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 << ($PMMOD_SHIFT + 3),
26 p => 1 << ($PMMOD_SHIFT + 4),
28 l => 1 << ($PMMOD_SHIFT + 5),
29 u => 1 << ($PMMOD_SHIFT + 6),
34 eval { # Ignore errors
37 my $terminal = Tgetent Term::Cap ({OSPEED => 9600}); # Avoid warning.
38 my $props = $ENV{PERL_RE_TC} || 'md,me,so,se,us,ue';
39 my @props = split /,/, $props;
40 my $colors = join "\t", map {$terminal->Tputs($_,1)} @props;
43 $ENV{PERL_RE_COLORS} = $colors;
46 $ENV{PERL_RE_COLORS} ||= qq'\t\t> <\t> <\t\t';
67 OFFSETSDBG => 0x040000,
69 OPTIMISEM => 0x100000,
74 $flags{ALL} = -1 & ~($flags{OFFSETS}|$flags{OFFSETSDBG}|$flags{BUFFERS});
75 $flags{All} = $flags{all} = $flags{DUMP} | $flags{EXECUTE};
76 $flags{Extra} = $flags{EXECUTE} | $flags{COMPILE} | $flags{GPOS};
77 $flags{More} = $flags{MORE} = $flags{All} | $flags{TRIEC} | $flags{TRIEM} | $flags{STATE};
78 $flags{State} = $flags{DUMP} | $flags{EXECUTE} | $flags{STATE};
79 $flags{TRIE} = $flags{DUMP} | $flags{EXECUTE} | $flags{TRIEC};
81 if (defined &DynaLoader::boot_DynaLoader) {
86 # We need to work for miniperl, because the XS toolchain uses Text::Wrap, which
92 # We call install() every time, as if we didn't, we wouldn't
93 # "see" any changes to the color environment var since
94 # the last time it was called.
96 # install() returns an integer, which if casted properly
97 # in C resolves to a structure containing the regexp
98 # hooks. Setting it to a random integer will guarantee
100 $^H{regcomp} = install();
111 Carp::carp("Useless use of \"re\" pragma");
114 foreach my $idx (0..$#_){
116 if ($s eq 'Debug' or $s eq 'Debugcolor') {
117 setcolor() if $s =~/color/i;
118 ${^RE_DEBUG_FLAGS} = 0 unless defined ${^RE_DEBUG_FLAGS};
119 for my $idx ($idx+1..$#_) {
120 if ($flags{$_[$idx]}) {
122 ${^RE_DEBUG_FLAGS} |= $flags{$_[$idx]};
124 ${^RE_DEBUG_FLAGS} &= ~ $flags{$_[$idx]};
128 Carp::carp("Unknown \"re\" Debug flag '$_[$idx]', possible flags: ",
129 join(", ",sort keys %flags ) );
132 _load_unload($on ? 1 : ${^RE_DEBUG_FLAGS});
134 } elsif ($s eq 'debug' or $s eq 'debugcolor') {
135 setcolor() if $s =~/color/i;
138 } elsif (exists $bitmask{$s}) {
139 $bits |= $bitmask{$s};
140 } elsif ($EXPORT_OK{$s}) {
142 re->export_to_level(2, 're', $s);
143 } elsif ($s =~ s/^\///) {
144 my $reflags = $^H{reflags} || 0;
149 if ($seen_dul && $seen_dul ne $_) {
152 qq 'The "$seen_dul" and "$_" flags '
156 $^H{reflags_dul} = $reflags{$_};
160 delete $^H{reflags_dul}
161 if defined $^H{reflags_dul}
162 && $^H{reflags_dul} == $reflags{$_};
164 } elsif (exists $reflags{$_}) {
166 ? $reflags |= $reflags{$_}
167 : ($reflags &= ~$reflags{$_});
171 qq'Unknown regular expression flag "$_"'
176 ($^H{reflags} = $reflags or defined $^H{reflags_dul})
178 : ($^H &= ~$flags_hint);
181 Carp::carp("Unknown \"re\" subpragma '$s' (known ones are: ",
182 join(', ', map {qq('$_')} 'debug', 'debugcolor', sort keys %bitmask),
196 $^H &= ~ bits(0, @_);
205 re - Perl pragma to alter regular expression behaviour
210 ($x) = ($^X =~ /^(.*)$/s); # $x is tainted here
212 $pat = '(?{ $foo = 1 })';
214 /foo${pat}bar/; # won't fail (when not under -T switch)
217 no re 'taint'; # the default
218 ($x) = ($^X =~ /^(.*)$/s); # $x is not tainted here
220 no re 'eval'; # the default
221 /foo${pat}bar/; # disallowed (with or without -T switch)
225 "FOO" =~ / foo /; # /ix implied
227 "FOO" =~ /foo/; # just /i implied
229 use re 'debug'; # output debugging info during
230 /^(.*)$/s; # compile and run time
233 use re 'debugcolor'; # same as 'debug', but with colored output
236 use re qw(Debug All); # Finer tuned debugging options.
