3 perlvar - Perl predefined variables
7 =head2 Predefined Names
9 The following names have special meaning to Perl. Most
10 punctuation names have reasonable mnemonics, or analogues in one of
11 the shells. Nevertheless, if you wish to use long variable names,
16 at the top of your program. This will alias all the short names to the
17 long names in the current package. Some even have medium names,
18 generally borrowed from B<awk>.
20 To go a step further, those variables that depend on the currently
21 selected filehandle may instead (and preferably) be set by calling an
22 object method on the FileHandle object. (Summary lines below for this
23 contain the word HANDLE.) First you must say
27 after which you may use either
35 Each of the methods returns the old value of the FileHandle attribute.
36 The methods each take an optional EXPR, which if supplied specifies the
37 new value for the FileHandle attribute in question. If not supplied,
38 most of the methods do nothing to the current value, except for
39 autoflush(), which will assume a 1 for you, just to be different.
41 A few of these variables are considered "read-only". This means that if
42 you try to assign to this variable, either directly or indirectly through
43 a reference, you'll raise a run-time exception.
45 The following list is ordered by scalar variables first, then the
46 arrays, then the hashes (except $^M was added in the wrong place).
47 This is somewhat obscured by the fact that %ENV and %SIG are listed as
48 $ENV{expr} and $SIG{expr}.
57 The default input and pattern-searching space. The following pairs are
60 while (<>) {...} # equivalent in only while!
61 while (defined($_ = <>)) {...}
72 Here are the places where Perl will assume $_ even if you
79 Various unary functions, including functions like ord() and int(), as well
80 as the all file tests (C<-f>, C<-d>) except for C<-t>, which defaults to
85 Various list functions like print() and unlink().
89 The pattern matching operations C<m//>, C<s///>, and C<tr///> when used
90 without an C<=~> operator.
94 The default iterator variable in a C<foreach> loop if no other
99 The implicit iterator variable in the grep() and map() functions.
103 The default place to put an input record when a C<E<lt>FHE<gt>>
104 operation's result is tested by itself as the sole criterion of a C<while>
105 test. Note that outside of a C<while> test, this will not happen.
109 (Mnemonic: underline is understood in certain operations.)
115 =item $E<lt>I<digits>E<gt>
117 Contains the subpattern from the corresponding set of parentheses in
118 the last pattern matched, not counting patterns matched in nested
119 blocks that have been exited already. (Mnemonic: like \digits.)
120 These variables are all read-only.
126 The string matched by the last successful pattern match (not counting
127 any matches hidden within a BLOCK or eval() enclosed by the current
128 BLOCK). (Mnemonic: like & in some editors.) This variable is read-only.
134 The string preceding whatever was matched by the last successful
135 pattern match (not counting any matches hidden within a BLOCK or eval
136 enclosed by the current BLOCK). (Mnemonic: C<`> often precedes a quoted
137 string.) This variable is read-only.
143 The string following whatever was matched by the last successful
144 pattern match (not counting any matches hidden within a BLOCK or eval()
145 enclosed by the current BLOCK). (Mnemonic: C<'> often follows a quoted
150 print "$`:$&:$'\n"; # prints abc:def:ghi
152 This variable is read-only.
154 =item $LAST_PAREN_MATCH
158 The last bracket matched by the last search pattern. This is useful if
159 you don't know which of a set of alternative patterns matched. For
162 /Version: (.*)|Revision: (.*)/ && ($rev = $+);
164 (Mnemonic: be positive and forward looking.)
165 This variable is read-only.
167 =item $MULTILINE_MATCHING
171 Set to 1 to do multi-line matching within a string, 0 to tell Perl
172 that it can assume that strings contain a single line, for the purpose
173 of optimizing pattern matches. Pattern matches on strings containing
174 multiple newlines can produce confusing results when "C<$*>" is 0. Default
175 is 0. (Mnemonic: * matches multiple things.) Note that this variable
176 influences the interpretation of only "C<^>" and "C<$>". A literal newline can
177 be searched for even when C<$* == 0>.
