3 # pragma for controlling the regex engine
8 our @ISA = qw(Exporter);
9 our @EXPORT_OK = qw(is_regexp regexp_pattern regmust);
10 our %EXPORT_OK = map { $_ => 1 } @EXPORT_OK;
12 # *** WARNING *** WARNING *** WARNING *** WARNING *** WARNING ***
14 # If you modify these values see comment below!
17 taint => 0x00100000, # HINT_RE_TAINT
18 eval => 0x00200000, # HINT_RE_EVAL
21 # - File::Basename contains a literal for 'taint' as a fallback. If
22 # taint is changed here, File::Basename must be updated as well.
24 # - ExtUtils::ParseXS uses a hardcoded
25 # BEGIN { $^H |= 0x00200000 }
26 # in it to allow re.xs to be built. So if 'eval' is changed here then
27 # ExtUtils::ParseXS must be changed as well.
29 # *** WARNING *** WARNING *** WARNING *** WARNING *** WARNING ***
32 eval { # Ignore errors
35 my $terminal = Tgetent Term::Cap ({OSPEED => 9600}); # Avoid warning.
36 my $props = $ENV{PERL_RE_TC} || 'md,me,so,se,us,ue';
37 my @props = split /,/, $props;
38 my $colors = join "\t", map {$terminal->Tputs($_,1)} @props;
41 $ENV{PERL_RE_COLORS} = $colors;
44 $ENV{PERL_RE_COLORS} ||= qq'\t\t> <\t> <\t\t';
64 OFFSETSDBG => 0x040000,
66 OPTIMISEM => 0x100000,
70 $flags{All} = $flags{all} = $flags{DUMP} | $flags{EXECUTE};
71 $flags{Extra} = $flags{EXECUTE} | $flags{COMPILE};
72 $flags{More} = $flags{MORE} = $flags{All} | $flags{TRIEC} | $flags{TRIEM} | $flags{STATE};
73 $flags{State} = $flags{DUMP} | $flags{EXECUTE} | $flags{STATE};
74 $flags{TRIE} = $flags{DUMP} | $flags{EXECUTE} | $flags{TRIEC};
80 if ( ! defined($installed) ) {
82 $installed = eval { XSLoader::load('re', $VERSION) } || 0;
83 $installed_error = $@;
92 die "'re' not installed!? ($installed_error)";
94 # We call install() every time, as if we didn't, we wouldn't
95 # "see" any changes to the color environment var since
96 # the last time it was called.
98 # install() returns an integer, which if casted properly
99 # in C resolves to a structure containing the regex
100 # hooks. Setting it to a random integer will guarantee
102 $^H{regcomp} = install();
114 Carp::carp("Useless use of \"re\" pragma");
116 foreach my $idx (0..$#_){
118 if ($s eq 'Debug' or $s eq 'Debugcolor') {
119 setcolor() if $s =~/color/i;
120 ${^RE_DEBUG_FLAGS} = 0 unless defined ${^RE_DEBUG_FLAGS};
121 for my $idx ($idx+1..$#_) {
122 if ($flags{$_[$idx]}) {
124 ${^RE_DEBUG_FLAGS} |= $flags{$_[$idx]};
126 ${^RE_DEBUG_FLAGS} &= ~ $flags{$_[$idx]};
130 Carp::carp("Unknown \"re\" Debug flag '$_[$idx]', possible flags: ",
131 join(", ",sort keys %flags ) );
134 _load_unload($on ? 1 : ${^RE_DEBUG_FLAGS});
136 } elsif ($s eq 'debug' or $s eq 'debugcolor') {
137 setcolor() if $s =~/color/i;
139 } elsif (exists $bitmask{$s}) {
140 $bits |= $bitmask{$s};
141 } elsif ($EXPORT_OK{$s}) {
144 re->export_to_level(2, 're', $s);
147 Carp::carp("Unknown \"re\" subpragma '$s' (known ones are: ",
148 join(', ', map {qq('$_')} 'debug', 'debugcolor', sort keys %bitmask),
162 $^H &= ~ bits(0, @_);
171 re - Perl pragma to alter regular expression behaviour
176 ($x) = ($^X =~ /^(.*)$/s); # $x is tainted here
178 $pat = '(?{ $foo = 1 })';
180 /foo${pat}bar/; # won't fail (when not under -T switch)
183 no re 'taint'; # the default
184 ($x) = ($^X =~ /^(.*)$/s); # $x is not tainted here
186 no re 'eval'; # the default
187 /foo${pat}bar/; # disallowed (with or without -T switch)
190 use re 'debug'; # output debugging info during
191 /^(.*)$/s; # compile and run time
194 use re 'debugcolor'; # same as 'debug', but with colored output
197 use re qw(Debug All); # Finer tuned debugging options.
198 use re qw(Debug More);
199 no re qw(Debug ALL); # Turn of all re debugging in this scope
201 use re qw(is_regexp regexp_pattern); # import utility functions
202 my ($pat,$mods)=regexp_pattern(qr/foo/i);
203 if (is_regexp($obj)) {
204 print "Got regexp: ",
205 scalar regexp_pattern($obj); # just as perl would stringify it
206 } # but no hassle with blessed re's.
208 (We use $^X in these examples because it's tainted by default.)
214 When C<use re 'taint'> is in effect, and a tainted string is the target
215 of a regex, the regex memories (or values returned by the m// operator
216 in list context) are tainted. This feature is useful when regex operations
217 on tainted data aren't meant to extract safe substrings, but to perform
218 other transformations.