237 use re qw(Debug More);
238 no re qw(Debug ALL); # Turn of all re debugging in this scope
240 use re qw(is_regexp regexp_pattern); # import utility functions
241 my ($pat,$mods)=regexp_pattern(qr/foo/i);
242 if (is_regexp($obj)) {
243 print "Got regexp: ",
244 scalar regexp_pattern($obj); # just as perl would stringify it
245 } # but no hassle with blessed re's.
247 (We use $^X in these examples because it's tainted by default.)
253 When C<use re 'taint'> is in effect, and a tainted string is the target
254 of a regexp, the regexp memories (or values returned by the m// operator
255 in list context) are tainted. This feature is useful when regexp operations
256 on tainted data aren't meant to extract safe substrings, but to perform
257 other transformations.
261 When C<use re 'eval'> is in effect, a regexp is allowed to contain
262 C<(?{ ... })> zero-width assertions and C<(??{ ... })> postponed
263 subexpressions, even if the regular expression contains
264 variable interpolation. That is normally disallowed, since it is a
265 potential security risk. Note that this pragma is ignored when the regular
266 expression is obtained from tainted data, i.e. evaluation is always
267 disallowed with tainted regular expressions. See L<perlre/(?{ code })>
268 and L<perlre/(??{ code })>.
270 For the purpose of this pragma, interpolation of precompiled regular
271 expressions (i.e., the result of C<qr//>) is I<not> considered variable
276 I<is> allowed if $pat is a precompiled regular expression, even
277 if $pat contains C<(?{ ... })> assertions or C<(??{ ... })> subexpressions.
281 When C<use re '/flags'> is specified, the given flags are automatically
282 added to every regular expression till the end of the lexical scope.
284 C<no re '/flags'> will turn off the effect of C<use re '/flags'> for the
287 For example, if you want all your regular expressions to have /msx on by
292 at the top of your code.
294 The /dul flags cancel each other out. So, in this example,
301 The second C<use re> does an implicit C<no re '/u'>.
303 Turning on the /l and /u flags with C<use re> takes precedence over the
304 C<locale> pragma and the 'unicode_strings' C<feature>, for regular
305 expressions. Turning off one of these flags when it is active reverts to
306 the behaviour specified by whatever other pragmata are in scope. For
309 use feature "unicode_strings";
310 no re "/u"; # does nothing
312 no re "/l"; # reverts to unicode_strings behaviour
316 When C<use re 'debug'> is in effect, perl emits debugging messages when
317 compiling and using regular expressions. The output is the same as that
318 obtained by running a C<-DDEBUGGING>-enabled perl interpreter with the
319 B<-Dr> switch. It may be quite voluminous depending on the complexity
320 of the match. Using C<debugcolor> instead of C<debug> enables a
321 form of output that can be used to get a colorful display on terminals
322 that understand termcap color sequences. Set C<$ENV{PERL_RE_TC}> to a
323 comma-separated list of C<termcap> properties to use for highlighting
324 strings on/off, pre-point part on/off.
325 See L<perldebug/"Debugging regular expressions"> for additional info.
327 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
328 lexically scoped, as the other directives are. However they have both
329 compile-time and run-time effects.
331 See L<perlmodlib/Pragmatic Modules>.
335 Similarly C<use re 'Debug'> produces debugging output, the difference
336 being that it allows the fine tuning of what debugging output will be
337 emitted. Options are divided into three groups, those related to
338 compilation, those related to execution and those related to special
339 purposes. The options are as follows:
343 =item Compile related options
349 Turns on all compile related debug options.
353 Turns on debug output related to the process of parsing the pattern.
357 Enables output related to the optimisation phase of compilation.
361 Detailed info about trie compilation.
365 Dump the final program out after it is compiled and optimised.
369 =item Execute related options
375 Turns on all execute related debug options.
379 Turns on debugging of the main matching loop.
383 Extra debugging of how tries execute.
387 Enable debugging of start point optimisations.
391 =item Extra debugging options
397 Turns on all "extra" debugging options.
401 Enable debugging the capture group storage during match. Warning,
402 this can potentially produce extremely large output.