179 Use of "C<$*>" is deprecated in modern Perls, supplanted by
180 the C</s> and C</m> modifiers on pattern matching.
182 =item input_line_number HANDLE EXPR
184 =item $INPUT_LINE_NUMBER
190 The current input line number for the last file handle from
191 which you read (or performed a C<seek> or C<tell> on). An
192 explicit close on a filehandle resets the line number. Because
193 "C<E<lt>E<gt>>" never does an explicit close, line numbers increase
194 across ARGV files (but see examples under eof()). Localizing C<$.> has
195 the effect of also localizing Perl's notion of "the last read
196 filehandle". (Mnemonic: many programs use "." to mean the current line
199 =item input_record_separator HANDLE EXPR
201 =item $INPUT_RECORD_SEPARATOR
207 The input record separator, newline by default. Works like B<awk>'s RS
208 variable, including treating empty lines as delimiters if set to the
209 null string. (Note: An empty line cannot contain any spaces or tabs.)
210 You may set it to a multi-character string to match a multi-character
211 delimiter, or to C<undef> to read to end of file. Note that setting it
212 to C<"\n\n"> means something slightly different than setting it to
213 C<"">, if the file contains consecutive empty lines. Setting it to
214 C<""> will treat two or more consecutive empty lines as a single empty
215 line. Setting it to C<"\n\n"> will blindly assume that the next input
216 character belongs to the next paragraph, even if it's a newline.
217 (Mnemonic: / is used to delimit line boundaries when quoting poetry.)
220 $_ = <FH>; # whole file now here
223 Remember: the value of $/ is a string, not a regexp. AWK has to be
224 better for something :-)
226 Setting $/ to a reference to an integer, scalar containing an integer, or
227 scalar that's convertable to an integer will attempt to read records
228 instead of lines, with the maximum record size being the referenced
231 $/ = \32768; # or \"32768", or \$var_containing_32768
235 will read a record of no more than 32768 bytes from FILE. If you're not
236 reading from a record-oriented file (or your OS doesn't have
237 record-oriented files), then you'll likely get a full chunk of data with
238 every read. If a record is larger than the record size you've set, you'll
239 get the record back in pieces.
241 On VMS, record reads are done with the equivalent of C<sysread>, so it's
242 best not to mix record and non-record reads on the same file. (This is
243 likely not a problem, as any file you'd want to read in record mode is
244 proably usable in line mode) Non-VMS systems perform normal I/O, so
245 it's safe to mix record and non-record reads of a file.
247 =item autoflush HANDLE EXPR
249 =item $OUTPUT_AUTOFLUSH
253 If set to nonzero, forces a flush right away and after every write or print on the
254 currently selected output channel. Default is 0 (regardless of whether
255 the channel is actually buffered by the system or not; C<$|> tells you
256 only whether you've asked Perl explicitly to flush after each write).
257 Note that STDOUT will typically be line buffered if output is to the
258 terminal and block buffered otherwise. Setting this variable is useful
259 primarily when you are outputting to a pipe, such as when you are running
260 a Perl script under rsh and want to see the output as it's happening. This
261 has no effect on input buffering.
262 (Mnemonic: when you want your pipes to be piping hot.)
264 =item output_field_separator HANDLE EXPR
266 =item $OUTPUT_FIELD_SEPARATOR
272 The output field separator for the print operator. Ordinarily the
273 print operator simply prints out the comma-separated fields you
274 specify. To get behavior more like B<awk>, set this variable
275 as you would set B<awk>'s OFS variable to specify what is printed
276 between fields. (Mnemonic: what is printed when there is a , in your
279 =item output_record_separator HANDLE EXPR
281 =item $OUTPUT_RECORD_SEPARATOR
287 The output record separator for the print operator. Ordinarily the
288 print operator simply prints out the comma-separated fields you
289 specify, with no trailing newline or record separator assumed.
