3 # pragma for controlling the regexp engine
8 our @ISA = qw(Exporter);
9 our @EXPORT_OK = ('regmust',
10 qw(is_regexp regexp_pattern
11 regname regnames regnames_count));
12 our %EXPORT_OK = map { $_ => 1 } @EXPORT_OK;
15 taint => 0x00100000, # HINT_RE_TAINT
16 eval => 0x00200000, # HINT_RE_EVAL
19 my $flags_hint = 0x02000000; # HINT_RE_FLAGS
22 m => 1 << ($PMMOD_SHIFT + 0),
23 s => 1 << ($PMMOD_SHIFT + 1),
24 i => 1 << ($PMMOD_SHIFT + 2),
25 x => 1 << ($PMMOD_SHIFT + 3),
26 xx => 1 << ($PMMOD_SHIFT + 4),
27 n => 1 << ($PMMOD_SHIFT + 5),
28 p => 1 << ($PMMOD_SHIFT + 6),
29 strict => 1 << ($PMMOD_SHIFT + 10),
39 eval { # Ignore errors
42 my $terminal = Tgetent Term::Cap ({OSPEED => 9600}); # Avoid warning.
43 my $props = $ENV{PERL_RE_TC} || 'md,me,so,se,us,ue';
44 my @props = split /,/, $props;
45 my $colors = join "\t", map {$terminal->Tputs($_,1)} @props;
48 $ENV{PERL_RE_COLORS} = $colors;
51 $ENV{PERL_RE_COLORS} ||= qq'\t\t> <\t> <\t\t';
73 OFFSETSDBG => 0x040000,
75 OPTIMISEM => 0x100000,
80 $flags{ALL} = -1 & ~($flags{OFFSETS}|$flags{OFFSETSDBG}|$flags{BUFFERS});
81 $flags{All} = $flags{all} = $flags{DUMP} | $flags{EXECUTE};
82 $flags{Extra} = $flags{EXECUTE} | $flags{COMPILE} | $flags{GPOS};
83 $flags{More} = $flags{MORE} =
84 $flags{All} | $flags{TRIEC} | $flags{TRIEM} | $flags{STATE};
85 $flags{State} = $flags{DUMP} | $flags{EXECUTE} | $flags{STATE};
86 $flags{TRIE} = $flags{DUMP} | $flags{EXECUTE} | $flags{TRIEC};
88 if (defined &DynaLoader::boot_DynaLoader) {
93 # We need to work for miniperl, because the XS toolchain uses Text::Wrap, which
99 # We call install() every time, as if we didn't, we wouldn't
100 # "see" any changes to the color environment var since
101 # the last time it was called.
103 # install() returns an integer, which if casted properly
104 # in C resolves to a structure containing the regexp
105 # hooks. Setting it to a random integer will guarantee
107 $^H{regcomp} = install();
116 my $turning_all_off = ! @_ && ! $on;
117 if ($turning_all_off) {
119 # Pretend were called with certain parameters, which are best dealt
121 push @_, keys %bitmask; # taint and eval
125 # Process each subpragma parameter
127 foreach my $idx (0..$#_){
129 if ($s eq 'Debug' or $s eq 'Debugcolor') {
130 setcolor() if $s =~/color/i;
131 ${^RE_DEBUG_FLAGS} = 0 unless defined ${^RE_DEBUG_FLAGS};
132 for my $idx ($idx+1..$#_) {
133 if ($flags{$_[$idx]}) {
135 ${^RE_DEBUG_FLAGS} |= $flags{$_[$idx]};
137 ${^RE_DEBUG_FLAGS} &= ~ $flags{$_[$idx]};
141 Carp::carp("Unknown \"re\" Debug flag '$_[$idx]', possible flags: ",
142 join(", ",sort keys %flags ) );
145 _load_unload($on ? 1 : ${^RE_DEBUG_FLAGS});
147 } elsif ($s eq 'debug' or $s eq 'debugcolor') {
148 setcolor() if $s =~/color/i;
151 } elsif (exists $bitmask{$s}) {
152 $bits |= $bitmask{$s};
153 } elsif ($EXPORT_OK{$s}) {
155 re->export_to_level(2, 're', $s);
156 } elsif ($s eq 'strict') {
158 $^H{reflags} |= $reflags{$s};
159 warnings::warnif('experimental::re_strict',
160 "\"use re 'strict'\" is experimental");
162 # Turn on warnings if not already done.
