print backtrace and die if
1430Carp was available. The third line will be executed only if Carp was
1431not available.
1432
19799a22 1433See L<perlfunc/die>, L<perlfunc/warn>, L<perlfunc/eval>, and
4438c4b7 1434L<warnings> for additional information.
68dc0745 1435
a0d0e21e 1436=back
55602bd2
IZ
1437
1438=head2 Error Indicators
1439
19799a22
GS
1440The variables C<$@>, C<$!>, C<$^E>, and C<$?> contain information
1441about different types of error conditions that may appear during
1442execution of a Perl program. The variables are shown ordered by
1443the "distance" between the subsystem which reported the error and
1444the Perl process. They correspond to errors detected by the Perl
1445interpreter, C library, operating system, or an external program,
1446respectively.
55602bd2
IZ
1447
1448To illustrate the differences between these variables, consider the
19799a22 1449following Perl expression, which uses a single-quoted string:
55602bd2 1450
19799a22 1451 eval q{
22d0716c
SB
1452 open my $pipe, "/cdrom/install |" or die $!;
1453 my @res = <$pipe>;
1454 close $pipe or die "bad pipe: $?, $!";
19799a22 1455 };
55602bd2
IZ
1456
1457After execution of this statement all 4 variables may have been set.
1458
19799a22
GS
1459C<$@> is set if the string to be C<eval>-ed did not compile (this
1460may happen if C<open> or C<close> were imported with bad prototypes),
1461or if Perl code executed during evaluation die()d . In these cases
1462the value of $@ is the compile error, or the argument to C<die>
4cb1c523 1463(which will interpolate C<$!> and C<$?>). (See also L<Fatal>,
19799a22
GS
1464though.)
1465
c47ff5f1 1466When the eval() expression above is executed, open(), C<< <PIPE> >>,
19799a22
GS
1467and C<close> are translated to calls in the C run-time library and
1468thence to the operating system kernel. C<$!> is set to the C library's
1469C<errno> if one of these calls fails.
1470
1471Under a few operating systems, C<$^E> may contain a more verbose
1472error indicator, such as in this case, "CDROM tray not closed."
14218588 1473Systems that do not support extended error messages leave C<$^E>
19799a22
GS
1474the same as C<$!>.
1475
1476Finally, C<$?> may be set to non-0 value if the external program
1477F</cdrom/install> fails. The upper eight bits reflect specific
1478error conditions encountered by the program (the program's exit()
1479value). The lower eight bits reflect mode of failure, like signal
1480death and core dump information See wait(2) for details. In
1481contrast to C<$!> and C<$^E>, which are set only if error condition
1482is detected, the variable C<$?> is set on each C<wait> or pipe
1483C<close>, overwriting the old value. This is more like C<$@>, which
1484on every eval() is always set on failure and cleared on success.
2b92dfce 1485
19799a22
GS
1486For more details, see the individual descriptions at C<$@>, C<$!>, C<$^E>,
1487and C<$?>.
2b92dfce
GS
1488
1489=head2 Technical Note on the Syntax of Variable Names
1490
19799a22
GS
1491Variable names in Perl can have several formats. Usually, they
1492must begin with a letter or underscore, in which case they can be
1493arbitrarily long (up to an internal limit of 251 characters) and
1494may contain letters, digits, underscores, or the special sequence
1495C<::> or C<'>. In this case, the part before the last C<::> or
1496C<'> is taken to be a I<package qualifier>; see L<perlmod>.
2b92dfce
GS
1497
1498Perl variable names may also be a sequence of digits or a single
1499punctuation or control character. These names are all reserved for
19799a22
GS
1500special uses by Perl; for example, the all-digits names are used
1501to hold data captured by backreferences after a regular expression
1502match. Perl has a special syntax for the single-control-character
1503names: It understands C<^X> (caret C<X>) to mean the control-C<X>
1504character. For example, the notation C<$^W> (dollar-sign caret
1505C<W>) is the scalar variable whose name is the single character
1506control-C<W>. This is better than typing a literal control-C<W>
1507into your program.
2b92dfce 1508
87275199 1509Finally, new in Perl 5.6, Perl variable names may be alphanumeric
19799a22
GS
1510strings that begin with control characters (or better yet, a caret).
1511These variables must be written in the form C<${^Foo}>; the braces
1512are not optional. C<${^Foo}> denotes the scalar variable whose
1513name is a control-C<F> followed by two C<o>'s. These variables are
1514reserved for future special uses by Perl, except for the ones that
1515begin with C<^_> (control-underscore or caret-underscore). No
1516control-character name that begins with C<^_> will acquire a special
1517meaning in any future version of Perl; such names may therefore be
1518used safely in programs. C<$^_> itself, however, I<is> reserved.
1519
1fcb18de
RGS
1520Perl identifiers that begin with digits, control characters, or
1521punctuation characters are exempt from the effects of the C<package>
1522declaration and are always forced to be in package C<main>; they are
1523also exempt from C<strict 'vars'> errors. A few other names are also
1524exempt in these ways:
2b92dfce
GS
1525
1526 ENV STDIN
1527 INC STDOUT
1528 ARGV STDERR
5b88253b 1529 ARGVOUT _
2b92dfce
GS
1530 SIG
1531
1532In particular, the new special C<${^_XYZ}> variables are always taken
19799a22 1533to be in package C<main>, regardless of any C<package> declarations
747fafda 1534presently in scope.
2b92dfce 1535
19799a22
GS
1536=head1 BUGS
1537
1538Due to an unfortunate accident of Perl's implementation, C<use
1539English> imposes a considerable performance penalty on all regular
1540expression matches in a program, regardless of whether they occur
1541in the scope of C<use English>. For that reason, saying C<use
1542English> in libraries is strongly discouraged. See the
1543Devel::SawAmpersand module documentation from CPAN
1577cd80 1544( http://www.cpan.org/modules/by-module/Devel/ )
19799a22 1545for more information.
2b92dfce 1546
19799a22
GS
1547Having to even think about the C<$^S> variable in your exception
1548handlers is simply wrong. C<$SIG{__DIE__}> as currently implemented
1549invites grievous and difficult to track down errors. Avoid it
1550and use an C<END{}> or CORE::GLOBAL::die override instead.