290 To get behavior more like B<awk>, set this variable as you would
291 set B<awk>'s ORS variable to specify what is printed at the end of the
292 print. (Mnemonic: you set "C<$\>" instead of adding \n at the end of the
293 print. Also, it's just like C<$/>, but it's what you get "back" from
296 =item $LIST_SEPARATOR
300 This is like "C<$,>" except that it applies to array values interpolated
301 into a double-quoted string (or similar interpreted string). Default
302 is a space. (Mnemonic: obvious, I think.)
304 =item $SUBSCRIPT_SEPARATOR
310 The subscript separator for multidimensional array emulation. If you
311 refer to a hash element as
317 $foo{join($;, $a, $b, $c)}
321 @foo{$a,$b,$c} # a slice--note the @
325 ($foo{$a},$foo{$b},$foo{$c})
327 Default is "\034", the same as SUBSEP in B<awk>. Note that if your
328 keys contain binary data there might not be any safe value for "C<$;>".
329 (Mnemonic: comma (the syntactic subscript separator) is a
330 semi-semicolon. Yeah, I know, it's pretty lame, but "C<$,>" is already
331 taken for something more important.)
333 Consider using "real" multidimensional arrays.
339 The output format for printed numbers. This variable is a half-hearted
340 attempt to emulate B<awk>'s OFMT variable. There are times, however,
341 when B<awk> and Perl have differing notions of what is in fact
342 numeric. The initial value is %.I<n>g, where I<n> is the value
343 of the macro DBL_DIG from your system's F<float.h>. This is different from
344 B<awk>'s default OFMT setting of %.6g, so you need to set "C<$#>"
345 explicitly to get B<awk>'s value. (Mnemonic: # is the number sign.)
347 Use of "C<$#>" is deprecated.
349 =item format_page_number HANDLE EXPR
351 =item $FORMAT_PAGE_NUMBER
355 The current page number of the currently selected output channel.
356 (Mnemonic: % is page number in B<nroff>.)
358 =item format_lines_per_page HANDLE EXPR
360 =item $FORMAT_LINES_PER_PAGE
364 The current page length (printable lines) of the currently selected
365 output channel. Default is 60. (Mnemonic: = has horizontal lines.)
367 =item format_lines_left HANDLE EXPR
369 =item $FORMAT_LINES_LEFT
373 The number of lines left on the page of the currently selected output
374 channel. (Mnemonic: lines_on_page - lines_printed.)
376 =item format_name HANDLE EXPR
382 The name of the current report format for the currently selected output
383 channel. Default is name of the filehandle. (Mnemonic: brother to
386 =item format_top_name HANDLE EXPR
388 =item $FORMAT_TOP_NAME
392 The name of the current top-of-page format for the currently selected
393 output channel. Default is name of the filehandle with _TOP
394 appended. (Mnemonic: points to top of page.)
396 =item format_line_break_characters HANDLE EXPR
398 =item $FORMAT_LINE_BREAK_CHARACTERS
402 The current set of characters after which a string may be broken to
403 fill continuation fields (starting with ^) in a format. Default is
404 S<" \n-">, to break on whitespace or hyphens. (Mnemonic: a "colon" in
405 poetry is a part of a line.)
407 =item format_formfeed HANDLE EXPR
409 =item $FORMAT_FORMFEED
413 What formats output to perform a form feed. Default is \f.
419 The current value of the write() accumulator for format() lines. A format
420 contains formline() commands that put their result into C<$^A>. After
421 calling its format, write() prints out the contents of C<$^A> and empties.
422 So you never actually see the contents of C<$^A> unless you call
423 formline() yourself and then look at it. See L<perlform> and
424 L<perlfunc/formline()>.
430 The status returned by the last pipe close, backtick (C<``>) command,
431 or system() operator. Note that this is the status word returned by the
432 wait() system call (or else is made up to look like it). Thus, the exit
433 value of the subprocess is actually (C<$? E<gt>E<gt> 8>), and C<$? & 127>
434 gives which signal, if any, the process died from, and C<$? & 128> reports
435 whether there was a core dump. (Mnemonic: similar to B<sh> and B<ksh>.)