163 if (! warnings::enabled('regexp')) {
165 warnings->import('regexp');
170 $^H{reflags} &= ~$reflags{$s} if $^H{reflags};
172 # Turn off warnings if we turned them on.
173 warnings->unimport('regexp') if $^H{re_strict};
181 } elsif ($s =~ s/^\///) {
182 my $reflags = $^H{reflags} || 0;
185 while ($s =~ m/( . )/gx) {
188 # The 'a' may be repeated; hide this from the rest of the
189 # code by counting and getting rid of all of them, then
190 # changing to 'aa' if there is a repeat.
192 my $sav_pos = pos $s;
193 my $a_count = $s =~ s/a//g;
194 pos $s = $sav_pos - 1; # -1 because got rid of the 'a'
198 qq 'The "a" flag may only appear a maximum of twice'
201 elsif ($a_count == 2) {
208 if ($seen_charset ne $_) {
210 qq 'The "$seen_charset" and "$_" flags '
216 qq 'The "$seen_charset" flag may not appear '
221 $^H{reflags_charset} = $reflags{$_};
225 delete $^H{reflags_charset}
226 if defined $^H{reflags_charset}
227 && $^H{reflags_charset} == $reflags{$_};
229 } elsif (exists $reflags{$_}) {
235 qq 'The "x" flag may only appear a maximum of twice'
238 elsif ($x_count == 2) {
239 $_ = 'xx'; # First time through got the /x
244 ? $reflags |= $reflags{$_}
245 : ($reflags &= ~$reflags{$_});
249 qq'Unknown regular expression flag "$_"'
254 ($^H{reflags} = $reflags or defined $^H{reflags_charset})
256 : ($^H &= ~$flags_hint);
259 Carp::carp("Unknown \"re\" subpragma '$s' (known ones are: ",
260 join(', ', map {qq('$_')} 'debug', 'debugcolor', sort keys %bitmask),
265 if ($turning_all_off) {
268 $^H{reflags_charset} = 0;
282 $^H &= ~ bits(0, @_);
291 re - Perl pragma to alter regular expression behaviour
296 ($x) = ($^X =~ /^(.*)$/s); # $x is tainted here
298 $pat = '(?{ $foo = 1 })';
300 /foo${pat}bar/; # won't fail (when not under -T
304 no re 'taint'; # the default
305 ($x) = ($^X =~ /^(.*)$/s); # $x is not tainted here
307 no re 'eval'; # the default
308 /foo${pat}bar/; # disallowed (with or without -T
312 use re 'strict'; # Raise warnings for more conditions
315 "FOO" =~ / foo /; # /ix implied
317 "FOO" =~ /foo/; # just /i implied
319 use re 'debug'; # output debugging info during
320 /^(.*)$/s; # compile and run time
323 use re 'debugcolor'; # same as 'debug', but with colored
327 use re qw(Debug All); # Same as "use re 'debug'", but you
328 # can use "Debug" with things other
330 use re qw(Debug More); # 'All' plus output more details
331 no re qw(Debug ALL); # Turn on (almost) all re debugging
334 use re qw(is_regexp regexp_pattern); # import utility functions
335 my ($pat,$mods)=regexp_pattern(qr/foo/i);
336 if (is_regexp($obj)) {
337 print "Got regexp: ",
338 scalar regexp_pattern($obj); # just as perl would stringify
339 } # it but no hassle with blessed
342 (We use $^X in these examples because it's tainted by default.)
348 When C<use re 'taint'> is in effect, and a tainted string is the target
349 of a regexp, the regexp memories (or values returned by the m// operator
350 in list context) are tainted. This feature is useful when regexp operations
351 on tainted data aren't meant to extract safe substrings, but to perform
352 other transformations.
356 When C<use re 'eval'> is in effect, a regexp is allowed to contain
357 C<(?{ ... })> zero-width assertions and C<(??{ ... })> postponed
358 subexpressions that are derived from variable interpolation, rather than
359 appearing literally within the regexp. That is normally disallowed, since
361 potential security risk. Note that this pragma is ignored when the regular
362 expression is obtained from tainted data, i.e. evaluation is always
363 disallowed with tainted regular expressions. See L<perlre/(?{ code })>
364 and L<perlre/(??{ code })>.
366 For the purpose of this pragma, interpolation of precompiled regular
367 expressions (i.e., the result of C<qr//>) is I<not> considered variable
372 I<is> allowed if $pat is a precompiled regular expression, even
373 if $pat contains C<(?{ ... })> assertions or C<(??{ ... })> subexpressions.
