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
20 eval { # Ignore errors
23 my $terminal = Tgetent Term::Cap ({OSPEED => 9600}); # Avoid warning.
24 my $props = $ENV{PERL_RE_TC} || 'md,me,so,se,us,ue';
25 my @props = split /,/, $props;
26 my $colors = join "\t", map {$terminal->Tputs($_,1)} @props;
29 $ENV{PERL_RE_COLORS} = $colors;
32 $ENV{PERL_RE_COLORS} ||= qq'\t\t> <\t> <\t\t';
53 OFFSETSDBG => 0x040000,
55 OPTIMISEM => 0x100000,
60 $flags{ALL} = -1 & ~($flags{OFFSETS}|$flags{OFFSETSDBG}|$flags{BUFFERS});
61 $flags{All} = $flags{all} = $flags{DUMP} | $flags{EXECUTE};
62 $flags{Extra} = $flags{EXECUTE} | $flags{COMPILE} | $flags{GPOS};
63 $flags{More} = $flags{MORE} = $flags{All} | $flags{TRIEC} | $flags{TRIEM} | $flags{STATE};
64 $flags{State} = $flags{DUMP} | $flags{EXECUTE} | $flags{STATE};
65 $flags{TRIE} = $flags{DUMP} | $flags{EXECUTE} | $flags{TRIEC};
67 if (defined &DynaLoader::boot_DynaLoader) {
69 XSLoader::load( __PACKAGE__, $VERSION);
72 # We need to work for miniperl, because the XS toolchain uses Text::Wrap, which
78 # We call install() every time, as if we didn't, we wouldn't
79 # "see" any changes to the color environment var since
80 # the last time it was called.
82 # install() returns an integer, which if casted properly
83 # in C resolves to a structure containing the regexp
84 # hooks. Setting it to a random integer will guarantee
86 $^H{regcomp} = install();
97 Carp::carp("Useless use of \"re\" pragma");
99 foreach my $idx (0..$#_){
101 if ($s eq 'Debug' or $s eq 'Debugcolor') {
102 setcolor() if $s =~/color/i;
103 ${^RE_DEBUG_FLAGS} = 0 unless defined ${^RE_DEBUG_FLAGS};
104 for my $idx ($idx+1..$#_) {
105 if ($flags{$_[$idx]}) {
107 ${^RE_DEBUG_FLAGS} |= $flags{$_[$idx]};
109 ${^RE_DEBUG_FLAGS} &= ~ $flags{$_[$idx]};
113 Carp::carp("Unknown \"re\" Debug flag '$_[$idx]', possible flags: ",
114 join(", ",sort keys %flags ) );
117 _load_unload($on ? 1 : ${^RE_DEBUG_FLAGS});
119 } elsif ($s eq 'debug' or $s eq 'debugcolor') {
120 setcolor() if $s =~/color/i;
123 } elsif (exists $bitmask{$s}) {
124 $bits |= $bitmask{$s};
125 } elsif ($EXPORT_OK{$s}) {
127 re->export_to_level(2, 're', $s);
130 Carp::carp("Unknown \"re\" subpragma '$s' (known ones are: ",
131 join(', ', map {qq('$_')} 'debug', 'debugcolor', sort keys %bitmask),
145 $^H &= ~ bits(0, @_);
154 re - Perl pragma to alter regular expression behaviour
159 ($x) = ($^X =~ /^(.*)$/s); # $x is tainted here
161 $pat = '(?{ $foo = 1 })';
163 /foo${pat}bar/; # won't fail (when not under -T switch)
166 no re 'taint'; # the default
167 ($x) = ($^X =~ /^(.*)$/s); # $x is not tainted here
169 no re 'eval'; # the default
170 /foo${pat}bar/; # disallowed (with or without -T switch)
173 use re 'debug'; # output debugging info during
174 /^(.*)$/s; # compile and run time
177 use re 'debugcolor'; # same as 'debug', but with colored output
180 use re qw(Debug All); # Finer tuned debugging options.
181 use re qw(Debug More);
182 no re qw(Debug ALL); # Turn of all re debugging in this scope
184 use re qw(is_regexp regexp_pattern); # import utility functions
185 my ($pat,$mods)=regexp_pattern(qr/foo/i);
186 if (is_regexp($obj)) {
187 print "Got regexp: ",
188 scalar regexp_pattern($obj); # just as perl would stringify it
189 } # but no hassle with blessed re's.
191 (We use $^X in these examples because it's tainted by default.)
197 When C<use re 'taint'> is in effect, and a tainted string is the target
198 of a regexp, the regexp memories (or values returned by the m// operator
199 in list context) are tainted. This feature is useful when regexp operations
200 on tainted data aren't meant to extract safe substrings, but to perform
201 other transformations.
205 When C<use re 'eval'> is in effect, a regexp is allowed to contain
206 C<(?{ ... })> zero-width assertions and C<(??{ ... })> postponed
207 subexpressions, even if the regular expression contains
208 variable interpolation. That is normally disallowed, since it is a
209 potential security risk. Note that this pragma is ignored when the regular
210 expression is obtained from tainted data, i.e. evaluation is always
211 disallowed with tainted regular expressions. See L<perlre/(?{ code })>
212 and L<perlre/(??{ code })>.
214 For the purpose of this pragma, interpolation of precompiled regular
215 expressions (i.e., the result of C<qr//>) is I<not> considered variable
220 I<is> allowed if $pat is a precompiled regular expression, even
221 if $pat contains C<(?{ ... })> assertions or C<(??{ ... })> subexpressions.