437 Additionally, if the C<h_errno> variable is supported in C, its value
438 is returned via $? if any of the C<gethost*()> functions fail.
440 Note that if you have installed a signal handler for C<SIGCHLD>, the
441 value of C<$?> will usually be wrong outside that handler.
443 Inside an C<END> subroutine C<$?> contains the value that is going to be
444 given to C<exit()>. You can modify C<$?> in an C<END> subroutine to
445 change the exit status of the script.
447 Under VMS, the pragma C<use vmsish 'status'> makes C<$?> reflect the
448 actual VMS exit status, instead of the default emulation of POSIX
451 Also see L<Error Indicators>.
459 If used in a numeric context, yields the current value of errno, with
460 all the usual caveats. (This means that you shouldn't depend on the
461 value of C<$!> to be anything in particular unless you've gotten a
462 specific error return indicating a system error.) If used in a string
463 context, yields the corresponding system error string. You can assign
464 to C<$!> to set I<errno> if, for instance, you want C<"$!"> to return the
465 string for error I<n>, or you want to set the exit value for the die()
466 operator. (Mnemonic: What just went bang?)
468 Also see L<Error Indicators>.
470 =item $EXTENDED_OS_ERROR
474 Error information specific to the current operating system. At
475 the moment, this differs from C<$!> under only VMS, OS/2, and Win32
476 (and for MacPerl). On all other platforms, C<$^E> is always just
479 Under VMS, C<$^E> provides the VMS status value from the last
480 system error. This is more specific information about the last
481 system error than that provided by C<$!>. This is particularly
482 important when C<$!> is set to B<EVMSERR>.
484 Under OS/2, C<$^E> is set to the error code of the last call to
485 OS/2 API either via CRT, or directly from perl.
487 Under Win32, C<$^E> always returns the last error information
488 reported by the Win32 call C<GetLastError()> which describes
489 the last error from within the Win32 API. Most Win32-specific
490 code will report errors via C<$^E>. ANSI C and UNIX-like calls
491 set C<errno> and so most portable Perl code will report errors
494 Caveats mentioned in the description of C<$!> generally apply to
495 C<$^E>, also. (Mnemonic: Extra error explanation.)
497 Also see L<Error Indicators>.
503 The Perl syntax error message from the last eval() command. If null, the
504 last eval() parsed and executed correctly (although the operations you
505 invoked may have failed in the normal fashion). (Mnemonic: Where was
506 the syntax error "at"?)
508 Note that warning messages are not collected in this variable. You can,
509 however, set up a routine to process warnings by setting C<$SIG{__WARN__}>
512 Also see L<Error Indicators>.
520 The process number of the Perl running this script. (Mnemonic: same
529 The real uid of this process. (Mnemonic: it's the uid you came I<FROM>,
530 if you're running setuid.)
532 =item $EFFECTIVE_USER_ID
538 The effective uid of this process. Example:
540 $< = $>; # set real to effective uid
541 ($<,$>) = ($>,$<); # swap real and effective uid
543 (Mnemonic: it's the uid you went I<TO>, if you're running setuid.)
544 Note: "C<$E<lt>>" and "C<$E<gt>>" can be swapped only on machines
545 supporting setreuid().
553 The real gid of this process. If you are on a machine that supports
554 membership in multiple groups simultaneously, gives a space separated
555 list of groups you are in. The first number is the one returned by
556 getgid(), and the subsequent ones by getgroups(), one of which may be
557 the same as the first number.
559 However, a value assigned to "C<$(>" must be a single number used to
560 set the real gid. So the value given by "C<$(>" should I<not> be assigned
561 back to "C<$(>" without being forced numeric, such as by adding zero.
563 (Mnemonic: parentheses are used to I<GROUP> things. The real gid is the
564 group you I<LEFT>, if you're running setgid.)