406 Enable enhanced TRIE debugging. Enhances both TRIEE
411 Enable debugging of states in the engine.
415 Enable debugging of the recursion stack in the engine. Enabling
416 or disabling this option automatically does the same for debugging
417 states as well. This output from this can be quite large.
421 Enable enhanced optimisation debugging and start point optimisations.
422 Probably not useful except when debugging the regexp engine itself.
426 Dump offset information. This can be used to see how regops correlate
427 to the pattern. Output format is
429 NODENUM:POSITION[LENGTH]
431 Where 1 is the position of the first char in the string. Note that position
432 can be 0, or larger than the actual length of the pattern, likewise length
437 Enable debugging of offsets information. This emits copious
438 amounts of trace information and doesn't mesh well with other
441 Almost definitely only useful to people hacking
442 on the offsets part of the debug engine.
446 =item Other useful flags
448 These are useful shortcuts to save on the typing.
454 Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS
458 Enable DUMP and all execute options. Equivalent to:
466 Enable TRIEM and all execute compile and execute options.
472 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
473 lexically scoped, as the other directives are. However they have both
474 compile-time and run-time effects.
476 =head2 Exportable Functions
478 As of perl 5.9.5 're' debug contains a number of utility functions that
479 may be optionally exported into the caller's namespace. They are listed
484 =item is_regexp($ref)
486 Returns true if the argument is a compiled regular expression as returned
487 by C<qr//>, false if it is not.
489 This function will not be confused by overloading or blessing. In
490 internals terms, this extracts the regexp pointer out of the
491 PERL_MAGIC_qr structure so it cannot be fooled.
493 =item regexp_pattern($ref)
495 If the argument is a compiled regular expression as returned by C<qr//>,
496 then this function returns the pattern.
498 In list context it returns a two element list, the first element
499 containing the pattern and the second containing the modifiers used when
500 the pattern was compiled.
502 my ($pat, $mods) = regexp_pattern($ref);
504 In scalar context it returns the same as perl would when stringifying a raw
505 C<qr//> with the same pattern inside. If the argument is not a compiled
506 reference then this routine returns false but defined in scalar context,
507 and the empty list in list context. Thus the following
509 if (regexp_pattern($ref) eq '(?^i:foo)')
511 will be warning free regardless of what $ref actually is.
513 Like C<is_regexp> this function will not be confused by overloading
514 or blessing of the object.
518 If the argument is a compiled regular expression as returned by C<qr//>,
519 then this function returns what the optimiser considers to be the longest
520 anchored fixed string and longest floating fixed string in the pattern.
522 A I<fixed string> is defined as being a substring that must appear for the
523 pattern to match. An I<anchored fixed string> is a fixed string that must
524 appear at a particular offset from the beginning of the match. A I<floating
525 fixed string> is defined as a fixed string that can appear at any point in
526 a range of positions relative to the start of the match. For example,
528 my $qr = qr/here .* there/x;
529 my ($anchored, $floating) = regmust($qr);
530 print "anchored:'$anchored'\nfloating:'$floating'\n";
537 Because the C<here> is before the C<.*> in the pattern, its position
538 can be determined exactly. That's not true, however, for the C<there>;
539 it could appear at any point after where the anchored string appeared.
540 Perl uses both for its optimisations, prefering the longer, or, if they are
543 B<NOTE:> This may not necessarily be the definitive longest anchored and
544 floating string. This will be what the optimiser of the Perl that you
545 are using thinks is the longest. If you believe that the result is wrong
546 please report it via the L<perlbug> utility.
548 =item regname($name,$all)
550 Returns the contents of a named buffer of the last successful match. If
551 $all is true, then returns an array ref containing one entry per buffer,
552 otherwise returns the first defined buffer.
556 Returns a list of all of the named buffers defined in the last successful
557 match. If $all is true, then it returns all names defined, if not it returns
558 only names which were involved in the match.
560 =item regnames_count()
562 Returns the number of distinct names defined in the pattern used
563 for the last successful match.
565 B<Note:> this result is always the actual number of distinct
566 named buffers defined, it may not actually match that which is
567 returned by C<regnames()> and related routines when those routines
568 have not been called with the $all parameter set.
574 L<perlmodlib/Pragmatic Modules>.