222 When C<use re 'eval'> is in effect, a regex is allowed to contain
223 C<(?{ ... })> zero-width assertions even if regular expression contains
224 variable interpolation. That is normally disallowed, since it is a
225 potential security risk. Note that this pragma is ignored when the regular
226 expression is obtained from tainted data, i.e. evaluation is always
227 disallowed with tainted regular expressions. See L<perlre/(?{ code })>.
229 For the purpose of this pragma, interpolation of precompiled regular
230 expressions (i.e., the result of C<qr//>) is I<not> considered variable
235 I<is> allowed if $pat is a precompiled regular expression, even
236 if $pat contains C<(?{ ... })> assertions.
240 When C<use re 'debug'> is in effect, perl emits debugging messages when
241 compiling and using regular expressions. The output is the same as that
242 obtained by running a C<-DDEBUGGING>-enabled perl interpreter with the
243 B<-Dr> switch. It may be quite voluminous depending on the complexity
244 of the match. Using C<debugcolor> instead of C<debug> enables a
245 form of output that can be used to get a colorful display on terminals
246 that understand termcap color sequences. Set C<$ENV{PERL_RE_TC}> to a
247 comma-separated list of C<termcap> properties to use for highlighting
248 strings on/off, pre-point part on/off.
249 See L<perldebug/"Debugging regular expressions"> for additional info.
251 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
252 lexically scoped, as the other directives are. However they have both
253 compile-time and run-time effects.
255 See L<perlmodlib/Pragmatic Modules>.
259 Similarly C<use re 'Debug'> produces debugging output, the difference
260 being that it allows the fine tuning of what debugging output will be
261 emitted. Options are divided into three groups, those related to
262 compilation, those related to execution and those related to special
263 purposes. The options are as follows:
267 =item Compile related options
273 Turns on all compile related debug options.
277 Turns on debug output related to the process of parsing the pattern.
281 Enables output related to the optimisation phase of compilation.
285 Detailed info about trie compilation.
289 Dump the final program out after it is compiled and optimised.
293 =item Execute related options
299 Turns on all execute related debug options.
303 Turns on debugging of the main matching loop.
307 Extra debugging of how tries execute.
311 Enable debugging of start point optimisations.
315 =item Extra debugging options
321 Turns on all "extra" debugging options.
325 Enable enhanced TRIE debugging. Enhances both TRIEE
330 Enable debugging of states in the engine.
334 Enable debugging of the recursion stack in the engine. Enabling
335 or disabling this option automatically does the same for debugging
336 states as well. This output from this can be quite large.
340 Enable enhanced optimisation debugging and start point optimisations.
341 Probably not useful except when debugging the regex engine itself.
345 Dump offset information. This can be used to see how regops correlate
346 to the pattern. Output format is
348 NODENUM:POSITION[LENGTH]
350 Where 1 is the position of the first char in the string. Note that position
351 can be 0, or larger than the actual length of the pattern, likewise length
356 Enable debugging of offsets information. This emits copious
357 amounts of trace information and doesn't mesh well with other
360 Almost definitely only useful to people hacking
361 on the offsets part of the debug engine.
365 =item Other useful flags
367 These are useful shortcuts to save on the typing.
373 Enable all compile and execute options at once.
377 Enable DUMP and all execute options. Equivalent to:
385 Enable TRIEM and all execute compile and execute options.
391 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
392 lexically scoped, as the other directives are. However they have both
393 compile-time and run-time effects.
395 =head2 Exportable Functions
397 As of perl 5.9.5 're' debug contains a number of utility functions that
398 may be optionally exported into the caller's namespace. They are listed
403 =item is_regexp($ref)
405 Returns true if the argument is a compiled regular expression as returned
406 by C<qr//>, false if it is not.
408 This function will not be confused by overloading or blessing. In
409 internals terms, this extracts the regexp pointer out of the
410 PERL_MAGIC_qr structure so it it cannot be fooled.
412 =item regexp_pattern($ref)
414 If the argument is a compiled regular expression as returned by C<qr//>,
415 then this function returns the pattern.
417 In list context it returns a two element list, the first element
418 containing the pattern and the second containing the modifiers used when
419 the pattern was compiled.
421 my ($pat, $mods) = regexp_pattern($ref);
423 In scalar context it returns the same as perl would when strigifying a raw
424 C<qr//> with the same pattern inside. If the argument is not a compiled
425 reference then this routine returns false but defined in scalar context,
426 and the empty list in list context. Thus the following
428 if (regexp_pattern($ref) eq '(?i-xsm:foo)')
430 will be warning free regardless of what $ref actually is.
432 Like C<is_regexp> this function will not be confused by overloading
433 or blessing of the object.
437 If the argument is a compiled regular expression as returned by C<qr//>,
438 then this function returns what the optimiser consiers to be the longest
439 anchored fixed string and longest floating fixed string in the pattern.
441 A I<fixed string> is defined as being a substring that must appear for the
442 pattern to match. An I<anchored fixed string> is a fixed string that must
443 appear at a particular offset from the beginning of the match. A I<floating
444 fixed string> is defined as a fixed string that can appear at any point in
445 a range of positions relative to the start of the match. For example,
447 my $qr = qr/here .* there/x;
448 my ($anchored, $floating) = regmust($qr);
449 print "anchored:'$anchored'\nfloating:'$floating'\n";
456 Because the C<here> is before the C<.*> in the pattern, its position
457 can be determined exactly. That's not true, however, for the C<there>;
458 it could appear at any point after where the anchored string appeared.
459 Perl uses both for its optimisations, prefering the longer, or, if they are
462 B<NOTE:> This may not necessarily be the definitive longest anchored and
463 floating string. This will be what the optimiser of the Perl that you
464 are using thinks is the longest. If you believe that the result is wrong
465 please report it via the L<perlbug> utility.
471 L<perlmodlib/Pragmatic Modules>.