566 =item $EFFECTIVE_GROUP_ID
572 The effective gid of this process. If you are on a machine that
573 supports membership in multiple groups simultaneously, gives a space
574 separated list of groups you are in. The first number is the one
575 returned by getegid(), and the subsequent ones by getgroups(), one of
576 which may be the same as the first number.
578 Similarly, a value assigned to "C<$)>" must also be a space-separated
579 list of numbers. The first number is used to set the effective gid, and
580 the rest (if any) are passed to setgroups(). To get the effect of an
581 empty list for setgroups(), just repeat the new effective gid; that is,
582 to force an effective gid of 5 and an effectively empty setgroups()
583 list, say C< $) = "5 5" >.
585 (Mnemonic: parentheses are used to I<GROUP> things. The effective gid
586 is the group that's I<RIGHT> for you, if you're running setgid.)
588 Note: "C<$E<lt>>", "C<$E<gt>>", "C<$(>" and "C<$)>" can be set only on
589 machines that support the corresponding I<set[re][ug]id()> routine. "C<$(>"
590 and "C<$)>" can be swapped only on machines supporting setregid().
596 Contains the name of the file containing the Perl script being
597 executed. On some operating systems
598 assigning to "C<$0>" modifies the argument area that the ps(1)
599 program sees. This is more useful as a way of indicating the
600 current program state than it is for hiding the program you're running.
601 (Mnemonic: same as B<sh> and B<ksh>.)
605 The index of the first element in an array, and of the first character
606 in a substring. Default is 0, but you could set it to 1 to make
607 Perl behave more like B<awk> (or Fortran) when subscripting and when
608 evaluating the index() and substr() functions. (Mnemonic: [ begins
611 As of Perl 5, assignment to "C<$[>" is treated as a compiler directive,
612 and cannot influence the behavior of any other file. Its use is
619 The version + patchlevel / 1000 of the Perl interpreter. This variable
620 can be used to determine whether the Perl interpreter executing a
621 script is in the right range of versions. (Mnemonic: Is this version
622 of perl in the right bracket?) Example:
624 warn "No checksumming!\n" if $] < 3.019;
626 See also the documentation of C<use VERSION> and C<require VERSION>
627 for a convenient way to fail if the Perl interpreter is too old.
633 The current value of the debugging flags. (Mnemonic: value of B<-D>
640 The maximum system file descriptor, ordinarily 2. System file
641 descriptors are passed to exec()ed processes, while higher file
642 descriptors are not. Also, during an open(), system file descriptors are
643 preserved even if the open() fails. (Ordinary file descriptors are
644 closed before the open() is attempted.) Note that the close-on-exec
645 status of a file descriptor will be decided according to the value of
646 C<$^F> at the time of the open, not the time of the exec.
650 The current set of syntax checks enabled by C<use strict> and other block
651 scoped compiler hints. See the documentation of C<strict> for more details.
657 The current value of the inplace-edit extension. Use C<undef> to disable
658 inplace editing. (Mnemonic: value of B<-i> switch.)
662 By default, running out of memory it is not trappable. However, if
663 compiled for this, Perl may use the contents of C<$^M> as an emergency
664 pool after die()ing with this message. Suppose that your Perl were
665 compiled with -DPERL_EMERGENCY_SBRK and used Perl's malloc. Then
669 would allocate a 64K buffer for use when in emergency. See the F<INSTALL>
670 file for information on how to enable this option. As a disincentive to
671 casual use of this advanced feature, there is no L<English> long name for
678 The name of the operating system under which this copy of Perl was
679 built, as determined during the configuration process. The value
680 is identical to C<$Config{'osname'}>.
686 The internal variable for debugging support. Different bits mean the
687 following (subject to change):
693 Debug subroutine enter/exit.
697 Line-by-line debugging.
701 Switch off optimizations.
705 Preserve more data for future interactive inspections.
709 Keep info about source lines on which a subroutine is defined.
713 Start with single-step on.
717 Note that some bits may be relevent at compile-time only, some at
718 run-time only. This is a new mechanism and the details may change.