225 When C<use re 'debug'> is in effect, perl emits debugging messages when
226 compiling and using regular expressions. The output is the same as that
227 obtained by running a C<-DDEBUGGING>-enabled perl interpreter with the
228 B<-Dr> switch. It may be quite voluminous depending on the complexity
229 of the match. Using C<debugcolor> instead of C<debug> enables a
230 form of output that can be used to get a colorful display on terminals
231 that understand termcap color sequences. Set C<$ENV{PERL_RE_TC}> to a
232 comma-separated list of C<termcap> properties to use for highlighting
233 strings on/off, pre-point part on/off.
234 See L<perldebug/"Debugging regular expressions"> for additional info.
236 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
237 lexically scoped, as the other directives are. However they have both
238 compile-time and run-time effects.
240 See L<perlmodlib/Pragmatic Modules>.
244 Similarly C<use re 'Debug'> produces debugging output, the difference
245 being that it allows the fine tuning of what debugging output will be
246 emitted. Options are divided into three groups, those related to
247 compilation, those related to execution and those related to special
248 purposes. The options are as follows:
252 =item Compile related options
258 Turns on all compile related debug options.
262 Turns on debug output related to the process of parsing the pattern.
266 Enables output related to the optimisation phase of compilation.
270 Detailed info about trie compilation.
274 Dump the final program out after it is compiled and optimised.
278 =item Execute related options
284 Turns on all execute related debug options.
288 Turns on debugging of the main matching loop.
292 Extra debugging of how tries execute.
296 Enable debugging of start point optimisations.
300 =item Extra debugging options
306 Turns on all "extra" debugging options.
310 Enable debugging the capture buffer storage during match. Warning,
311 this can potentially produce extremely large output.
315 Enable enhanced TRIE debugging. Enhances both TRIEE
320 Enable debugging of states in the engine.
324 Enable debugging of the recursion stack in the engine. Enabling
325 or disabling this option automatically does the same for debugging
326 states as well. This output from this can be quite large.
330 Enable enhanced optimisation debugging and start point optimisations.
331 Probably not useful except when debugging the regexp engine itself.
335 Dump offset information. This can be used to see how regops correlate
336 to the pattern. Output format is
338 NODENUM:POSITION[LENGTH]
340 Where 1 is the position of the first char in the string. Note that position
341 can be 0, or larger than the actual length of the pattern, likewise length
346 Enable debugging of offsets information. This emits copious
347 amounts of trace information and doesn't mesh well with other
350 Almost definitely only useful to people hacking
351 on the offsets part of the debug engine.
355 =item Other useful flags
357 These are useful shortcuts to save on the typing.
363 Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS
367 Enable DUMP and all execute options. Equivalent to:
375 Enable TRIEM and all execute compile and execute options.
381 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
382 lexically scoped, as the other directives are. However they have both
383 compile-time and run-time effects.
385 =head2 Exportable Functions
387 As of perl 5.9.5 're' debug contains a number of utility functions that
388 may be optionally exported into the caller's namespace. They are listed
393 =item is_regexp($ref)
395 Returns true if the argument is a compiled regular expression as returned
396 by C<qr//>, false if it is not.
398 This function will not be confused by overloading or blessing. In
399 internals terms, this extracts the regexp pointer out of the
400 PERL_MAGIC_qr structure so it it cannot be fooled.
402 =item regexp_pattern($ref)
404 If the argument is a compiled regular expression as returned by C<qr//>,
405 then this function returns the pattern.
407 In list context it returns a two element list, the first element
408 containing the pattern and the second containing the modifiers used when
409 the pattern was compiled.
411 my ($pat, $mods) = regexp_pattern($ref);
413 In scalar context it returns the same as perl would when stringifying a raw
414 C<qr//> with the same pattern inside. If the argument is not a compiled
415 reference then this routine returns false but defined in scalar context,
416 and the empty list in list context. Thus the following
418 if (regexp_pattern($ref) eq '(?i-xsm:foo)')
420 will be warning free regardless of what $ref actually is.
422 Like C<is_regexp> this function will not be confused by overloading
423 or blessing of the object.
427 If the argument is a compiled regular expression as returned by C<qr//>,
428 then this function returns what the optimiser considers to be the longest
429 anchored fixed string and longest floating fixed string in the pattern.
431 A I<fixed string> is defined as being a substring that must appear for the
432 pattern to match. An I<anchored fixed string> is a fixed string that must
433 appear at a particular offset from the beginning of the match. A I<floating
434 fixed string> is defined as a fixed string that can appear at any point in
435 a range of positions relative to the start of the match. For example,
437 my $qr = qr/here .* there/x;
438 my ($anchored, $floating) = regmust($qr);
439 print "anchored:'$anchored'\nfloating:'$floating'\n";
446 Because the C<here> is before the C<.*> in the pattern, its position
447 can be determined exactly. That's not true, however, for the C<there>;
448 it could appear at any point after where the anchored string appeared.
449 Perl uses both for its optimisations, prefering the longer, or, if they are
452 B<NOTE:> This may not necessarily be the definitive longest anchored and
453 floating string. This will be what the optimiser of the Perl that you
454 are using thinks is the longest. If you believe that the result is wrong
455 please report it via the L<perlbug> utility.
457 =item regname($name,$all)
459 Returns the contents of a named buffer of the last successful match. If
460 $all is true, then returns an array ref containing one entry per buffer,
461 otherwise returns the first defined buffer.
465 Returns a list of all of the named buffers defined in the last successful
466 match. If $all is true, then it returns all names defined, if not it returns
467 only names which were involved in the match.
469 =item regnames_count()
471 Returns the number of distinct names defined in the pattern used
472 for the last successful match.
474 B<Note:> this result is always the actual number of distinct
475 named buffers defined, it may not actually match that which is
476 returned by C<regnames()> and related routines when those routines
477 have not been called with the $all parameter set.
483 L<perlmodlib/Pragmatic Modules>.