377 Note that this is an experimental feature which may be changed or removed in a
380 When C<use re 'strict'> is in effect, stricter checks are applied than
381 otherwise when compiling regular expressions patterns. These may cause more
382 warnings to be raised than otherwise, and more things to be fatal instead of
383 just warnings. The purpose of this is to find and report at compile time some
384 things, which may be legal, but have a reasonable possibility of not being the
385 programmer's actual intent. This automatically turns on the C<"regexp">
386 warnings category (if not already on) within its scope.
388 As an example of something that is caught under C<"strict'>, but not
389 otherwise, is the pattern
393 The C<"\x"> construct without curly braces should be followed by exactly two
394 hex digits; this one is followed by three. This currently evaluates as
399 that is, the character whose code point value is C<0xAB>, followed by the
400 letter C<C>. But since C<C> is a a hex digit, there is a reasonable chance
405 that is the single character at C<0xABC>. Under C<'strict'> it is an error to
406 not follow C<\x> with exactly two hex digits. When not under C<'strict'> a
407 warning is generated if there is only one hex digit, and no warning is raised
408 if there are more than two.
410 It is expected that what exactly C<'strict'> does will evolve over time as we
411 gain experience with it. This means that programs that compile under it in
412 today's Perl may not compile, or may have more or fewer warnings, in future
413 Perls. There is no backwards compatibility promises with regards to it. Also
414 there are already proposals for an alternate syntax for enabling it. For
415 these reasons, using it will raise a C<experimental::re_strict> class warning,
416 unless that category is turned off.
418 Note that if a pattern compiled within C<'strict'> is recompiled, say by
419 interpolating into another pattern, outside of C<'strict'>, it is not checked
420 again for strictness. This is because if it works under strict it must work
425 When C<use re '/I<flags>'> is specified, the given I<flags> are automatically
426 added to every regular expression till the end of the lexical scope.
427 I<flags> can be any combination of
442 C<no re '/I<flags>'> will turn off the effect of C<use re '/I<flags>'> for the
445 For example, if you want all your regular expressions to have /msxx on by
450 at the top of your code.
452 The character set C</adul> flags cancel each other out. So, in this example,
459 the second C<use re> does an implicit C<no re '/u'>.
463 use re "/xx"; # Doubled-x
465 use re "/x"; # Single x from here on
468 Turning on one of the character set flags with C<use re> takes precedence over the
469 C<locale> pragma and the 'unicode_strings' C<feature>, for regular
470 expressions. Turning off one of these flags when it is active reverts to
471 the behaviour specified by whatever other pragmata are in scope. For
474 use feature "unicode_strings";
475 no re "/u"; # does nothing
477 no re "/l"; # reverts to unicode_strings behaviour
481 When C<use re 'debug'> is in effect, perl emits debugging messages when
482 compiling and using regular expressions. The output is the same as that
483 obtained by running a C<-DDEBUGGING>-enabled perl interpreter with the
484 B<-Dr> switch. It may be quite voluminous depending on the complexity
485 of the match. Using C<debugcolor> instead of C<debug> enables a
486 form of output that can be used to get a colorful display on terminals
487 that understand termcap color sequences. Set C<$ENV{PERL_RE_TC}> to a
488 comma-separated list of C<termcap> properties to use for highlighting
489 strings on/off, pre-point part on/off.
490 See L<perldebug/"Debugging Regular Expressions"> for additional info.
492 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
493 lexically scoped, as the other directives are. However they have both
494 compile-time and run-time effects.
496 See L<perlmodlib/Pragmatic Modules>.
500 Similarly C<use re 'Debug'> produces debugging output, the difference
501 being that it allows the fine tuning of what debugging output will be
502 emitted. Options are divided into three groups, those related to
503 compilation, those related to execution and those related to special
504 purposes. The options are as follows:
508 =item Compile related options
514 Turns on all non-extra compile related debug options.
518 Turns on debug output related to the process of parsing the pattern.
522 Enables output related to the optimisation phase of compilation.
526 Detailed info about trie compilation.
530 Dump the final program out after it is compiled and optimised.
534 Dump the flags associated with the program
538 Print output intended for testing the internals of the compile process
542 =item Execute related options
548 Turns on all non-extra execute related debug options.
552 Turns on debugging of the main matching loop.
556 Extra debugging of how tries execute.
560 Enable debugging of start-point optimisations.
564 =item Extra debugging options
570 Turns on all "extra" debugging options.
574 Enable debugging the capture group storage during match. Warning,
575 this can potentially produce extremely large output.