722 The result of evaluation of the last successful L<perlre/C<(?{ code })>>
723 regular expression assertion. (Excluding those used as switches.) May
728 Current state of the interpreter. Undefined if parsing of the current
729 module/eval is not finished (may happen in $SIG{__DIE__} and
730 $SIG{__WARN__} handlers). True if inside an eval, otherwise false.
736 The time at which the script began running, in seconds since the
737 epoch (beginning of 1970). The values returned by the B<-M>, B<-A>,
738 and B<-C> filetests are
745 The current value of the warning switch, either TRUE or FALSE.
746 (Mnemonic: related to the B<-w> switch.)
748 =item $EXECUTABLE_NAME
752 The name that the Perl binary itself was executed as, from C's C<argv[0]>.
756 contains the name of the current file when reading from E<lt>E<gt>.
760 The array @ARGV contains the command line arguments intended for the
761 script. Note that C<$#ARGV> is the generally number of arguments minus
762 one, because C<$ARGV[0]> is the first argument, I<NOT> the command name. See
763 "C<$0>" for the command name.
767 The array @INC contains the list of places to look for Perl scripts to
768 be evaluated by the C<do EXPR>, C<require>, or C<use> constructs. It
769 initially consists of the arguments to any B<-I> command line switches,
770 followed by the default Perl library, probably F</usr/local/lib/perl>,
771 followed by ".", to represent the current directory. If you need to
772 modify this at runtime, you should use the C<use lib> pragma
773 to get the machine-dependent library properly loaded also:
775 use lib '/mypath/libdir/';
780 Within a subroutine the array @_ contains the parameters passed to that
781 subroutine. See L<perlsub>.
785 The hash %INC contains entries for each filename that has
786 been included via C<do> or C<require>. The key is the filename you
787 specified, and the value is the location of the file actually found.
788 The C<require> command uses this array to determine whether a given file
789 has already been included.
791 =item %ENV $ENV{expr}
793 The hash %ENV contains your current environment. Setting a
794 value in C<ENV> changes the environment for child processes.
796 =item %SIG $SIG{expr}
798 The hash %SIG is used to set signal handlers for various
801 sub handler { # 1st argument is signal name
803 print "Caught a SIG$sig--shutting down\n";
808 $SIG{'INT'} = \&handler;
809 $SIG{'QUIT'} = \&handler;
811 $SIG{'INT'} = 'DEFAULT'; # restore default action
812 $SIG{'QUIT'} = 'IGNORE'; # ignore SIGQUIT
814 The %SIG array contains values for only the signals actually set within
815 the Perl script. Here are some other examples:
817 $SIG{"PIPE"} = Plumber; # SCARY!!
818 $SIG{"PIPE"} = "Plumber"; # assumes main::Plumber (not recommended)
819 $SIG{"PIPE"} = \&Plumber; # just fine; assume current Plumber
820 $SIG{"PIPE"} = Plumber(); # oops, what did Plumber() return??
822 The one marked scary is problematic because it's a bareword, which means
823 sometimes it's a string representing the function, and sometimes it's
824 going to call the subroutine call right then and there! Best to be sure
825 and quote it or take a reference to it. *Plumber works too. See L<perlsub>.
827 If your system has the sigaction() function then signal handlers are
828 installed using it. This means you get reliable signal handling. If
829 your system has the SA_RESTART flag it is used when signals handlers are
830 installed. This means that system calls for which it is supported
831 continue rather than returning when a signal arrives. If you want your
832 system calls to be interrupted by signal delivery then do something like
835 use POSIX ':signal_h';
838 sigaction SIGALRM, new POSIX::SigAction sub { $alarm = 1 }
839 or die "Error setting SIGALRM handler: $!\n";
843 Certain internal hooks can be also set using the %SIG hash. The
844 routine indicated by C<$SIG{__WARN__}> is called when a warning message is
845 about to be printed. The warning message is passed as the first
846 argument. The presence of a __WARN__ hook causes the ordinary printing
847 of warnings to STDERR to be suppressed. You can use this to save warnings
848 in a variable, or turn warnings into fatal errors, like this:
850 local $SIG{__WARN__} = sub { die $_[0] };
853 The routine indicated by C<$SIG{__DIE__}> is called when a fatal exception
854 is about to be thrown. The error message is passed as the first
855 argument. When a __DIE__ hook routine returns, the exception
856 processing continues as it would have in the absence of the hook,
857 unless the hook routine itself exits via a C<goto>, a loop exit, or a die().
