3 # pragma for controlling the regex 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 regex
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 regex, the regex memories (or values returned by the m// operator
199 in list context) are tainted. This feature is useful when regex 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 regex is allowed to contain
206 C<(?{ ... })> zero-width assertions even if regular expression contains
207 variable interpolation. That is normally disallowed, since it is a
208 potential security risk. Note that this pragma is ignored when the regular
209 expression is obtained from tainted data, i.e. evaluation is always
210 disallowed with tainted regular expressions. See L<perlre/(?{ code })>.
212 For the purpose of this pragma, interpolation of precompiled regular
213 expressions (i.e., the result of C<qr//>) is I<not> considered variable
218 I<is> allowed if $pat is a precompiled regular expression, even
219 if $pat contains C<(?{ ... })> assertions.
223 When C<use re 'debug'> is in effect, perl emits debugging messages when
224 compiling and using regular expressions. The output is the same as that
225 obtained by running a C<-DDEBUGGING>-enabled perl interpreter with the
226 B<-Dr> switch. It may be quite voluminous depending on the complexity
227 of the match. Using C<debugcolor> instead of C<debug> enables a
228 form of output that can be used to get a colorful display on terminals
229 that understand termcap color sequences. Set C<$ENV{PERL_RE_TC}> to a
230 comma-separated list of C<termcap> properties to use for highlighting
231 strings on/off, pre-point part on/off.
232 See L<perldebug/"Debugging regular expressions"> for additional info.
234 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
235 lexically scoped, as the other directives are. However they have both
236 compile-time and run-time effects.
238 See L<perlmodlib/Pragmatic Modules>.
242 Similarly C<use re 'Debug'> produces debugging output, the difference
243 being that it allows the fine tuning of what debugging output will be
244 emitted. Options are divided into three groups, those related to
245 compilation, those related to execution and those related to special
246 purposes. The options are as follows:
250 =item Compile related options
256 Turns on all compile related debug options.
260 Turns on debug output related to the process of parsing the pattern.
264 Enables output related to the optimisation phase of compilation.
268 Detailed info about trie compilation.
272 Dump the final program out after it is compiled and optimised.
276 =item Execute related options
282 Turns on all execute related debug options.
286 Turns on debugging of the main matching loop.
290 Extra debugging of how tries execute.
294 Enable debugging of start point optimisations.
298 =item Extra debugging options
304 Turns on all "extra" debugging options.
308 Enable debugging the capture buffer storage during match. Warning,
309 this can potentially produce extremely large output.
313 Enable enhanced TRIE debugging. Enhances both TRIEE
318 Enable debugging of states in the engine.
322 Enable debugging of the recursion stack in the engine. Enabling
323 or disabling this option automatically does the same for debugging
324 states as well. This output from this can be quite large.
328 Enable enhanced optimisation debugging and start point optimisations.
329 Probably not useful except when debugging the regex engine itself.
333 Dump offset information. This can be used to see how regops correlate
334 to the pattern. Output format is
336 NODENUM:POSITION[LENGTH]
338 Where 1 is the position of the first char in the string. Note that position
339 can be 0, or larger than the actual length of the pattern, likewise length
344 Enable debugging of offsets information. This emits copious
345 amounts of trace information and doesn't mesh well with other
348 Almost definitely only useful to people hacking
349 on the offsets part of the debug engine.
353 =item Other useful flags
355 These are useful shortcuts to save on the typing.
361 Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS
365 Enable DUMP and all execute options. Equivalent to:
373 Enable TRIEM and all execute compile and execute options.
379 As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
380 lexically scoped, as the other directives are. However they have both
381 compile-time and run-time effects.
383 =head2 Exportable Functions
385 As of perl 5.9.5 're' debug contains a number of utility functions that
386 may be optionally exported into the caller's namespace. They are listed
391 =item is_regexp($ref)
393 Returns true if the argument is a compiled regular expression as returned
394 by C<qr//>, false if it is not.
396 This function will not be confused by overloading or blessing. In
397 internals terms, this extracts the regexp pointer out of the
398 PERL_MAGIC_qr structure so it it cannot be fooled.
400 =item regexp_pattern($ref)
402 If the argument is a compiled regular expression as returned by C<qr//>,
403 then this function returns the pattern.
405 In list context it returns a two element list, the first element
406 containing the pattern and the second containing the modifiers used when
407 the pattern was compiled.
409 my ($pat, $mods) = regexp_pattern($ref);
411 In scalar context it returns the same as perl would when strigifying a raw
412 C<qr//> with the same pattern inside. If the argument is not a compiled
413 reference then this routine returns false but defined in scalar context,
414 and the empty list in list context. Thus the following
416 if (regexp_pattern($ref) eq '(?i-xsm:foo)')
418 will be warning free regardless of what $ref actually is.
420 Like C<is_regexp> this function will not be confused by overloading
421 or blessing of the object.
425 If the argument is a compiled regular expression as returned by C<qr//>,
426 then this function returns what the optimiser consiers to be the longest
427 anchored fixed string and longest floating fixed string in the pattern.
429 A I<fixed string> is defined as being a substring that must appear for the
430 pattern to match. An I<anchored fixed string> is a fixed string that must
431 appear at a particular offset from the beginning of the match. A I<floating
432 fixed string> is defined as a fixed string that can appear at any point in
433 a range of positions relative to the start of the match. For example,
435 my $qr = qr/here .* there/x;
436 my ($anchored, $floating) = regmust($qr);
437 print "anchored:'$anchored'\nfloating:'$floating'\n";
444 Because the C<here> is before the C<.*> in the pattern, its position
445 can be determined exactly. That's not true, however, for the C<there>;
446 it could appear at any point after where the anchored string appeared.
447 Perl uses both for its optimisations, prefering the longer, or, if they are
450 B<NOTE:> This may not necessarily be the definitive longest anchored and
451 floating string. This will be what the optimiser of the Perl that you
452 are using thinks is the longest. If you believe that the result is wrong
453 please report it via the L<perlbug> utility.
455 =item regname($name,$all)
457 Returns the contents of a named buffer of the last successful match. If
458 $all is true, then returns an array ref containing one entry per buffer,
459 otherwise returns the first defined buffer.
463 Returns a list of all of the named buffers defined in the last successful
464 match. If $all is true, then it returns all names defined, if not it returns
465 only names which were involved in the match.
467 =item regnames_count()
469 Returns the number of distinct names defined in the pattern used
470 for the last successful match.
472 B<Note:> this result is always the actual number of distinct
473 named buffers defined, it may not actually match that which is
474 returned by C<regnames()> and related routines when those routines
475 have not been called with the $all parameter set.
481 L<perlmodlib/Pragmatic Modules>.