579 Enable enhanced TRIE debugging. Enhances both TRIEE
584 Enable debugging of states in the engine.
588 Enable debugging of the recursion stack in the engine. Enabling
589 or disabling this option automatically does the same for debugging
590 states as well. This output from this can be quite large.
594 Enable debugging of the \G modifier.
598 Enable enhanced optimisation debugging and start-point optimisations.
599 Probably not useful except when debugging the regexp engine itself.
603 Dump offset information. This can be used to see how regops correlate
604 to the pattern. Output format is
606 NODENUM:POSITION[LENGTH]
608 Where 1 is the position of the first char in the string. Note that position
609 can be 0, or larger than the actual length of the pattern, likewise length
614 Enable debugging of offsets information. This emits copious
615 amounts of trace information and doesn't mesh well with other
618 Almost definitely only useful to people hacking
619 on the offsets part of the debug engine.
624 =item Other useful flags
626 These are useful shortcuts to save on the typing.
632 Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS.
633 (To get every single option without exception, use both ALL and EXTRA, or
634 starting in 5.30 on a C<-DDEBUGGING>-enabled perl interpreter, use
635 the B<-Drv> command-line switches.)
639 Enable DUMP and all execute options. Equivalent to:
647 Enable the options enabled by "All", plus STATE, TRIEC, and TRIEM.
653 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
654 lexically scoped, as are the other directives. However they have both
655 compile-time and run-time effects.
657 =head2 Exportable Functions
659 As of perl 5.9.5 're' debug contains a number of utility functions that
660 may be optionally exported into the caller's namespace. They are listed
665 =item is_regexp($ref)
667 Returns true if the argument is a compiled regular expression as returned
668 by C<qr//>, false if it is not.
670 This function will not be confused by overloading or blessing. In
671 internals terms, this extracts the regexp pointer out of the
672 PERL_MAGIC_qr structure so it cannot be fooled.
674 =item regexp_pattern($ref)
676 If the argument is a compiled regular expression as returned by C<qr//>,
677 then this function returns the pattern.
679 In list context it returns a two element list, the first element
680 containing the pattern and the second containing the modifiers used when
681 the pattern was compiled.
683 my ($pat, $mods) = regexp_pattern($ref);
685 In scalar context it returns the same as perl would when stringifying a raw
686 C<qr//> with the same pattern inside. If the argument is not a compiled
687 reference then this routine returns false but defined in scalar context,
688 and the empty list in list context. Thus the following
690 if (regexp_pattern($ref) eq '(?^i:foo)')
692 will be warning free regardless of what $ref actually is.
694 Like C<is_regexp> this function will not be confused by overloading
695 or blessing of the object.
699 If the argument is a compiled regular expression as returned by C<qr//>,
700 then this function returns what the optimiser considers to be the longest
701 anchored fixed string and longest floating fixed string in the pattern.
703 A I<fixed string> is defined as being a substring that must appear for the
704 pattern to match. An I<anchored fixed string> is a fixed string that must
705 appear at a particular offset from the beginning of the match. A I<floating
706 fixed string> is defined as a fixed string that can appear at any point in
707 a range of positions relative to the start of the match. For example,
709 my $qr = qr/here .* there/x;
710 my ($anchored, $floating) = regmust($qr);
711 print "anchored:'$anchored'\nfloating:'$floating'\n";
718 Because the C<here> is before the C<.*> in the pattern, its position
719 can be determined exactly. That's not true, however, for the C<there>;
720 it could appear at any point after where the anchored string appeared.
721 Perl uses both for its optimisations, preferring the longer, or, if they are
724 B<NOTE:> This may not necessarily be the definitive longest anchored and
725 floating string. This will be what the optimiser of the Perl that you
726 are using thinks is the longest. If you believe that the result is wrong
727 please report it via the L<perlbug> utility.
729 =item regname($name,$all)
731 Returns the contents of a named buffer of the last successful match. If
732 $all is true, then returns an array ref containing one entry per buffer,
733 otherwise returns the first defined buffer.
737 Returns a list of all of the named buffers defined in the last successful
738 match. If $all is true, then it returns all names defined, if not it returns
739 only names which were involved in the match.
741 =item regnames_count()
743 Returns the number of distinct names defined in the pattern used
744 for the last successful match.
746 B<Note:> this result is always the actual number of distinct
747 named buffers defined, it may not actually match that which is
748 returned by C<regnames()> and related routines when those routines
749 have not been called with the $all parameter set.
755 L<perlmodlib/Pragmatic Modules>.