858 The C<__DIE__> handler is explicitly disabled during the call, so that you
859 can die from a C<__DIE__> handler. Similarly for C<__WARN__>.
861 Note that the C<$SIG{__DIE__}> hook is called even inside eval()ed
862 blocks/strings. See L<perlfunc/die> and L<perlvar/$^S> for how to
865 Note that C<__DIE__>/C<__WARN__> handlers are very special in one
866 respect: they may be called to report (probable) errors found by the
867 parser. In such a case the parser may be in inconsistent state, so
868 any attempt to evaluate Perl code from such a handler will probably
869 result in a segfault. This means that calls which result/may-result
870 in parsing Perl should be used with extreme causion, like this:
872 require Carp if defined $^S;
873 Carp::confess("Something wrong") if defined &Carp::confess;
874 die "Something wrong, but could not load Carp to give backtrace...
875 To see backtrace try starting Perl with -MCarp switch";
877 Here the first line will load Carp I<unless> it is the parser who
878 called the handler. The second line will print backtrace and die if
879 Carp was available. The third line will be executed only if Carp was
882 See L<perlfunc/die>, L<perlfunc/warn> and L<perlfunc/eval> for
887 =head2 Error Indicators
889 The variables L<$@>, L<$!>, L<$^E>, and L<$?> contain information about
890 different types of error conditions that may appear during execution of
891 Perl script. The variables are shown ordered by the "distance" between
892 the subsystem which reported the error and the Perl process, and
893 correspond to errors detected by the Perl interpreter, C library,
894 operating system, or an external program, respectively.
896 To illustrate the differences between these variables, consider the
897 following Perl expression:
900 open PIPE, "/cdrom/install |";
902 close PIPE or die "bad pipe: $?, $!";
905 After execution of this statement all 4 variables may have been set.
907 $@ is set if the string to be C<eval>-ed did not compile (this may happen if
908 C<open> or C<close> were imported with bad prototypes), or if Perl
909 code executed during evaluation die()d (either implicitly, say,
910 if C<open> was imported from module L<Fatal>, or the C<die> after
911 C<close> was triggered). In these cases the value of $@ is the compile
912 error, or C<Fatal> error (which will interpolate C<$!>!), or the argument
913 to C<die> (which will interpolate C<$!> and C<$?>!).
915 When the above expression is executed, open(), C<<PIPEE<gt>>, and C<close>
916 are translated to C run-time library calls. $! is set if one of these
917 calls fails. The value is a symbolic indicator chosen by the C run-time
918 library, say C<No such file or directory>.
920 On some systems the above C library calls are further translated
921 to calls to the kernel. The kernel may have set more verbose error
922 indicator that one of the handful of standard C errors. In such cases $^E
923 contains this verbose error indicator, which may be, say, C<CDROM tray not
924 closed>. On systems where C library calls are identical to system calls
925 $^E is a duplicate of $!.
927 Finally, $? may be set to non-C<0> value if the external program
928 C</cdrom/install> fails. Upper bits of the particular value may reflect
929 specific error conditions encountered by this program (this is
930 program-dependent), lower-bits reflect mode of failure (segfault, completion,
931 etc.). Note that in contrast to $@, $!, and $^E, which are set only
932 if error condition is detected, the variable $? is set on each C<wait> or
933 pipe C<close>, overwriting the old value.
935 For more details, see the individual descriptions at L<$@>, L<$!>, L<$^E>,